A couple of interesting articles out today. First, from the NYT, an article on why Melo and the Knicks need one another, with insight from Jeff Van Gundy:
NYTimes Article
Next, CNNSport Money discusses NBA player ROI, and finds Langston Galloway to be the third most valuable player (in terms of ROI) in the league this past season:
NBA Player ROI
First article goes into speaking how Blatt would be a great fit for the team, and I agree. He's accomplished a lot, and given Phil's transitional role with the team, going forward Mills and Blatt working together would be a great situation. The only issue would be Dolan's reluctance to interfere once Phil does leave.
As for the player ROI bit, it makes some sense, although I wouldn't put too much credence into it without understanding what weight individual stats have on the metric, or how it's calculated in general. But it also brings to mind an occurrence that's unique to the NBA - Of the three major sports (I'm not including hockey because I'm not very familiar with it), the NBA is the only league that constrains individual salaries, rather than only total team salaries. Thus, max contract player ROI can fluctuate quite a bit, and I'd be curious to learn/develop some type of metric that can help determine at what point giving a player a max deal is a good decision. That'll probably be a project I work on a few months down the road.
#Tankbis.
He's best suited as a backup at this point. I'm not sure if his ego will allow for that to happen so I think a return to Dallas could be likely. There are a handful of teams that could use him as a starter and the prominent ones are Dallas, NY and BKN.
The crazy thing is, KOQ is still a good signing, despite him underperforming. He's locked in for 3 more years at $4M per, and his issues aren't talent - they're mental. If they can help him with that stuff, then he's a pretty good backup.
Quote:
Williams is opting out. Not an option for the Knicks (absolutely hated NY) but still notable.
He's best suited as a backup at this point. I'm not sure if his ego will allow for that to happen so I think a return to Dallas could be likely. There are a handful of teams that could use him as a starter and the prominent ones are Dallas, NY and BKN.
I'm assuming it's a foregone conclusion he remains in Dallas. Just saw it on the scroll and thought it was worth mentioning.
Uhhhhhh
Melo requesting less playing time so the young guys could get experience makes him me-first? Melo showing and talking about his willingness to move the ball to other players on the team (as I've referenced before, end of the SA game) makes him selfish?
What's the weather like back in 2008?
Funny, I'm perfecting my R and Python skills at the moment. My own personal project that I'll be working on more immediately is developing a metric to better understand a player's value based on his role within a team. If you've noticed from any of my basketball posts, I put very little credence in stats like PER, as well as all-in-one metrics in general. Look up "The Case for Dennis Rodman" to understand why a bit better.
My baseline thinking is that basketball, like soccer, is a fluid game. In theory, you could put 5 guys around the 3 point line, just like you could play a 2-2-6 in soccer. Similarly, players can have varying roles within a team. The best example would be point guard and a point forward. They serve the same purpose - playmaking - but they don't necessarily function the same way on the court.
Anyway, the project would basically entail understanding the roles and what stats are more important depending on the role being played. Then, developing separate role metrics to examine a player's ability within said role. Then, finally, utilizing historical player stats to see how good each player would be in a particular role.
As for the max player thing, it would require defining a metric like Win Shares and weighing it against the percentage a player's salary weighs against the cap.
Quote:
That's a pretty cool project to work on, mind sharing on how you are gonna go about doing it? I just started to learn Python and I am getting a degree in analytics next school year so I'm looking into doing some sort of fun project in finance, real estate, sports, etc.
Funny, I'm perfecting my R and Python skills at the moment. My own personal project that I'll be working on more immediately is developing a metric to better understand a player's value based on his role within a team. If you've noticed from any of my basketball posts, I put very little credence in stats like PER, as well as all-in-one metrics in general. Look up "The Case for Dennis Rodman" to understand why a bit better.
My baseline thinking is that basketball, like soccer, is a fluid game. In theory, you could put 5 guys around the 3 point line, just like you could play a 2-2-6 in soccer. Similarly, players can have varying roles within a team. The best example would be point guard and a point forward. They serve the same purpose - playmaking - but they don't necessarily function the same way on the court.
Anyway, the project would basically entail understanding the roles and what stats are more important depending on the role being played. Then, developing separate role metrics to examine a player's ability within said role. Then, finally, utilizing historical player stats to see how good each player would be in a particular role.
As for the max player thing, it would require defining a metric like Win Shares and weighing it against the percentage a player's salary weighs against the cap.
Yeah, I'm also against PER to an extent. Not sure if you heard of the book "Mathletics", was written by a former professor at my school and explains the faultiness in stats like PER. Not sure if you heard of a YouTube channel called "BBALLBREAKDOWN" but they did something on what the average NBA player stats look like and analyzed which players are overpaid. To be honest they did a shitty job (most of the stuff they make seems rudimentary and lacking in depth analysis) as they only did basic stats and neglected defense all together outside of blocks and steals (they listed Derrick Favors and Avery Bradley as two of the most overpaid "average" players). Seems like what you are doing is more inclusive of impact rather than stats.
Link - ( New Window )