Will Analytics ever work in the NFL like other sports?

in the opener, but decided to not make that opening post too long and will instead post as a response:

Quote:

Where is the NFL’s Version of Moneyball? Measuring player success in football is extremely difficult There is a reason that baseball is the poster child for sports analytics. It is a binary game. On every play, you either succeed or fail. There is very, very little room for gray area in this regard. Basketball is not too far behind. There exists an extra layer of subjectivity when compared to baseball, but still, a relatively binary nature is still present. Football is quite the opposite. Think of a running back that takes a broken play, where he should be tackled for a loss, and turns it into a 3-yard gain after making multiple defender miss. Then compare that to a play where the blocking is perfect, or the defense completely misses their assignment, and he runs for 50 yards without ever being touched. Which play is truly more indicative of the type of player the running back is? The answer is almost surely the former, but translating that into something quantifiable is a tall task. There are few comparable grey areas of success to a heroic 3 yard run in other sports. They are pass/fail. Did you get a hit or did you get out? Did you make the shot or did you miss? In football, examples of subjective success are nearly endless. What about an offensive lineman who drives his man 4 yards off the ball, but as the running back hits the hole, he ever-so slightly slips off the block and the defender makes an arm tackle for a 3-4 yard gain. I would look at that play and say that the OL was the reason a positive gain was achieved in the first place and getting that much movement off the line of scrimmage was an incredibly impressive individual effort. Whereas, somebody else grading that play might think I’m crazy and the entire blame would be placed on the OL whose defender made the play. Furthermore, if you are not part of the staff that drew up the play, you often will not know what was trying to be accomplished. As a former offensive lineman, I know all too well that many times you can be doing your job just fine but something in the play goes array, the ball bounces where you weren’t expecting it go and your man makes the play. To an un-educated eye, the offensive lineman clearly is at fault, but should he be? Well, the answer there is yes, because it’s always the offensive line’s fault, but the point is, if you don’t know the objective or scheme of the play, placing blame or credit can be extremely difficult. Variables, Variables, Variables When trying to obtain actionable insights from data, the make-up of your data is incredibly important. The easiest data to extract actual insights from will include a massive sample size with a limited number of variables. The ‘story’ of these numbers will not be hard to follow and even the most basic analytic tools can be useful when data is constructed like this. At a high level, compared to the other major sports, football has the most variables (22 players on the field) combined with the smallest sample size (16 games). Going even further, simply having the most players on the field does not come close to telling the full story when it comes to the amount variables at work on a given football play. Each player on the field has a job on each play, and depending on the play, each player will have an actual effect on the play. In this way, there is no other sport like it. In baseball, the only people involved on a given play are the hitter, pitcher, catcher and whoever the ball is hit toward. In basketball (especially the NBA), while team defensive positioning is important, the game, at its core, still comes down to the man with ball and the man covering him. Similar to my sentiment regarding understanding the concept of a play, with the amount people influencing a play, it can become extremely hard to determine who should be credited or blamed with the success or failure of a given play. Another variable that is unique to football is the lack of uniformity within the construction of a given play or series. In baseball, each play starts the exact same way, with absolutely no variation. In basketball, despite the fluid nature of the game, a natural up and down progression is followed. This is not the case in football. Where you are on the field, what the down and distance is, and whether it is a run play or pass play are all completely unique variables that exist only in football. Each play has a certain context around it that must be accounted for and because of that, gathering usable data is extremely difficult. A run that only gains 1 yard would, 95% of the time, be considered failure. But, if it was 3rd and inches, the opposite is true. Outside of extremely rare circumstances, missing a shot in basketball or getting an out in baseball, is always considered a negative outcome for the offense/hitting team. There are so many contextual variables for each play that occurs on a football field, variables that both do not exist in any other sports and variables that, when trying to gather usable data, must be quantified and accounted for. Where Do We Go From Here? All that was said above is meant to serve as an explanation for why we do not hear the analytics success stories that we hear in other sports in the NFL. This does not mean that analytics will not make the proverbial leap and will never become a fixture of decision making processes in NFL front offices and coaching booths. In fact, I believe the opposite is true. It is simply a matter of time until analytics finds its footing and the NFL has its version of Moneyball. With that said, the question must be asked: where are the most likely spots for us to see analytics shine through in the game of football? 1. League-wide changes in style of play We have seen this occur time after time in the NFL even without the helping hand of big data. A football game in the mid-70’s seldom resembles a modern-day NFL game and this evolution will only continue to unfold. What will be the NFL version of three’s and layups? Could it be an exacerbation of the current wide-open NFL offense that moves closer and closer to a college offense? Or will it be a steady dose of run plays and play actions passes? The answer will lie in the numbers. 2. Quarterback evaluation Many advanced metrics, while still not finely tuned, currently exist around the quarterback position. As mentioned previously, using numbers to evaluate football players is incredibly difficult and, in one sentence, is the reason why this article is being written. But, if there is to be a position where the code is truly cracked, it will be the quarterback. Some metric will emerge that quantifies how ‘good’ every throw is based on the pressure the quarterback is under, multiplied by the space available to put the ball, plus the distance the ball is thrown (or something along those lines). Naming the top 5-7 quarterbacks in the league is not a difficult task. Conversely, differentiating between the 9th and 15th best quarterbacks in the league is nearly impossible. Who is really better between Kirk Cousins and Case Keenum? Alex Smith or Blake Bortles? These answers often change on a yearly basis given scheme and surrounding talent; currently, these questions are a crap-shoot for NFL front offices, but someday, with the aid of advanced numbers, more clarity will emerge. 3. Play-calling This will be one of the final frontiers of analytics in the game in football. Coaches already carry around their “chart” which, using numbers, lays out when they should/shouldn’t go for it on 4th down, when they can begin to take a knee along with other various situational breakdowns. This is Version 1.0 of what analytics driven play calling will look like. Someday, every play-caller in the NFL will have a line in their headset that patches them through to their analytics ‘guy’ who can provide the statistically best play for every situation that is encountered on the football field. There will be a synthesis of what currently exists and a new analytics driven play calling system. How much teams lean on analytics to drive their play calling will differ greatly amongst different teams, but in a game of trends, where everyone is looking for an edge, you are crazy to think that teams will not get creative with this type of stuff. The early success of these initiatives will be crucial in determining how widespread it becomes, but all it will take is one success story, and teams will be all but forced to test their luck. Analytics has not found its true place in the game of football. The reason for this is more about the way in which the game is played, not the people involved in it. The NFL will have its version of Moneyball, the question is not if, but instead, when and where. Those answers are not clear, but someday we’ll know and hopefully, when that happens, it pushes the game forward in an interesting and innovative way that makes the game better for us, the fans. Also, hopefully we get a full feature movie so the best football movie in the last 10 years is no longer Draft Day, please and thank you.

.
Where is the NFL’s Version of Moneyball? Inside the Pylon 2018 - ( New Window )

the more players, the less impact it will have. It will always have a place, and it will evolve and get better, but it works in baseball due to lack of 1 v 1 matchups, less positions, and overall infinitely fewer situations to account for.

Analytics largest role in football is with scouting, IMO.

IMO.

I’ve been banging on.


By the time you get enough data the league has already evolved again making the prior data gathered useless. Because of the amount of variables the amount of data that needs to be gathered is astronomical. So unless the game stays static (it won’t) you are always going to need the film to go hand in hand. There’s clearly a spot for it (big fan of using it for in game decision making with coaches discretion) but to use it for pure scouting purposes is tough.

...have looked at the odds for years.
Even at the HS level, we have been asking ourselves, "why punt?" Why k/o deep?

Go for it, every k/o should be onsides...why not?

I have always just thought of this as scouting on an individual basis each week or rather, macro scouting in the offseason to come up with a philosophy (rather than a game plan).

For me, at the end of the day, you're still trying to find out what you do well, what your opponent struggles to do well and marry the two things in your game plan.

I believe that some of the talk of analytics is getting guys away from traditional modes of doing things, to be sure. But I also believe that some of it is simple semantics, A new name for what we've always called breaking down cut ups and scouting. Using the information in season and off season to evaluate trends both within your organization and league wide alike.

In comment 15273667 Zeke's Alibi said:

Quote:

I’ve been banging on. By the time you get enough data the league has already evolved again making the prior data gathered useless. Because of the amount of variables the amount of data that needs to be gathered is astronomical. So unless the game stays static (it won’t) you are always going to need the film to go hand in hand. There’s clearly a spot for it (big fan of using it for in game decision making with coaches discretion) but to use it for pure scouting purposes is tough.

I found the first two on the list to be the biggest reasons I don't think analytics can ever fully take over the NFL: Measuring individual player success being difficult, and all of the variables.

They cited single games or calls min their criticisms. This demonstrates a lack of understanding that is a strong indictment of the veracity any other claims made on the topic from that particular author.

In comment 15273675 Thegratefulhead said:

Quote:

They cited single games or calls min their criticisms. This demonstrates a lack of understanding that is a strong indictment of the veracity any other claims made on the topic from that particular author.

Counterpoint to that, and perhaps a reason why the NFL can't fit the same mold as baseball and basketball as cited by the Inside the Pylon number... There are significantly less games played in a season in the NFL vs. the other two:

17 games (now) vs. 162 and 82

So single games have a MUCH bigger impact statistically than the other two.

.

Next question...

would be if they eliminated tackling/contact. And I'm not saying that to be funny, I'm saying that would eliminate a lot of variables that make it difficult.

In comment 15273663 UConn4523 said:

Quote:

the more players, the less impact it will have. It will always have a place, and it will evolve and get better, but it works in baseball due to lack of 1 v 1 matchups, less positions, and overall infinitely fewer situations to account for. Analytics largest role in football is with scouting, IMO.

I think it's biggest role is in play selection. Fans love latching onto plays that backfire because the "analytics" told the coach to do XYZ, but coaches trusting their "guts" likely failed even more often. Plenty of examples in baseball where a manager left the pitcher in too long because his gut told him to. Or PH with a lefty to get a "favorable" matchup.

Getting back to the NFL, coaches are far too conservative and their use of TOs at the ends of halves/games generally sucks. Analytics is already helping here. And that doesn't mean always going for it on 4th or never kicking PATs. But analytics gives coaches concrete information that they can use when making these choices. One of the most encouraging things about Judge is he understood this and (IMO) did a great job incorporating analytics into his in game decision making. Especially compared to Shurmur who seemed to want to use analytics, but often seemed confused about how to do so.

Full contact vs. minimal contact.

In the NFL, you have 11 people making physical contact with 11 other people every play, often violently. This leads to unpredictable outcomes that are harder to measure, I believe.

...simply make it 7v7. (Basketball on grass).

Fun to watch but absolutely not "football."

In comment 15273694 Brown_Hornet said:

Quote:

...simply make it 7v7. (Basketball on grass). Fun to watch but absolutely not "football."

That's what the arena league is for.

and it is a valuable tool in making good decisions

but the beauty of football is that there is a raw human element to it. If you lose that human element, team work and individual matchups -- and create a robotic composite -- you may as well just stick with Madden if that's what you are into

I think have a place in the game for specific decisions such as:

- going for it on 4th down based on distance and location on the field vs punting

- going for it on 4th down based on distance and location on the field vs kicking a FG

- kicking an XP vs going for 2

- clock management (teams are so bad at this - I think analytics can help)

- even some game planning - analytics won't execute, but they can tell you the plays that have the most success vs a team or defensive scheme (for example)

but I don't think analytics to evaluate players will ever be THAT meaningful because it's too hard IMO to extract the individual from the team to isolate their performance in a vacuum.

You have WAR in baseball and I do believe WAR is a good indicator of individual performance, I don't think in team sports like hockey or football you can truly put a "win share" or "goal share" on one person because there are too many variables, too many moving parts, and too many unknowns to the evaluator - it's why I'm not a huge fan of PFF ratings.

but I do think analytics could "work" in the NFL - in many specific areas.

That's it. It's like this boogeyman to some people despite the fact that it has always been around to some extent. A coach citing rushing attempts as a factor in winning games is actually using (misusing) analytics. He's using data to support an action he believes helps him win football games. The biggest change, beyond the amount of data, is that we're a lot better at distilling and processing that data than we used to be. Going back to that coaching example, anybody with even a basic understanding of how to process stats can now tell you why that is a misinterpretation of that data.

At this point, it largely depends on the application. In terms of assigning positional value, draft picks value, maneuvering the cap, it's a lot more useful. You're basically using economics, which managing a salary cap is. So the economic decisions of the FO should be driven by analytics, obviously, they shouldn't make every single decision based on the data but it should be the engine. There are egos, team chemistry involved in personnel decisions as well. It's why in a vacuum extending Saquan is an awful decision but if he's truley a team leader, among the best at what he does and you have an ascending football team...there's some collateral damage in letting a guy like that walk. In-game coaching decisions should be driven by the data as well but there is certainly more grey area there. "Coaching discretion" thing is a crutch used by coaches to really only use analytics when it supports their gut. It's usually fear driven when they shy away as well, which is an emotion based decision. I think the backbone of in-game decisions should be supported by analytics with exceptions here and there...it's largely the opposite right now. Also, that data should be looking at the large macro view and what the coach's specific team does well. If you're an awful short yardage team, you should be less inclined to go for it in a short yardage situation even if the data gives a slight edge to going for it. In terms of scouting and valuing specific players based on talent? This is where you need to focus on the tape. I think it's largely useless scouting QBs, so much of that positional is either cerebral or intangible. With the other positions, it's probably useful in establishing disqualifying baselines (ie. WRs running slower than 4.7) than using it to drive your decisions. There's a lot of discretion in terms of how each player fits in a system as well that the numbers really can't tell you.

In comment 15273700 pjcas18 said:

Quote:

I think have a place in the game for specific decisions such as: - going for it on 4th down based on distance and location on the field vs punting - going for it on 4th down based on distance and location on the field vs kicking a FG - kicking an XP vs going for 2 - clock management (teams are so bad at this - I think analytics can help) - even some game planning - analytics won't execute, but they can tell you the plays that have the most success vs a team or defensive scheme (for example) but I don't think analytics to evaluate players will ever be THAT meaningful because it's too hard IMO to extract the individual from the team to isolate their performance in a vacuum. You have WAR in baseball and I do believe WAR is a good indicator of individual performance, I don't think in team sports like hockey or football you can truly put a "win share" or "goal share" on one person because there are too many variables, too many moving parts, and too many unknowns to the evaluator - it's why I'm not a huge fan of PFF ratings. but I do think analytics could "work" in the NFL - in many specific areas.

You highlighting the term "work" made me realize I think my subject misrepresented my question.

Because they already "work".

I guess what I'm asking is will they take over as the primary evaluation tool in all facets the way they have in baseball, and now seemingly basketball.

The question should have been that instead of working in synergy as they do now, actually overtaking scouting and film study as the primary indicator of success. If so, what does that look like and how far is too far?

"analytics" is the stupidest most misused phrase in all sports. It's just data and data has always been used and will always be used. With better technology there is better data. It's the case in all sports just with different technologies/data. And better data almost never replaces the need for expert analysis/integration/understanding of that data.

when computing advanced I don't think anybody wondered if weather forecasters should use the new data/technology or stick with almanacs and doing calculations by hand. or whether expertise was needed to model that data effectively.

No data is perfect but generally speaking technology advancements leading to better data is the way of the modern world.

In comment 15273682 Britt in VA said:

Quote:

would be if they eliminated tackling/contact. And I'm not saying that to be funny, I'm saying that would eliminate a lot of variables that make it difficult.

What does 'taking over' mean?

Analytics are about finding inefficiencies in a sport or even within a game and taking advantage of them. Take the example from one of the articles - the Packers success on 2nd and short by passing. If analytics show that defenses tend to stack the box in these situations expecting a run, then it makes sense for the offense to counter via deep passes.

And if enough teams start taking deep shots on 2nd and short plays, then Ds will counter and Os will be better off with a different strategy.

At its core, analytics is a more advanced and accurate study of teams' tendencies. They often look at the same things that traditional scouts do, they just put hard numbers and accurate probabilities around these things.

Scout - Player A is great at getting separation on underneath routes
Analytics - Player A gets 1.5 yards of separation on routes downfield, but 3 yards of separation on option routes in the middle of the field

Scout - Team A likes to run it on first down
Analytics - Team A runs in 63.3% of the time on first down and play action on 22.5%.

Scout - McAdoo loves 11 personnel without motion
Analytics - McAdoo uses 11 personnel on 91% of plays and 93% of this time has zero motion.

but I used it because that's how it was used in the articles I posted discussing it.

So if there is a catch all term, that would be it. Easily identifiable subject.

In comment 15273697 Britt in VA said:

Quote:

In comment 15273694 Brown_Hornet said: Quote: ...simply make it 7v7. (Basketball on grass). Fun to watch but absolutely not "football." That's what the arena league is for.

Almost, just, as you said, remove the contact.

and it really misses the mark on what/who it is supposed to serve.

So when I everything from sports and everything else get wrapped up in data driven assessments I get skeptical.

The nature of the game makes it easier to isolate individual value (Pitcher vs. Batter matchup) than other sports. It still doesn't tell the complete story but analytics in baseball are just far more meaningful than other sports.

I don't see the NFL as all that different from the NBA or NHL. All three sports have been heavily influenced by analytics in terms of style over the last decade. It's already happening. It won't ever be like the MLB due to all the variables happening on every single play over the course of a game, but the influence of analytics will continue to grow.

NFL NextGen stats is kind of the surface for what I think will be the future of the NFL stats (and other sports like NBA/NHL/MLB). The combination of video scouting with detailed data. In the future we will know that Antonio Brown gets open so often because he is able to bend his ankle X degrees on cuts and go Y MPH in the first 5 yards after making a cut. Football has always been a game of angles, data that can help understand those angles will be huge for coaches/players.

In comment 15273714 Osi Osi Osi OyOyOy said:

Quote:

NFL NextGen stats is kind of the surface for what I think will be the future of the NFL stats (and other sports like NBA/NHL/MLB). The combination of video scouting with detailed data. In the future we will know that Antonio Brown gets open so often because he is able to bend his ankle X degrees on cuts and go Y MPH in the first 5 yards after making a cut. Football has always been a game of angles, data that can help understand those angles will be huge for coaches/players.

This is probably the most fascinating and useful way it could be applied. Or as I read in another article, how to tired a Defensive Back out with certain routes. That is useful and interesting.

In comment 15273713 Britt in VA said:

Quote:

and it really misses the mark on what/who it is supposed to serve. So when I everything from sports and everything else get wrapped up in data driven assessments I get skeptical.

data is only ever a tool - and tools are usually only as good as those using them.

imo "analytics" come down to the same fundamental things that traditional scouting and coaching decisions came from - the quality of the scout/coach. There may be some teams who gain a tech/resource edge but generally those are short lived. spygate was basically "illegal analytics".

means different things to different people. For example, I don't think knowing the 4th and 1 conversion rate inside the 20 is very analytical. Its just a basic equation of successful conversions divided by number of instances. That's more "common sense"/logic than actual analytics.

Analytics for me is far more complex, where machine learning is incorporated to model a vast array of data and scenarios. This is where it gets interesting, but very tricky for football because there's so many wildcard factors that can't be accounted for.

In comment 15273717 Eric on Li said:

Quote:

In comment 15273713 Britt in VA said: Quote: and it really misses the mark on what/who it is supposed to serve. So when I everything from sports and everything else get wrapped up in data driven assessments I get skeptical. data is only ever a tool - and tools are usually only as good as those using them. imo "analytics" come down to the same fundamental things that traditional scouting and coaching decisions came from - the quality of the scout/coach. There may be some teams who gain a tech/resource edge but generally those are short lived. spygate was basically "illegal analytics".

This tracks. And that's a good example on Spygate. Thanks.

I think analytics applied to football makes it better where it has nearly ruined baseball in my opinion. Going for it more on 4th down and trying for a 2 point conversion adds to excitement, the shift and launch angle swings that result in more strikeouts make baseball boring.

In comment 15273725 Thegratefulhead said:

Quote:

I think analytics applied to football makes it better where it has nearly ruined baseball in my opinion. Going for it more on 4th down and trying for a 2 point conversion adds to excitement, the shift and launch angle swings that result in more strikeouts make baseball boring.

So to that point, what would be going to far in regards to using this in football. You just described something specific in baseball that has gotten worse for you because of this, what is that point in football?

.

In comment 15273718 UConn4523 said:

Quote:

means different things to different people. For example, I don't think knowing the 4th and 1 conversion rate inside the 20 is very analytical. Its just a basic equation of successful conversions divided by number of instances. That's more "common sense"/logic than actual analytics. Analytics for me is far more complex, where machine learning is incorporated to model a vast array of data and scenarios. This is where it gets interesting, but very tricky for football because there's so many wildcard factors that can't be accounted for.

This is spot on. I'll add another comment - even to ones fighting to be "experts" in the field misuse the term.

That's my main gripe about PFF or guys like Warren Sharp. They use the term analytics as a buzzword - as a possible panecea, then they go on and draw conclusions from datasets that aren't analytical, but rather subjective.

The PFF rating system is presented to fans as a measure of analytics when in reality, it is a bunch of guesstimations formulated into a scaleable figure. That's not data - it is subjective opinions put into a form that is presented as hard data.

as somebody that doesn't follow baseball.

From the the rise of Moneyball to now (perhaps the fall), what are the lessons to learn and pittfalls to avoid as these metrics and their uses become more prevalent? Are the Eagles an example of this, from their Superbowl victory to now?

In comment 15273714 Osi Osi Osi OyOyOy said:

Quote:

The nature of the game makes it easier to isolate individual value (Pitcher vs. Batter matchup) than other sports. It still doesn't tell the complete story but analytics in baseball are just far more meaningful than other sports.

This is exactly right. None of the other Big 4 sports have a one-on-one competition as the bedrock of game play. It just naturally lends itself to more detailed statistical analysis. There are too many variables in the other sports.

In comment 15273726 Britt in VA said:

Quote:

In comment 15273725 Thegratefulhead said: Quote: I think analytics applied to football makes it better where it has nearly ruined baseball in my opinion. Going for it more on 4th down and trying for a 2 point conversion adds to excitement, the shift and launch angle swings that result in more strikeouts make baseball boring. So to that point, what would be going to far in regards to using this in football. You just described something specific in baseball that has gotten worse for you because of this, what is that point in football?

I don't know.

If analytics applied to football made it boring?

A hypothetical:

If analytics applied to football reduced the average game to a combined 16 points. Something like that. If it devalues a traditionally important position, I don't care.

In comment 15273731 Thegratefulhead said:

Quote:

In comment 15273726 Britt in VA said: Quote: In comment 15273725 Thegratefulhead said: Quote: I think analytics applied to football makes it better where it has nearly ruined baseball in my opinion. Going for it more on 4th down and trying for a 2 point conversion adds to excitement, the shift and launch angle swings that result in more strikeouts make baseball boring. So to that point, what would be going to far in regards to using this in football. You just described something specific in baseball that has gotten worse for you because of this, what is that point in football? I don't know. If analytics applied to football made it boring? A hypothetical: If analytics applied to football reduced the average game to a combined 16 points. Something like that. If it devalues a traditionally important position, I don't care.

Or perhaps the opposite, and the increased the average game score to a combined 100+ points and teams scored on nearly every drive making it routine...?

In comment 15273733 Britt in VA said:

Quote:

In comment 15273731 Thegratefulhead said: Quote: In comment 15273726 Britt in VA said: Quote: In comment 15273725 Thegratefulhead said: Quote: I think analytics applied to football makes it better where it has nearly ruined baseball in my opinion. Going for it more on 4th down and trying for a 2 point conversion adds to excitement, the shift and launch angle swings that result in more strikeouts make baseball boring. So to that point, what would be going to far in regards to using this in football. You just described something specific in baseball that has gotten worse for you because of this, what is that point in football? I don't know. If analytics applied to football made it boring? A hypothetical: If analytics applied to football reduced the average game to a combined 16 points. Something like that. If it devalues a traditionally important position, I don't care. Or perhaps the opposite, and the increased the average game score to a combined 100+ points and teams scored on nearly every drive making it routine...?

I don't know about that. One of my favorite games of all time was Marino vs O'Brien. I generally enjoy big plays and games that feature statements and answers.

In comment 15273734 Thegratefulhead said:

Quote:

In comment 15273733 Britt in VA said: Quote: In comment 15273731 Thegratefulhead said: Quote: In comment 15273726 Britt in VA said: Quote: In comment 15273725 Thegratefulhead said: Quote: I think analytics applied to football makes it better where it has nearly ruined baseball in my opinion. Going for it more on 4th down and trying for a 2 point conversion adds to excitement, the shift and launch angle swings that result in more strikeouts make baseball boring. So to that point, what would be going to far in regards to using this in football. You just described something specific in baseball that has gotten worse for you because of this, what is that point in football? I don't know. If analytics applied to football made it boring? A hypothetical: If analytics applied to football reduced the average game to a combined 16 points. Something like that. If it devalues a traditionally important position, I don't care. Or perhaps the opposite, and the increased the average game score to a combined 100+ points and teams scored on nearly every drive making it routine...? I don't know about that. One of my favorite games of all time was Marino vs O'Brien. I generally enjoy big plays and games that feature statements and answers.

Isn't the uniqueness of a performance like that what makes it exciting though? You remember it because it was unique. What if every game, every week somewhat resembled that performance?

for a very long time. Really enjoyed the attempts to make the game more efficient. But the unintended consequence, IMV, is the game is bordering on being unwatchable due to that. I don't like the shifts, the one dimensional aspect of hitting, the approach to pitching, etc. The romance of the game is lost on me...

And now I've lost more and more interest in the NBA, as it has embraced analytics, with the emphasis on three point shooting and spreading the floor for more 1x1.

In the NFL, analytics are working right now in game preparation and in-game decision making. As many above have noted, 4th down decisions, 2 point decisions, offensive formations, etc are all derivative of analytics. These are decisions based on large data samples - to better understand probability and outcomes - and further exploiting the rules to gain advantages. Like, for example, the discussions we have on the impacts of the passing rules and how that dictates QB play and offense.

Analytics have also impacted conditioning and nutrition as more teams understand health and how to optimize performance. And behavioral and psychological health data has and is being studied to determine preferred personalities traits for team building.

So it's in the bloodstream right now of the NFL and will continue to grow. How it impacts the quality of the game is going to be the real test for me...

In comment 15273736 bw in dc said:

Quote:

for a very long time. Really enjoyed the attempts to make the game more efficient. But the unintended consequence, IMV, is the game is bordering on being unwatchable due to that. I don't like the shifts, the one dimensional aspect of hitting, the approach to pitching, etc. The romance of the game is lost on me... And now I've lost more and more interest in the NBA, as it has embraced analytics, with the emphasis on three point shooting and spreading the floor for more 1x1. In the NFL, analytics are working right now in game preparation and in-game decision making. As many above have noted, 4th down decisions, 2 point decisions, offensive formations, etc are all derivative of analytics. These are decisions based on large data samples - to better understand probability and outcomes - and further exploiting the rules to gain advantages. Like, for example, the discussions we have on the impacts of the passing rules and how that dictates QB play and offense. Analytics have also impacted conditioning and nutrition as more teams understand health and how to optimize performance. And behavioral and psychological health data has and is being studied to determine preferred personalities traits for team building. So it's in the bloodstream right now of the NFL and will continue to grow. How it impacts the quality of the game is going to be the real test for me...

Very well said.

That was a great answer.

In comment 15273735 Britt in VA said:

Quote:

In comment 15273734 Thegratefulhead said: Quote: In comment 15273733 Britt in VA said: Quote: In comment 15273731 Thegratefulhead said: Quote: In comment 15273726 Britt in VA said: Quote: In comment 15273725 Thegratefulhead said: Quote: I think analytics applied to football makes it better where it has nearly ruined baseball in my opinion. Going for it more on 4th down and trying for a 2 point conversion adds to excitement, the shift and launch angle swings that result in more strikeouts make baseball boring. So to that point, what would be going to far in regards to using this in football. You just described something specific in baseball that has gotten worse for you because of this, what is that point in football? I don't know. If analytics applied to football made it boring? A hypothetical: If analytics applied to football reduced the average game to a combined 16 points. Something like that. If it devalues a traditionally important position, I don't care. Or perhaps the opposite, and the increased the average game score to a combined 100+ points and teams scored on nearly every drive making it routine...? I don't know about that. One of my favorite games of all time was Marino vs O'Brien. I generally enjoy big plays and games that feature statements and answers. Isn't the uniqueness of a performance like that what makes it exciting though? You remember it because it was unique. What if every game, every week somewhat resembled that performance?

Maybe, in general I like more scoring and big plays but readily concede too much could be a problem. The problem is that I don't know what too much is, until I feel it.

In comment 15273736 bw in dc said:

Quote:

for a very long time. Really enjoyed the attempts to make the game more efficient. But the unintended consequence, IMV, is the game is bordering on being unwatchable due to that. I don't like the shifts, the one dimensional aspect of hitting, the approach to pitching, etc. The romance of the game is lost on me... And now I've lost more and more interest in the NBA, as it has embraced analytics, with the emphasis on three point shooting and spreading the floor for more 1x1. In the NFL, analytics are working right now in game preparation and in-game decision making. As many above have noted, 4th down decisions, 2 point decisions, offensive formations, etc are all derivative of analytics. These are decisions based on large data samples - to better understand probability and outcomes - and further exploiting the rules to gain advantages. Like, for example, the discussions we have on the impacts of the passing rules and how that dictates QB play and offense. Analytics have also impacted conditioning and nutrition as more teams understand health and how to optimize performance. And behavioral and psychological health data has and is being studied to determine preferred personalities traits for team building. So it's in the bloodstream right now of the NFL and will continue to grow. How it impacts the quality of the game is going to be the real test for me...

FINALLY! Your sanity has returned..Great post..Actual mic drop, imv

In comment 15273741 Thegratefulhead said:

Quote:

In comment 15273735 Britt in VA said: Quote: In comment 15273734 Thegratefulhead said: Quote: In comment 15273733 Britt in VA said: Quote: In comment 15273731 Thegratefulhead said: Quote: In comment 15273726 Britt in VA said: Quote: In comment 15273725 Thegratefulhead said: Quote: I think analytics applied to football makes it better where it has nearly ruined baseball in my opinion. Going for it more on 4th down and trying for a 2 point conversion adds to excitement, the shift and launch angle swings that result in more strikeouts make baseball boring. So to that point, what would be going to far in regards to using this in football. You just described something specific in baseball that has gotten worse for you because of this, what is that point in football? I don't know. If analytics applied to football made it boring? A hypothetical: If analytics applied to football reduced the average game to a combined 16 points. Something like that. If it devalues a traditionally important position, I don't care. Or perhaps the opposite, and the increased the average game score to a combined 100+ points and teams scored on nearly every drive making it routine...? I don't know about that. One of my favorite games of all time was Marino vs O'Brien. I generally enjoy big plays and games that feature statements and answers. Isn't the uniqueness of a performance like that what makes it exciting though? You remember it because it was unique. What if every game, every week somewhat resembled that performance? Maybe, in general I like more scoring and big plays but readily concede too much could be a problem. The problem is that I don't know what too much is, until I feel it.

I guess my concern is... By the time you feel and it turns you off, it will be too late.

Not that there's any closing of pandora's box, or putting the toothpaste back in the tube at this point, but to bw's point. I do not like watching Arena Football. I hope it does not go that far.

I also embraced analytics in baseball but I barely watch games now. When it was stacking players that got on base and scoring increased, I was good. When it took the nuances of the game away, no good.

Maybe that is the answer.

If analytics were to fundamentally change the way the game is played and the result was less excitement, I would be against it. If it changes the game and adds to the excitement, I would be for it.

I have enjoyed reading the thread.

In comment 15273706 AcesUp said:

Quote:

That's it. It's like this boogeyman to some people despite the fact that it has always been around to some extent. A coach citing rushing attempts as a factor in winning games is actually using (misusing) analytics. He's using data to support an action he believes helps him win football games. The biggest change, beyond the amount of data, is that we're a lot better at distilling and processing that data than we used to be. Going back to that coaching example, anybody with even a basic understanding of how to process stats can now tell you why that is a misinterpretation of that data. At this point, it largely depends on the application. In terms of assigning positional value, draft picks value, maneuvering the cap, it's a lot more useful. You're basically using economics, which managing a salary cap is. So the economic decisions of the FO should be driven by analytics, obviously, they shouldn't make every single decision based on the data but it should be the engine. There are egos, team chemistry involved in personnel decisions as well. It's why in a vacuum extending Saquan is an awful decision but if he's truley a team leader, among the best at what he does and you have an ascending football team...there's some collateral damage in letting a guy like that walk. In-game coaching decisions should be driven by the data as well but there is certainly more grey area there. "Coaching discretion" thing is a crutch used by coaches to really only use analytics when it supports their gut. It's usually fear driven when they shy away as well, which is an emotion based decision. I think the backbone of in-game decisions should be supported by analytics with exceptions here and there...it's largely the opposite right now. Also, that data should be looking at the large macro view and what the coach's specific team does well. If you're an awful short yardage team, you should be less inclined to go for it in a short yardage situation even if the data gives a slight edge to going for it. In terms of scouting and valuing specific players based on talent? This is where you need to focus on the tape. I think it's largely useless scouting QBs, so much of that positional is either cerebral or intangible. With the other positions, it's probably useful in establishing disqualifying baselines (ie. WRs running slower than 4.7) than using it to drive your decisions. There's a lot of discretion in terms of how each player fits in a system as well that the numbers really can't tell you.

I missed this post earlier, but it does highlight the delicate balance between the human element and the data driven element. Thanks.

I think it will continue to evolve. One thing I would worry about is too much reliance in game situations. Coaches have had moments of over thinking (analysis) which can spill over to players.

Seahawks vs. Patriots Super Bowl. Hawks at the goal line and good ole Ernie Adams alerted Bill to a play the Hawks had used in that situation. Interception and game over. SB champs. Here you relied on a man considered a genius. So in special situations I think it has more value. Organized data accessible at a moments notice certainly should be implemented but not play by play. Punts, penalties, going for it or not, etc. certainly a stronger place for it.

In comment 15273752 Lines of Scrimmage said:

Quote:

I think it will continue to evolve. One thing I would worry about is too much reliance in game situations. Coaches have had moments of over thinking (analysis) which can spill over to players. Seahawks vs. Patriots Super Bowl. Hawks at the goal line and good ole Ernie Adams alerted Bill to a play the Hawks had used in that situation. Interception and game over. SB champs. Here you relied on a man considered a genius. So in special situations I think it has more value. Organized data accessible at a moments notice certainly should be implemented but not play by play. Punts, penalties, going for it or not, etc. certainly a stronger place for it.

I was literally just thinking about this, but in a different way than your example.

For instance, instead of the biggest spot on the biggest stage, what about this scenario:

You are in the second quarter of Week 3 in against a division rival at home down 7-3. You have 4th and 2 from your own 45 yard line. The data says to go for it, but despite the better odds you fail. The other team has great field position and goes on to score a TD to go up 14-3 and they never look back.

That changes the complexion of a game, big time. What if instead, you coffin corner punt it and pin them inside their own five yard line? New game, new story.

To take that one step further, it's now Week 17. You miss the playoffs by one game.

I know you can trace a million lines back to some point in a season, but if you just trusted your gut and played it conservatively there, maybe things are different.

It has made all the games less interesting to watch. I find I am watching less and less sports due to the over-reliance on all this statistical shit. And I do mean shit.

have a way of correcting themselves due to the sheer amount of games for them to even out.

In the NFL, EVERY game is critical. Mistakes are magnified big time because every game counts.

In comment 15273665 US1 Giants said:

Quote:

IMO.

I hear you. To an uncanny degree, baseball as it is played today has been "solved." Do this thing, get that result; and to get that result, you must do this thing. You don't see competing ideas of how to win, like power vs speed, because it's now clear from analytics what works and what doesn't.

Even with some rule changes, like banning the extreme shift and limiting the size of pitching staffs, even moving the mound back (an idea worth trying IMO) that's not likely to change. Not even a re-formulation of the ball or bigger outfields would change that. Personally, I think gloves should be much smaller, but that wouldn't really change the analytics-driven approach. So is there some undiscovered revolutionary change (maybe in hitting technique) that would open things up? Maybe, but until someone discovers it, it's hard to know.

It would be interesting to use analytics to try to discover what could change the game in positive ways -- more action, shorter time-of-game, more balls in play, etc.

... Analytics aren't changing the game Because they're not new.

The word analytics and the extent to which fans use analytics are what is new to the game.

The changes in the game have come from rule changes.

That said, in following along with the other conversations, I like high scoring and games too but I don't like them any more than low scoring games. In fact I prefer a mix.

I guess this goes back to The importance of points because I believe the importance of not allowing them is equal to the importance of scoring them.
The difference The difference however is that the rule changes have The difference however is that the rule changes have made it harder to defend and easier to score... But that's a different conversation.

In comment 15273773 81_Great_Dane said:

Quote:

So is there some undiscovered revolutionary change (maybe in hitting technique) that would open things up? Maybe, but until someone discovers it, it's hard to know. ...

How about hitting the ball to the opposite field? That's a hitting technique that's not all that hard, or at least comparatively to hitting generally.

You make the best decision based on the information you have. Just because it failed doesn't mean it was the wrong decision. There will always be gamesmanship. The opponent, theoretically has access to the same analytics. It is why even though passing is better play than a run, you still must run or are too predictable and easily defeated. Because of this, I don't think we have to worry about analytics ruining football the way it has baseball.

If you need analytics you shouldn't be in or around the game period. Whatever happened to buckling chin strap and kicking ass. Stop w this computer bs makes me sick and angry. Every sport has been stripped of it's manhood.

In comment 15273757 Britt in VA said:

Quote:

In comment 15273752 Lines of Scrimmage said: Quote: I think it will continue to evolve. One thing I would worry about is too much reliance in game situations. Coaches have had moments of over thinking (analysis) which can spill over to players. Seahawks vs. Patriots Super Bowl. Hawks at the goal line and good ole Ernie Adams alerted Bill to a play the Hawks had used in that situation. Interception and game over. SB champs. Here you relied on a man considered a genius. So in special situations I think it has more value. Organized data accessible at a moments notice certainly should be implemented but not play by play. Punts, penalties, going for it or not, etc. certainly a stronger place for it. I was literally just thinking about this, but in a different way than your example. For instance, instead of the biggest spot on the biggest stage, what about this scenario: You are in the second quarter of Week 3 in against a division rival at home down 7-3. You have 4th and 2 from your own 45 yard line. The data says to go for it, but despite the better odds you fail. The other team has great field position and goes on to score a TD to go up 14-3 and they never look back. That changes the complexion of a game, big time. What if instead, you coffin corner punt it and pin them inside their own five yard line? New game, new story. To take that one step further, it's now Week 17. You miss the playoffs by one game. I know you can trace a million lines back to some point in a season, but if you just trusted your gut and played it conservatively there, maybe things are different.

You're cherry picking examples and assuming that trusting the gut is "flawless". It's not. It may be more conservative, but that doesn't make it better.

What if, in your example the coach goes with his gut and the P shanks it? Or kicks a low liner that is returned for a TD (see Desean Jackson)?

In comment 15273759 Britt in VA said:

Quote:

have a way of correcting themselves due to the sheer amount of games for them to even out. In the NFL, EVERY game is critical. Mistakes are magnified big time because every game counts.

The main point of analytics is to provide more detail to minimize "mistakes". Being overly conservative can also be a mistake.

In comment 15273849 giants#1 said:

Quote:

In comment 15273757 Britt in VA said: Quote: In comment 15273752 Lines of Scrimmage said: Quote: I think it will continue to evolve. One thing I would worry about is too much reliance in game situations. Coaches have had moments of over thinking (analysis) which can spill over to players. Seahawks vs. Patriots Super Bowl. Hawks at the goal line and good ole Ernie Adams alerted Bill to a play the Hawks had used in that situation. Interception and game over. SB champs. Here you relied on a man considered a genius. So in special situations I think it has more value. Organized data accessible at a moments notice certainly should be implemented but not play by play. Punts, penalties, going for it or not, etc. certainly a stronger place for it. I was literally just thinking about this, but in a different way than your example. For instance, instead of the biggest spot on the biggest stage, what about this scenario: You are in the second quarter of Week 3 in against a division rival at home down 7-3. You have 4th and 2 from your own 45 yard line. The data says to go for it, but despite the better odds you fail. The other team has great field position and goes on to score a TD to go up 14-3 and they never look back. That changes the complexion of a game, big time. What if instead, you coffin corner punt it and pin them inside their own five yard line? New game, new story. To take that one step further, it's now Week 17. You miss the playoffs by one game. I know you can trace a million lines back to some point in a season, but if you just trusted your gut and played it conservatively there, maybe things are different. You're cherry picking examples and assuming that trusting the gut is "flawless". It's not. It may be more conservative, but that doesn't make it better. What if, in your example the coach goes with his gut and the P shanks it? Or kicks a low liner that is returned for a TD (see Desean Jackson)?

I'm not talking about a gut feeling, I'm talking about conventional wisdom that you shouldn't go for it on 4th down in your own territory that early in the game. And there is a reason it is conventional wisdom, because it's been done for a long time. Now I'm not saying that said conventional wisdom is so rigid that it shouldn't change, but....

There must be some balance between playing the odds and conventional wisdom.

In comment 15273850 giants#1 said:

Quote:

In comment 15273759 Britt in VA said: Quote: have a way of correcting themselves due to the sheer amount of games for them to even out. In the NFL, EVERY game is critical. Mistakes are magnified big time because every game counts. The main point of analytics is to provide more detail to minimize "mistakes". Being overly conservative can also be a mistake.

Conversely, going for it on 4th down in your own territory early in a game can backfire, as well. So that could equally be considered a mistake.

doesn't mean its correct or that it gives your team the best chance to win. Conventional wisdom in baseball thought bunts were valuable. Yet in virtually every situation, bunts are a bad idea with one of the few times it makes sense is the artificial extra innings crap you now have.

.

It's already started, and it's only going to progress.

One thing I don't understand is the idea video isn't used anymore in baseball, it absolutely is. In fact all defensive saberstats are done after the fact with video (although statcast may allow closer to real-time analysis there). Batters and pitchers still watch video too, of themselves and opponents.

For in game decisions, baseball will always be different become of the 1v1 nature of hitter versus pitcher. Where you're going to see the big change in football is in analytics driving player acquisition.

For example in MLB spin rate is a big thing for pitchers, so we could see a similar stat become a big deal for QBs

with analytics to see how things might progress in football. As an example, here's an article about Liverpool.
Liverpool analytics - ( New Window )

In comment 15273934 Scooter185 said:

Quote:

For example in MLB spin rate is a big thing for pitchers, so we could see a similar stat become a big deal for QBs

My son is a D1 baseball prospect out of high school and the data collected at prospect events (like Perfect Game) are pretty amazing. He's a position player only, so they look at hitting categories like launch angle, maximum barrel speed, exit velo, impact momentum (barrel speed + weight of the bat), max acceleration (how fast can you reach your top bat speed). They then comp those to your graduating class, other D1 prospects, etc.

Ironically, however, the final evaluation grade you are given is based mostly on the subjective view of the scouts on hand... ;)

"Like Other Sports" is probably the best question.

The insights I have from the data science and predictive analytics world/community when I ask about football:

- All decisions are made on some sort of data, on a spectrum from past experience to sophisticated modeling. It's wise to use the best data at your disposal always.

- The notion football is too complex to model gets a hardy laugh, much more complex outcomes are modeled accurately.

- The immediate-term advances in football will be about positional value and resource allocation.

- The medium-term advances will be about tracking scenarios that work. The hardiest laughs are reserved for the notion "You don't know the play call or responsibilities."

- If you know the scenarios that ultimately work, you build your plays based on that.

- Football doesn't lend itself to grading individual players or actions like baseball. Grades are stupid.

football is a different sport than the others

In comment 15274020 christian said:

Quote:

"Like Other Sports" is probably the best question. The insights I have from the data science and predictive analytics world/community when I ask about football: - All decisions are made on some sort of data, on a spectrum from past experience to sophisticated modeling. It's wise to use the best data at your disposal always. - The notion football is too complex to model gets a hardy laugh, much more complex outcomes are modeled accurately. - The immediate-term advances in football will be about positional value and resource allocation. - The medium-term advances will be about tracking scenarios that work. The hardiest laughs are reserved for the notion "You don't know the play call or responsibilities." - If you know the scenarios that ultimately work, you build your plays based on that. - Football doesn't lend itself to grading individual players or actions like baseball. Grades are stupid.

Two questions about this.

The first is: Who is laughing? The guys doing the modeling? I wouldn't think they would say it can't be done, right? But what are we talking about here? Are we talking about choosing players, like you mentioned? Are we talking outcomes to plays? Or are we talking about predicting which way the ball bounces in the air either after a hit or drop, or somebody missing a block causing the play to go awry?

The second question, is in regards to actually building your play around what works. Is that an outcome we want? With all 32 teams building plays around the same data? Because that is a concern of people that have been fans of other sports and have now been turned off by this.

In comment 15274063 Britt in VA said:

Quote:

Two questions about this. The first is: Who is laughing? The guys doing the modeling? I wouldn't think they would say it can't be done, right? But what are we talking about here? Are we talking about choosing players, like you mentioned? Are we talking outcomes to plays? Or are we talking about predicting which way the ball bounces in the air either after a hit or drop, or somebody missing a block causing the play to go awry? The second question, is in regards to actually building your play around what works. Is that an outcome we want? With all 32 teams building plays around the same data? Because that is a concern of people that have been fans of other sports and have now been turned off by this.

RE: what can modeled -- all of it. The usefulness is on a spectrum. It's quite easy to predict the way a ball will bounce. In practice, not so useful because the human mind and body can't do much fast enough to do anything as a result.

Propensity of making mistakes. Absolutely, especially the circumstances most commonly preceding making the mistake. Coaches are already doing this, or at least trying. Coaching is nothing more than taking data (memories) and trying apply it to create the best future outcome.

With play design and play calls. Again, coaches are already doing, or at least trying. The first decision a coach makes when designing or calling a play is "will this work."

Narrowing the confidence interval is the game. This is where creativity and enguiniety come in -- if you don't want to boxed in by the data, break the mold.

But the answer isn't ignore the data - because you can be certain some of your competitors won't.

In comment 15273706 AcesUp said:

Quote:

That's it. It's like this boogeyman to some people despite the fact that it has always been around to some extent. A coach citing rushing attempts as a factor in winning games is actually using (misusing) analytics. He's using data to support an action he believes helps him win football games. The biggest change, beyond the amount of data, is that we're a lot better at distilling and processing that data than we used to be. Going back to that coaching example, anybody with even a basic understanding of how to process stats can now tell you why that is a misinterpretation of that data.

I think this is a very good post, and points out very simply how "analytics" have always been included in football decision-making, but the ability to process large pools of data is what is improving, and the willingness to embrace actual probability is increasing among coaches and front offices. That latter point may very well be the result of some data-driven trends in other sports becoming more widely accepted, to the extent that coaches might not be second-guessed quite as much now as they were previously, and that certainly has influenced some coaching decisions in the past.

I think we should also be aware that this thread may very well have been started with a purpose that is not just simple curiosity.

I have always been aproponent that analytics shouls help push a team in a direction as far as the style of football is concerned. And that is the number 1 item on that the author of one of Britt's articles mentions. THe determination of style, should then be a driving factor in your team build process.

We have far greater tools and capacity for analyzing data than ever before. You would be surprised at just how well some of these tools work. And they work in ways that are difficult to understand or explain. They find features that we as humans would not expect to be impactful, but are actually very impactful.

As somebody who has built portfolio performance and risk analytics engines, the parallels between team building within the salary cap, viewing the athletes as assets with concurrent liabilities such as salary and injury risk, as a process that has a lot of similarities to build close ended portfolios with financial securities as assets with concurrent liabilities such as cost, and the inherent risk of the security. Both are driven by market dynamics and macro trends, football has free agency and salary cap inflation/deflation respectively. And if you want to talk about variable, portfolio management has many many many more variables than football, and yet there are many successful models that drive portfolio manager investing. If it can be done for portfolio management, then it can be done for football.

.McL — I completely agree. The asset management part is not that complex.

...” y’all are funny.

In comment 15274239 Brown_Hornet said:

Quote:

...” y’all are funny.

It's a technical term, I son' expect you to understand.

Belichek made perhaps the most anti-"anlytics" decision ever in a SB by not calling a timeout. I would almost guarantee the probabilities and by the book analytics say that calling a timeout there is 100% no questions asked decision.

But Belichek didn't do that because he saw that Pete Carroll and the Seahawks appeared to be disorganized in calling the next play. He rolled the dice and made a judgement to just stay out of the way and hope his opponent made a mistake.

Success will always require a blend of good information and good decision making (and luck).

In comment 15274244 .McL. said:

Quote:

In comment 15274239 Brown_Hornet said: Quote: ...” y’all are funny. It's a technical term, I son' expect you to understand.

I'm familiar with the term.

It's just funny that you guys think it applies here.

As I have said it applies to fans and pundits...

You're expectations of me or not my concern.

In comment 15274271 Brown_Hornet said:

Quote:

In comment 15274244 .McL. said: Quote: In comment 15274239 Brown_Hornet said: Quote: ...” y’all are funny. It's a technical term, I son' expect you to understand. I'm familiar with the term. It's just funny that you guys think it applies here. As I have said it applies to fans and pundits... You're expectations of me or not my concern.

Are you saying teams don't run statistical models? Because usually the teams are way ahead of what the public actually gets in terms of information. They also often have their own proprietary models/stats that we never see

... I think fans especially those there educated with regards to analytics overestimate what the coaching staff is actually doing with the information.

Are they using it... absolutely but they've been using it since the 1970s.

I believe that there is someone on every NFL team possibly even a group of people that compile this information and that is a relatively new function of the NFL machine.
That said I believe thesr statistics/analytics/averages/trends, whatever you decide you want to call them, are used to support what the coaching staff already sees by studying their opponents and their own teams.

I do not believe that analytics are guiding teams to model Their offensive or defensive coach and philosophies to something different than what they believe their players are best suited for.

In comment 15274271 Brown_Hornet said:

Quote:

In comment 15274244 .McL. said: Quote: In comment 15274239 Brown_Hornet said: Quote: ...” y’all are funny. It's a technical term, I son' expect you to understand. I'm familiar with the term. It's just funny that you guys think it applies here. As I have said it applies to fans and pundits... You're expectations of me or not my concern.

So you’re just going to go run the 65 Toss-Power Trap because it worked for Hank Stram back in 1970?

Ignore the rest of this new fangled computer stuff?

...

In comment 15274020 christian said:

Quote:

- The notion football is too complex to model gets a hardy laugh, much more complex outcomes are modeled accurately. - The medium-term advances will be about tracking scenarios that work. The hardiest laughs are reserved for the notion "You don't know the play call or responsibilities."

The two points above may give insight as to why there is a divide in the expectations of those who talk about analytics extensively and the people who are in the analytics departments in the NFL. If the data science community is supposedly getting the biggest chuckles on those two points, then they have an equally loud laugh coming back at them from NFL team facilities

The prevailing thought inside the game is that football IS too complex to model to be able to gain a reasonable advantage in. There are a massive amount of permutations when dealing with the individual and coordinated movements of 11 players, with strategies and schemes that change continually. Trends that change year from year. The goal hasn't been to model the game - it has been to drill down to areas that may lead to an advantage - or at least better decision-making. That's why the short-term implementation has been in areas of kicking or going for it and areas not on the field like nutrition, exercise, etc.

When this topic has been brought up - the earliest answer I've heard of when a useable model could be implemented was 25 years. I've actually heard more people answer that it never will happen vs. those that give a time frame (even if I disagree with that). For some reason, a comparison to chess gets floated out there a lot. Maybe because people like to talk about the "perfect model". The differences aren't just stark between football and chess, they are massive. There aren't a set number of spaces to move to and 11 players can move any direction as many steps as possible and still achieve a positive outcome. All 22 players can move the same distance two plays in a row with different outcomes.

The idea that the data science community thinks that not knowing the playcall or responsibilities is immaterial to analyzing the game is one dismissed in football circles, and not only has that notion been discussed by football people, both Bill B and Ernie Adams have commented about this exact point. Little Bill doesn't believe in any evaluations of his players by outside sources because they don't know the intent of the play or the responsibilities of each player. He said it is integral to not only understanding the outcome, but how to learn from it and how to better design things in the future - as well as taking into account who missed an assignment, which caused the play to fail. Not knowing that specifics of the play leads to an incomplete conclusion - and even a conclusion that ends up being completely different from the one drawn by those that knew the specifics

Even in this case of Bill vs the analytics "experts", it appears he's playing chess while they are playing tic-tac-toe.

In comment 15274280 Jimmy Googs said:

Quote:

In comment 15274271 Brown_Hornet said: Quote: In comment 15274244 .McL. said: Quote: In comment 15274239 Brown_Hornet said: Quote: ...” y’all are funny. It's a technical term, I son' expect you to understand. I'm familiar with the term. It's just funny that you guys think it applies here. As I have said it applies to fans and pundits... You're expectations of me or not my concern. So you’re just going to go run the 65 Toss-Power Trap because it worked for Hank Stram back in 1970? Ignore the rest of this new fangled computer stuff?

You have clearly read none of what I have posted regarding analytics coaching or anything else...
... That's it I love the Hank Stram reset.

When you have decided to actually read what I have said regarding the subject we can discuss the topic at hand.

for the y’all crowd. And analytics have been around for a while.

Did you have more?

In comment 15274250 Eric on Li said:

Quote:

Belichek made perhaps the most anti-"anlytics" decision ever in a SB by not calling a timeout. I would almost guarantee the probabilities and by the book analytics say that calling a timeout there is 100% no questions asked decision. But Belichek didn't do that because he saw that Pete Carroll and the Seahawks appeared to be disorganized in calling the next play. He rolled the dice and made a judgement to just stay out of the way and hope his opponent made a mistake. Success will always require a blend of good information and good decision making (and luck).

Interesting point. It seemed like a total gut feel. Because from a special I saw on that game, coaches on the sideline were asking BB several times if he wanted to use the TO. And BB just calmly said no.

But that's the art vs science of football. What do you trust more in that situation? BB's 40+ years of pro coaching? Or the analytics? Or maybe that's a poor example because BB is an exception to any rule due to his genius and experience...

The real story to that situation was the interception by Butler and the preparation New England had done for that specific goal line formation.

In comment 15274302 Jimmy Googs said:

Quote:

for the y’all crowd. And analytics have been around for a while. Did you have more?

I’m not even sure what you just said...
... do you have a problem with the word y’all?

I’ve tried with you...I’m done.

In comment 15274308 bw in dc said:

Quote:

In comment 15274250 Eric on Li said: Quote: Belichek made perhaps the most anti-"anlytics" decision ever in a SB by not calling a timeout. I would almost guarantee the probabilities and by the book analytics say that calling a timeout there is 100% no questions asked decision. But Belichek didn't do that because he saw that Pete Carroll and the Seahawks appeared to be disorganized in calling the next play. He rolled the dice and made a judgement to just stay out of the way and hope his opponent made a mistake. Success will always require a blend of good information and good decision making (and luck). Interesting point. It seemed like a total gut feel. Because from a special I saw on that game, coaches on the sideline were asking BB several times if he wanted to use the TO. And BB just calmly said no. But that's the art vs science of football. What do you trust more in that situation? BB's 40+ years of pro coaching? Or the analytics? Or maybe that's a poor example because BB is an exception to any rule due to his genius and experience... The real story to that situation was the interception by Butler and the preparation New England had done for that specific goal line formation.

Thought i heard him talk about this once live...BB expected Seattle to run and they had their goal line defense in. But when they saw Seattle adjust with a 3 WR package, New England took out a LB and added Butler right before that play.

BB was fine with his alignment for a goal line stand. So no need for a timeout that would allow Seattle to adjust again.

In comment 15274310 Brown_Hornet said:

Quote:

In comment 15274302 Jimmy Googs said: Quote: for the y’all crowd. And analytics have been around for a while. Did you have more? I’m not even sure what you just said... ... do you have a problem with the word y’all? I’ve tried with you...I’m done.

Ok, just matriculating through this thread....

Don't confuse model with magic eight ball that will predict the perfect play and print out the final score for you too.

Data modeling produces micro and macro insights that assist with everything from play selection/probabilities, health and safety, and self scouting and assessment.

The break throughs in baseball came when the industry at large let go of insisting the things they were doing led to success, and instead looked at successful outcomes and worked backward to understand why.

This will happen in football. And the aids will get better and better. It's not going to take 25 years. Anyone making technical predictions on a 25 year horizon is the one we'll all probably be laughing at.

This is an industry 10 years ago with basically no modern analytical footprint. I'll happily re-visit this topic every year and fact check this 25 year claim.

In comment 15274308 bw in dc said:

Quote:

In comment 15274250 Eric on Li said: Quote: Belichek made perhaps the most anti-"anlytics" decision ever in a SB by not calling a timeout. I would almost guarantee the probabilities and by the book analytics say that calling a timeout there is 100% no questions asked decision. But Belichek didn't do that because he saw that Pete Carroll and the Seahawks appeared to be disorganized in calling the next play. He rolled the dice and made a judgement to just stay out of the way and hope his opponent made a mistake. Success will always require a blend of good information and good decision making (and luck). Interesting point. It seemed like a total gut feel. Because from a special I saw on that game, coaches on the sideline were asking BB several times if he wanted to use the TO. And BB just calmly said no. But that's the art vs science of football. What do you trust more in that situation? BB's 40+ years of pro coaching? Or the analytics? Or maybe that's a poor example because BB is an exception to any rule due to his genius and experience... The real story to that situation was the interception by Butler and the preparation New England had done for that specific goal line formation.

if the Pats call a time out that pass likely never gets thrown. I would guess Caroll would have talked through and realized there was plenty of time on the clock and runs Lynch into another championship on the next play. Even if only to make sure they didn't throw an incomplete pass, stop the clock, and give Brady the ball with time left.

perhaps the most expert usage of data is knowing when to ignore the data.

Eric — that’s a great point.

If the probability model tells you the next card will be a king, but you can see with your damn eyes it’s an ace, believe your eyes.

Analytics can reveal trends and probabilities but the outcome of any given play or even any single game is unknowable in advance.

To take an example from earlier in the thread, punting or going for it on 4th and 2 from your own 45: Analytics can tell you that if you consistently go for it, you increase your chances of winning over the season. In the example above, where you go for it, fail, lose and then maybe as the playoffs by one game, it’s fallacious to blame the single decision to go for it for missing the playoffs. Maybe consistently going for it on 4th and short won you a couple of games. Maybe you would have been out of contention if not for that strategy.

No one ever said it works every time. Games are still games. The smaller the sample size, like one play, the more random the result. And unlikely things happen all the time. Pitchers hit home runs. It rains in the desert. But any given pitcher is unlikely to hit a lot of home runs in a season. On any given spot in the desert, most of the time it’s dry. Same with analytics. Do this a lot, shift the odds a bit in your favor. But on any one play, crazy shit can still happen.

a millions lines that can be drawn back to points in the season that could have gone differently.

But I also mentioned that there are only 16 games (now 17) in a season, so the margin for error is razor thin in missing the playoffs vs not.

I’m not sure it makes much sense to take risks like the scenario I mentioned that early in the season in hopes that if it fails it just evens out. What if you do it again next game and now you’ve lost two in a row? When do you abandon ship on the odds in situations like that? How long you willing to wait for it to “even out”?

Seemed like all his decisions backed fired on him late in the game that year. If he played it safe the other team scored at the end to beat him, if he was aggressive then the team didn’t execute and left points behind.

I think a lot of that resulted in sentiments from posters that say TC lost his fastball or the game was passing him by...

In comment 15274406 Jimmy Googs said:

Quote:

Seemed like all his decisions backed fired on him late in the game that year. If he played it safe the other team scored at the end to beat him, if he was aggressive then the team didn’t execute and left points behind. I think a lot of that resulted in sentiments from posters that say TC lost his fastball or the game was passing him by...

He was in a tough spot that year. He had the 6th ranked points scoring offense and the 31st ranked defense in points allowed.

I don't recall the number exactly, but I think the Giants lost 6 games in 2015 where they were tied and had the lead with two minutes to go.

In the infamous Beckham meltdown game, the Giants were down 21 in the 4th quarter against the undefeated Panthers. Manning threw 3 TD's in a single quarter to tie the game with 50 seconds left.

The Panthers came out after the kickoff, started from their own 20, and casually dinked and dunked their way right down to a walk off FG for the victory.

There were many like that game. Analytics can't cure a defense that bad.

not and, obviously.

to do the opposite. They were down at the goal line, 1st and goal, with under two minutes to go. Instead of just scoring. They tried to get cute and run time off the clock but ended up having to kick a FG. Once again, defense caved and they lost.

Tom coach scared that year, with good reason. He knew his defense couldn't stop anybody.

But that's all for another thread.

Back to "analytics".

In comment 15274392 Britt in VA said:

Quote:

a millions lines that can be drawn back to points in the season that could have gone differently. But I also mentioned that there are only 16 games (now 17) in a season, so the margin for error is razor thin in missing the playoffs vs not. I’m not sure it makes much sense to take risks like the scenario I mentioned that early in the season in hopes that if it fails it just evens out. What if you do it again next game and now you’ve lost two in a row? When do you abandon ship on the odds in situations like that? How long you willing to wait for it to “even out”?

1. It depends on your view of the team. If you're the Chiefs, you can and arguably should be a little more risk averse. If you're a middle of the pack team, you should likely be more aggressive in your decision making. Hence the old "conventional wisdom" that good teams don't need to resort to trick plays.


2. Not sure what specific scenario you are alluding to, but from 1998-2016, teams converted 65.7% of 4th and 1 plays. So it's riskier not going for it.


Link - ( New Window )

Imagine the not so distant future when technology and data analysis gives the defensive coordinator better, more objective insights as to why the defense sucked. I bet that makes the defense better : )

In comment 15274439 christian said:

Quote:

Imagine the not so distant future when technology and data analysis gives the defensive coordinator better, more objective insights as to why the defense sucked. I bet that makes the defense better : )

You still need the athletes.

In comment 15274435 Britt in VA said:

Quote:

But that's all for another thread. Back to "analytics".

I don't recall the specific game management decisions, but late game scenarios like those are exactly where analytics can make a difference today. Proper use of TOs and knowing when to try and burn some extra secs can have a huge impact on close games. Just compare Shurmur and Judge's use of TOs in the final 5 minutes of games to see how it matters. Or look at some of Vrabel's decisions the last few seasons (taking penalties to burn extra clock).

yet the Giants still went from 31st on defense in the NFL to 2nd.

In comment 15274442 Britt in VA said:

Quote:

In comment 15274439 christian said: Quote: Imagine the not so distant future when technology and data analysis gives the defensive coordinator better, more objective insights as to why the defense sucked. I bet that makes the defense better : ) You still need the athletes.

Of course you do. Has anyone ever, remotely, intimidated that you don’t?

In comment 15274438 Britt in VA said:

Quote:

Quote: 2. Not sure what specific scenario you are alluding to, but from 1998-2016, teams converted 65.7% of 4th and 1 plays. So it's riskier not going for it.

No one said it doesn't. Aren't the articles you posted about 'game neutral' situations though? Not 4th quarter decisions.

that thinks we are old and stuck in our ways and can't see the big picture of analytics.

It would be foolish to deny their impact and what is coming.

However, and this is not just football related... We must tread carefully on how reliant we are or aren't on them. They are a tool. They should always be a tool. Context will ALWAYS be super important.

People should not let their passion for them cloud their big picture. And people that are wary of them should accept the benefits that the tools provide.

There is a balance that must be achieved.

In comment 15274435 Britt in VA said:

Quote:

But that's all for another thread. Back to "analytics".

Well, this is relevant as to the topic. Whatever the Giants were or were not using in 2015 as linked to poor results. And basically the opposite occurred in 2016 when the performances of the offense and defense flipped.

In comment 15274450 christian said:

Quote:

In comment 15274442 Britt in VA said: Quote: In comment 15274439 christian said: Quote: Imagine the not so distant future when technology and data analysis gives the defensive coordinator better, more objective insights as to why the defense sucked. I bet that makes the defense better : ) You still need the athletes. Of course you do. Has anyone ever, remotely, intimidated that you don’t?

Well you were responding to me saying that the Giants in that example had the 31st ranked defense in points allowed, and that perhaps having more analytics at their disposal could improve that. So I think you kind of did intimate it, there.

But as I pointed out, the Giants went from 31st to 2nd the following year with the same DC and same scheme, and it wasn't analytics related.

hands, footwork, leverage, or angles he's taking coming out of his stance, if he's a lazy piece of crap and unwilling to apply them or work at them it still doesn't matter.

And that's what the analytics can't account for.

In comment 15274279 Brown_Hornet said:

Quote:

... I think fans especially those there educated with regards to analytics overestimate what the coaching staff is actually doing with the information. Are they using it... absolutely but they've been using it since the 1970s. I believe that there is someone on every NFL team possibly even a group of people that compile this information and that is a relatively new function of the NFL machine. That said I believe thesr statistics/analytics/averages/trends, whatever you decide you want to call them, are used to support what the coaching staff already sees by studying their opponents and their own teams. I do not believe that analytics are guiding teams to model Their offensive or defensive coach and philosophies to something different than what they believe their players are best suited for.

They're not guiding the teams to model their philosophies yet. Within 10 years I'm confident they will.

Also, I see some people pointing to one specific play like it disproves analytically driven gameplanning. Even in baseball sometimes the bullpen matchup doesn't work and he gives up a HR. That doesn't mean the analytics didn't work, just that the the 10% chance hit instead of the 90% one. The idea is that in the long run teams will win more than they lose, not that they'll win every at bat.

In comment 15274483 Scooter185 said:

Quote:

In comment 15274279 Brown_Hornet said: Quote: ... I think fans especially those there educated with regards to analytics overestimate what the coaching staff is actually doing with the information. Are they using it... absolutely but they've been using it since the 1970s. I believe that there is someone on every NFL team possibly even a group of people that compile this information and that is a relatively new function of the NFL machine. That said I believe thesr statistics/analytics/averages/trends, whatever you decide you want to call them, are used to support what the coaching staff already sees by studying their opponents and their own teams. I do not believe that analytics are guiding teams to model Their offensive or defensive coach and philosophies to something different than what they believe their players are best suited for. They're not guiding the teams to model their philosophies yet. Within 10 years I'm confident they will. Also, I see some people pointing to one specific play like it disproves analytically driven gameplanning. Even in baseball sometimes the bullpen matchup doesn't work and he gives up a HR. That doesn't mean the analytics didn't work, just that the the 10% chance hit instead of the 90% one. The idea is that in the long run teams will win more than they lose, not that they'll win every at bat.

I didn't say it disproved anything. I said having only 17 games and a razor thin margin for error when it comes to making the playoffs because of it makes it harder to wait for it to "even out".

Baseball can afford to wait that out because they play 162 games. One loss can change an NFL team's fortune. That doesn't happen in baseball. In the NFL every win is magnified and "the long run" ain't long enough.

In comment 15274466 Britt in VA said:

Quote:

hands, footwork, leverage, or angles he's taking coming out of his stance, if he's a lazy piece of crap and unwilling to apply them or work at them it still doesn't matter. And that's what the analytics can't account for.

You're tilting at windmills at this point.

And scouts can't account for that stuff either otherwise Flowers wouldn't have been a 1st round pick...

In comment 15274491 giants#1 said:

Quote:

In comment 15274466 Britt in VA said: Quote: hands, footwork, leverage, or angles he's taking coming out of his stance, if he's a lazy piece of crap and unwilling to apply them or work at them it still doesn't matter. And that's what the analytics can't account for. You're tilting at windmills at this point. And scouts can't account for that stuff either otherwise Flowers wouldn't have been a 1st round pick...

Why is that tilting at windmills?

Is it not relevant to point out that the game is still played by humans and human nature is imperfect or inconsistent from day to day and that the math has no way of accounting for that?

Britt - you're starting to rail against arguments no one is making.

Take a look at the first point I posted yesterday -- all decisions are made on data, the key is to use the best set.

In your example of Flowers -- sure in the near future you'll have a data set from sophisticated telemetry, that will give insights on his physical performance.

If you're the Giants you'll also have a better psychological profile (if you haven't take a look at the changes Judge has made in the player development department), and data about how his profile projects.

And you'll have your own data, you experiences with lazy people.

The smart decision is probably pulling from all those data sets to make a decision on his likely success.

In comment 15274500 christian said:

Quote:

Britt - you're starting to rail against arguments no one is making. Take a look at the first point I posted yesterday -- all decisions are made on data, the key is to use the best set. In your example of Flowers -- sure in the near future you'll have a data set from sophisticated telemetry, that will give insights on his physical performance. If you're the Giants you'll also have a better psychological profile (if you haven't take a look at the changes Judge has made in the player development department), and data about how his profile projects. And you'll have your own data, you experiences with lazy people. The smart decision is probably pulling from all those data sets to make a decision on his likely success.

I can see this. Player selection. Athletic measurements and psychological profiles. I can see data being useful for that.

Salary cap, who to keep, injuries, etc... I can see all of that too.

Where I'm getting lost is on the field of play. And that's what I'm mainly talking about. The data affecting the natural way the game is played. Like baseball.

In comment 15274488 Britt in VA said:

Quote:

They're not guiding the teams to model their philosophies yet. Within 10 years I'm confident they will. Also, I see some people pointing to one specific play like it disproves analytically driven gameplanning. Even in baseball sometimes the bullpen matchup doesn't work and he gives up a HR. That doesn't mean the analytics didn't work, just that the the 10% chance hit instead of the 90% one. The idea is that in the long run teams will win more than they lose, not that they'll win every at bat. I didn't say it disproved anything. I said having only 17 games and a razor thin margin for error when it comes to making the playoffs because of it makes it harder to wait for it to "even out". Baseball can afford to wait that out because they play 162 games. One loss can change an NFL team's fortune. That doesn't happen in baseball. In the NFL every win is magnified and "the long run" ain't long enough.

Again, a slimmer margin of error means you need to do even more to maximize your chances of winning.

And comparing 16 (17) games to a 162 game season is misleading. A player gets 4-5 ABs/game in baseball and roughly 600 per season. There's ~60 plays (on each side) in an NFL game. That works out to over 1000 plays for a season, more than enough for individual plays to normalize themselves.

In comment 15274498 Britt in VA said:

Quote:

In comment 15274491 giants#1 said: Quote: In comment 15274466 Britt in VA said: Quote: hands, footwork, leverage, or angles he's taking coming out of his stance, if he's a lazy piece of crap and unwilling to apply them or work at them it still doesn't matter. And that's what the analytics can't account for. You're tilting at windmills at this point. And scouts can't account for that stuff either otherwise Flowers wouldn't have been a 1st round pick... Why is that tilting at windmills? Is it not relevant to point out that the game is still played by humans and human nature is imperfect or inconsistent from day to day and that the math has no way of accounting for that?

Because nobody that sees value in analytics is arguing that there is NO human element to the game. Or even that the math can predict what happens on a specific play. Analytics is about collecting and analyzing the data to maximize your chances of success in a game. Even with the simplicity of baseball, you can't predict ABs or even individual games.

In comment 15274506 giants#1 said:

Quote:

In comment 15274488 Britt in VA said: Quote: They're not guiding the teams to model their philosophies yet. Within 10 years I'm confident they will. Also, I see some people pointing to one specific play like it disproves analytically driven gameplanning. Even in baseball sometimes the bullpen matchup doesn't work and he gives up a HR. That doesn't mean the analytics didn't work, just that the the 10% chance hit instead of the 90% one. The idea is that in the long run teams will win more than they lose, not that they'll win every at bat. I didn't say it disproved anything. I said having only 17 games and a razor thin margin for error when it comes to making the playoffs because of it makes it harder to wait for it to "even out". Baseball can afford to wait that out because they play 162 games. One loss can change an NFL team's fortune. That doesn't happen in baseball. In the NFL every win is magnified and "the long run" ain't long enough. Again, a slimmer margin of error means you need to do even more to maximize your chances of winning. And comparing 16 (17) games to a 162 game season is misleading. A player gets 4-5 ABs/game in baseball and roughly 600 per season. There's ~60 plays (on each side) in an NFL game. That works out to over 1000 plays for a season, more than enough for individual plays to normalize themselves.

But we're not talking 1 vs 1 in football. We're talking 11 vs 11. That's a lot more variables on any given play in football that can affect other players around them regardless of what any individual player is doing. For instance, since we've talked so much about QB's. Trying to measure Daniel Jones on a play where his left tackle whiffs a block, his 1st read at WR gets stoned at the LOS in press coverage and his RB which is his outlet pass has to try and pick up the missed block all affects Daniel Jones. But on the stat sheet, all that shows up is Daniel Jones sack fumble, right? Is that an accurate measurement of Daniel Jones the individual?

In comment 15274505 Britt in VA said:

Quote:

I can see this. Player selection. Athletic measurements and psychological profiles. I can see data being useful for that. Salary cap, who to keep, injuries, etc... I can see all of that too. Where I'm getting lost is on the field of play. And that's what I'm mainly talking about. The data affecting the natural way the game is played. Like baseball.

Ok, take this example. You're playing the Chiefs and down 14-0 late in the 1st quarter. You know they're going to score more points. It's 4th and one from the Chiefs 32. Does your 'gut' tell you to kick the 49 yard FG or go for the 1st?

Even without analytics, it should be an easy call to play aggressive and go for the first (and if successful the TD). Where analytics helps, is when you're not playing a high powered offensive team and the score is closer. If instead you're facing the Jets and only down 7 your gut probably says to play more conservatively (risk averse) and take the FG. Analytics tells you, you still have a better chance of maximizing success by going for it (i.e. your gut is wrong).

In comment 15274518 giants#1 said:

Quote:

In comment 15274505 Britt in VA said: Quote: I can see this. Player selection. Athletic measurements and psychological profiles. I can see data being useful for that. Salary cap, who to keep, injuries, etc... I can see all of that too. Where I'm getting lost is on the field of play. And that's what I'm mainly talking about. The data affecting the natural way the game is played. Like baseball. Ok, take this example. You're playing the Chiefs and down 14-0 late in the 1st quarter. You know they're going to score more points. It's 4th and one from the Chiefs 32. Does your 'gut' tell you to kick the 49 yard FG or go for the 1st? Even without analytics, it should be an easy call to play aggressive and go for the first (and if successful the TD). Where analytics helps, is when you're not playing a high powered offensive team and the score is closer. If instead you're facing the Jets and only down 7 your gut probably says to play more conservatively (risk averse) and take the FG. Analytics tells you, you still have a better chance of maximizing success by going for it (i.e. your gut is wrong).

That seems reasonable. But I'm not talking about taking 3 points vs. 7 points. I'm talking about being at midfield or your own territory in that situation when you do it. Basically, you punt and you can stay in the game. You go for it and fail, it might be lights out. Or, you go for it and get it, but there is still no guarantee that you score on the drive. You just potentially set up for another decision. And the more 4th and 1's you end up going for, sooner or later...

63% is a lot, but it's still not that far from 50/50 when it comes to the amount of times any single team might have to make a decision on 4th and 1 in a season.

In comment 15274511 Britt in VA said:

Quote:

But we're not talking 1 vs 1 in football. We're talking 11 vs 11. That's a lot more variables on any given play in football that can affect other players around them regardless of what any individual player is doing. For instance, since we've talked so much about QB's. Trying to measure Daniel Jones on a play where his left tackle whiffs a block, his 1st read at WR gets stoned at the LOS in press coverage and his RB which is his outlet pass has to try and pick up the missed block all affects Daniel Jones. But on the stat sheet, all that shows up is Daniel Jones sack fumble, right? Is that an accurate measurement of Daniel Jones the individual?

1. Those are traditional metrics, not advanced analytics

2. Who in their right mind would try and judge a player, especially a QB, based on a single play?

This extreme example aside, analytics would actually bolster your case. Just look at Next Gen Stats and their metric that tells you how much separation a WR/TE has on plays he's targeted. But even this stat needs some context - e.g. are the windows tight because the WRs can't generate separation (see 2020 Giants) or is the QB making poor reads?

If the data isn't there yet, then the smart teams would combine that data with scouting breakdowns to make that determination. But the next step would be to develop the algorithms for determining that so your scouts can focus elsewhere.

In comment 15274523 Britt in VA said:

Quote:

That seems reasonable. But I'm not talking about taking 3 points vs. 7 points. I'm talking about being at midfield or your own territory in that situation when you do it. Basically, you punt and you can stay in the game. You go for it and fail, it might be lights out. Or, you go for it and get it, but there is still no guarantee that you score on the drive. You just potentially set up for another decision. And the more 4th and 1's you end up going for, sooner or later... 63% is a lot, but it's still not that far from 50/50 when it comes to the amount of times any single team might have to make a decision on 4th and 1 in a season.

You punt and it can be lights out. You seem to assume that you're going to kick a perfect punt that pins them deep and then your D will follow it up with a 3 and out.

go for it on 4th and 1. But analytics provide you with additional data that you can use to make the decision. It can't predict what happens on the play, but it can tell you how much a successful outcome improves your chances of winning and quantify how risky of a decision it is.

In comment 15274505 Britt in VA said:

Quote:

Where I'm getting lost is on the field of play. And that's what I'm mainly talking about. The data affecting the natural way the game is played. Like baseball.

Take 2 of the major revelations in baseball in the analytics era:

1) Roughly 85% of the time a ground ball is hit to the middle or pull side of the batter

2) Bunting with a runner on first decreases your expected runs per inning by 20%

You can micro dissect infinite what ifs, and outliers, and hypotheticals -- but over 10s of thousands of real life experiments -- this holds true.

What should a baseball manager do? Ignore the data because it impacts the "natural way."

So now take football -- two emerging theories are the offensive win rate of play action and 4 WR sets is substantively higher than other scenarios. Proving this out won't take 25 years of data modeling -- it's a months long exercise.

Let's say you run the data -- and after 10s of thousands real world experiments it holds true.

What should a football coach do? Ignore the data because it impacts the "natural way."

In comment 15274570 christian said:

Quote:

In comment 15274505 Britt in VA said: Quote: Where I'm getting lost is on the field of play. And that's what I'm mainly talking about. The data affecting the natural way the game is played. Like baseball. Take 2 of the major revelations in baseball in the analytics era: 1) Roughly 85% of the time a ground ball is hit to the middle or pull side of the batter 2) Bunting with a runner on first decreases your expected runs per inning by 20% You can micro dissect infinite what ifs, and outliers, and hypotheticals -- but over 10s of thousands of real life experiments -- this holds true. What should a baseball manager do? Ignore the data because it impacts the "natural way."

Clearly bunt since the pitcher is up and he has a 0.030 lifetime average! :D

but your reliance on the data may actually come at a detriment.

Quote:

So now take football -- two emerging theories are the offensive win rate of play action and 4 WR sets is substantively higher than other scenarios. Proving this out won't take 25 years of data modeling -- it's a months long exercise.

Whereas in baseball, certain theories won't change under baseball players change significantly, including their fundamentals, in football, the win rate will be significantly altered by defensive adjustments, both in the short term and over time.

So modeling something may be a months long exercise and the information be obsolete in another few months time. So where have you gotten to?

The question isn't if modeling will help. It is how much will it help, how much will it hurt and how much will it actually evolve the game.

And you have people pitching to NFL teams that they can build analytic programs that will deliver "optimal strategy" or a proven, winning game theory now. That's the point that many in the NFL believe will take 25 years and many believe will never happen.

The data science community also needs to be leery too because they are a lot of snake oil salesmen trying to sell teams on things they can't deliver, which adds noise to the legitimate companies that can bring in systems that will aid in certain areas.

on talking about where laughter comes from:

Quote:

- The notion football is too complex to model gets a hardy laugh, much more complex outcomes are modeled accurately. - The medium-term advances will be about tracking scenarios that work. The hardiest laughs are reserved for the notion "You don't know the play call or responsibilities."

Wasn't too long ago that many people here who discuss analytics found Gettleman's description of "computah people" insulting. Putting the shoe on the other foot and laughing about something that the league has an opposite take on probably shouldn't happen.

That air of superiority or the geeks vs. jocks angle isn't helping to advance analytics in the league right now.

if offenses start using play action 75% of the time, Ds will adjust and not bite as often (or at all). But smart offenses will still take advantage of the tendency (really instincts) of LBs/Ss to take a false step until that correction occurs.

In comment 15274598 giants#1 said:

Quote:

if offenses start using play action 75% of the time, Ds will adjust and not bite as often (or at all). But smart offenses will still take advantage of the tendency (really instincts) of LBs/Ss to take a false step until that correction occurs.

Hasn't that been happening all along, already though? For decades?

Analytics is being used in soccer to great effect all over the world. Same number of players as football, but I'd imagine a far greater range of possible movements. In soccer any player (including the goalie) can go anywhere on the field and perform the same offensive and defensive functions. If it can happen there, it can happen in football where the roles are far more defined and restricted.

A lot of this thread doesn't even sound like it's about analytics. It reads more like pining for the good old days and shunning evolution.

In comment 15274612 Go Terps said:

Quote:

Analytics is being used in soccer to great effect all over the world. Same number of players as football, but I'd imagine a far greater range of possible movements. In soccer any player (including the goalie) can go anywhere on the field and perform the same offensive and defensive functions. If it can happen there, it can happen in football where the roles are far more defined and restricted. A lot of this thread doesn't even sound like it's about analytics. It reads more like pining for the good old days and shunning evolution.

I started this thread using 3 current articles that, I felt, talk about analytics in football from all sides and angles. Then I asked three questions about the future.

You can't talk about the future of the sport without discussing the past, how things have always been done, and how much (or how little) these innovations will change that in the future.

It was a genuine attempt to understand and try to find a middle ground in the discourse, here, and there has been nothing inflammatory or abrasive, from anybody as far as I can see.

where does all of this data come from that you are using to predict future results?

The past.

In comment 15274589 FatMan in Charlotte said:

Quote:

but your reliance on the data may actually come at a detriment. Quote: So now take football -- two emerging theories are the offensive win rate of play action and 4 WR sets is substantively higher than other scenarios. Proving this out won't take 25 years of data modeling -- it's a months long exercise. Whereas in baseball, certain theories won't change under baseball players change significantly, including their fundamentals, in football, the win rate will be significantly altered by defensive adjustments, both in the short term and over time. So modeling something may be a months long exercise and the information be obsolete in another few months time. So where have you gotten to? The question isn't if modeling will help. It is how much will it help, how much will it hurt and how much will it actually evolve the game. And you have people pitching to NFL teams that they can build analytic programs that will deliver "optimal strategy" or a proven, winning game theory now. That's the point that many in the NFL believe will take 25 years and many believe will never happen. The data science community also needs to be leery too because they are a lot of snake oil salesmen trying to sell teams on things they can't deliver, which adds noise to the legitimate companies that can bring in systems that will aid in certain areas.

I'm confident the operations leaders at the NFL level are savvy enough to separate dubious claims from helpful job aids, which is what data analysis is after all.

Like I posted above -- I'm more than happy to track and debate the progress and the usefulness of data analysis every year for the next 25 years.

Football is it not the infinitesimally complicated set of outcomes it appears some of your friends who work in the NFL believe it to be.

No one in my field believes data analysis is a crystal ball or prediction engine. What we do believe is there are fundamental limits to the variables, and just like in baseball if you work backward from successful outcomes, there are components you can identify over diverse and large data sets.

These insights will work in concert with other factors.

The biggest contribution, just like it was in baseball, will likely be the elimination of brain dead conventional wisdoms that frequently don't work.

In comment 15274596 FatMan in Charlotte said:

Quote:

Wasn't too long ago that many people here who discuss analytics found Gettleman's description of "computah people" insulting. Putting the shoe on the other foot and laughing about something that the league has an opposite take on probably shouldn't happen. That air of superiority or the geeks vs. jocks angle isn't helping to advance analytics in the league right now.

I'm happy the Giants appear to have evolved to join the fray with the geeks. They'll be better for it.

hypocrisy..

For example, analytics is overused in baseball, in my opinion. It started with the A's and Billy Beane2 decades ago. However,it really was used in some form for a while. For example, the 90s Yankee dynasty didn't have an analytics department, per se, but they were also making a priority getting on base and making contact, just not to the extreme as the As because they didn't have money factoring in to the decisions.

Flash forward to present day. Most, if not all, teams are employing some level of analytics. But, there are always winners and losers. Does that mean it didn't work for the loser? Not necessarily. Then again, it is oversimplified. I hear this a lot in dissecting last year's Yankees-Rays playoff series. On one hand, analytics didn't "win" because both teams employed analytics. Likewise, analytics didn't lose. However, the part that is being missed is that the 2 teams have different forms of analytics. The Rays are still placing emphasis on stats like OBA and other metrics because they don't have a power packed lineup. In contrast, the Yankees analytics is more focused on launch angle, exit velocity, etc. While this has worked, in large part, in the regular season, it has failed them in the playoffs when you generally face more good pitching.

In comment 15274505 Britt in VA said:

Quote:

In comment 15274500 christian said: Quote: Britt - you're starting to rail against arguments no one is making. Take a look at the first point I posted yesterday -- all decisions are made on data, the key is to use the best set. In your example of Flowers -- sure in the near future you'll have a data set from sophisticated telemetry, that will give insights on his physical performance. If you're the Giants you'll also have a better psychological profile (if you haven't take a look at the changes Judge has made in the player development department), and data about how his profile projects. And you'll have your own data, you experiences with lazy people. The smart decision is probably pulling from all those data sets to make a decision on his likely success. I can see this. Player selection. Athletic measurements and psychological profiles. I can see data being useful for that. Salary cap, who to keep, injuries, etc... I can see all of that too. Where I'm getting lost is on the field of play. And that's what I'm mainly talking about. The data affecting the natural way the game is played. Like baseball.

Britt, I said this earlier...
And it is the #1 item listed in the areas most likely to be affected by analytics.
Analytics is not going to be a good tool to determine the outcome of a single player or a single player or even a single game.
It can point you in a direction saying that a certain style of play is more effective, and that players with certain profile flourish in that style of play.
So to go back to your 2015/2016 comments. Perhaps some analytics would have provided some insight into the team construction and suggested that there were flaws on defense that could have been corrected in the 2015 off season.
But even then, it is only going to point you in a direction. It won't provide final definite answers, just improved odds of making better decisions along the way.
The accumulation of many slightly better decisions should lead to more wins.