Raptors: Career-Arc Regression Model Estimator with Local Optimization (CARMELO)

[mcHAPPY]

ESPN is currently promoting this:

*FiveThirtyEight's Career-Arc Regression Model Estimator with Local Optimization (CARMELO) is a system that forecasts a player's future performance. The similarity score is an index measuring how comparable one player is to another, scaled such that a score of zero is average similarity and 100 is the highest possible degree of similarity. Plus-minus is the number of points per 100 possessions that a player contributed to his team, relative to an average NBA player. All player ages are as of Feb. 1, 2016. Minutes played is prorated to 82 games.

http://projects.fivethirtyeight.com/carmelo/

How does it work?

The basic premise of CARMELO is simple. For each current NBA player, CARMELO identifies similar players throughout modern NBA history1 and uses their careers to forecast the current player’s future.

According to CARMELO, for example, Washington Wizards point guard John Wall, through this point in his career, is similar to former NBA players Isiah Thomas, Jason Kidd, Steve Francis and Kenny Anderson. Kidd continued to improve as a player through his mid-to-late 20s, while Thomas had a long peak and led the Detroit Pistons to two championships. So both are favorable comps for Wall. Francis and Anderson are less favorable. So while Wall has the potential to develop into a superstar, he’s not a sure thing.

http://fivethirtyeight.com/features/...player-career/

Player: Category

Kyle Lowry: All-Star

DeMar DeRozan: Average Starter

DeMarre Carroll: Key Role Player

Patrick Patterson: Key Role Player

Jonas Valanciunas: Rotation Player

Corey Joseph: Rotation Player

Terrence Ross: Rotation Player

James Johnson: Key Role Player

Luis Scola: Defensive Specialist

Bismack Biyombo: Rotation Player

Delon Wright: OK prospect

Norman Powell: Marginal prospect

Bruno Caboclo: WHO?????

Anthony Bennett: Project

Bebe: WHO?????

So … should I bet on these things?

Hmm. Umm. Probably not? FiveThirtyEight’s relatively simple, RPM-based projections performed quite well last year, edging out Vegas along with most other projection systems. In theory, based on our back-testing, CARMELO should be slightly more accurate still, improving on the simple RPM projections by about 10 percent. But back-testing is not the same thing as seeing how predictions perform in the real world against truly unknown data. Rookie forecasting models can be buggy, moreover. I’d probably hold off until the system has at least a year or two of experience under its belt.

http://fivethirtyeight.com/features/...player-career/

Lots to debate there!

[MixxAOR]

Sounds about right

[The Great One]

Really cool.

[The Great One]

Carmelo Anthony: Good starter!!
Andrea Bargnani: Scrub!!

LMAO!

[Shantz]

Ah yes, defensive specialist, Luis Scola.

[Scraptor]

I really enjoy advanced stats but these models suffer so much from Garbage In = Garbage Out.

I mean, 2 of the top 10 players I'd compare Jonas to are definitely Michael Sweetney and Eddy Curry. Lol.

[mcHAPPY]

neil paine, our senior sports writer, and i had a lot of debates (which echoed long-running arguments within the broader basketball stat-geek community) about which advanced statistics to use for carmelo before deciding on this bpm/rpm blend. What settled the debate was that the bpm/rpm blend did better than alternatives like per and win shares in a variety of out-of-sample testing.

However, no all-in-one advanced stat is magic, and this is a source of systematic uncertainty in any nba projection system. If it seems as though carmelo “loves” or “hates” a certain player, it may be because of how bpm and rpm rate the player. for instance, both bpm and rpm rate the raptors’ jonas valanciunas poorly compared with statistics like per. So if valanciunas’s forecast seems pessimistic to you, it’s not because carmelo expects his performance to decline (in fact, carmelo has him getting a little better). It’s because bpm and rpm didn’t evaluate valanciunas as being all that good to begin with.

fyi

[Mindlessness]

I'm enjoying the name.

[Just Is]

Carmelo Anthony: Good starter!!
Andrea Bargnani: Scrub!!

LMAO!

I died at this. I thought you were paraphrasing. It LITERALLY says that.

[FireDT]

Derrick Rose: ¯\_(ツ)_/¯

[Chr1s1anL]

interesting

[Mindlessness]

Derrick Rose: ¯\_(ツ)_/¯

It actually says that, by the way.

Also, Derrick Favors is listed as a borderline all star. Can we get him like, yesterday?

[KHD]

this is kind of a problematic projection tool in that the vast majority of comparables to any nba player are not going to be star players or even good players in the league.

In a small amount of playing around with it, it is highly pessimistic about almost every prediction except for guys who put up amazing stats really early on.

basically: if you're young, you probably won't be a star player, statistically speaking. if you're older, you'll probably decline.

[big boi]

Pretty sweet name. Seriously this is an ESPN product? I imagine CA and his entourage would have something to say about it.

[Scraptor]

http://fivethirtyeight.com/features/...rs-in-the-nba/

Scan this list of franchise players and tell me this model isn't broken. I don't care how smart Nate Silver and co are. If your model produces a list that tells you Elfrid Payton and Marcus Smart will be the 13th/14th most valuable players over the next six seasons, you should check your math.