Abusing "Win Shares" to predict our record

[BlazerCaravan]

So the season's nearly done, and we had a lot of injuries... but we got to find out how good our team is, what their performance ability looks like. And looking at our lineup, and their ability to contribute to wins, I came up with a basic formula to predict our wins next year, provided players perform at the rate they did this year.

Starters:
Roy (0.182 WS per 48min)
Miller (0.134)
Batum (0.188)
Oden (0.216)
Aldridge (0.147)

Bench:
Cunningham (0.128)
Fernandez (0.103)
Webster (0.104)
Bayless (0.102)
Camby (0.153)

Now, we adjust they WS/game by applying min/game (these are approximations based loosely on actual minutes per game):

Starters:
Roy (36min) - 0.1365
Miller (32min) - 0.0893
Batum (24) - 0.0940
Oden (24) - 0.108
Aldridge (36) - 0.11025

Total = 0.53805 wins per game.

Bench:
Cunningham (10) - 0.0267
Fernandez (18) - 0.038625
Webster (18) - 0.0390
Bayless (18) - 0.03825
Camby (24) - 0.0765

Total = 0.219075 wins per game

This 10-man lineup gets 0.757125 wins per game. Assuming no injuries (HA!), that means we win that percentage of the time, on average.

Over an 82-game season, we'll get 62 wins next year if we do nothing and our players perform at the same level as this year. Essentially, this is the record we could have had this year... essentially.

 

[maxiep]

I thought we'd win 58 games with our opening day roster. I think Miller is better than Blake, we don't miss Outlaw much and Camby is better than Howard or Przybilla. So improving by four games if we stay healthy sounds pretty reasonable to me.

Great stuff!

 

[Nikolokolus]

So assuming there are basketball gods, does this mean they turn their wrathful gaze away from us next year and pick on somebody else? One can only hope.

 

[Minstrel]

One has to factor in some injuries. A perfectly healthy team is nearly impossible for any franchise, let alone one that has some players with red flags.

But I don't know how I'd do that...Oden's injuries are so hard to use to project forward. They've been enormous in terms of games missed, but not as enormous in terms of being chronic or likely to repeat. Taking an average of time missed so far is likely to greatly underestimate his games played going forward, but assuming no injury recurrence is probably going to overestimate his games played.

 

[BrianFromWA]

So Blake Griffin will NEVER PLAY AGAIN!?!? What a bust!

 

[mook]

Assuming we re-sign Camby, and Roy and Oden return 100%, and assuming we don't sustain any major new season-ending injuries, 61-62 wins is probably what I would've predicted anyway. I had us at 60 this year before the season started.

That's a lot of assumptions, though.

 

[PtldPlatypus]

I decided to change your minute allocations, giving Miller 36 mpg, Oden, Camby, and Batum 30 each, and reducing the rest of the bench guys. Now we're a 65-win team.

 

[BlazerCaravan]

Using total minutes instead of per game would allow you to account for games missed. But that would require more time than I have at work today.

Basically, there's 19,680 minutes in a season (48 * 5 * 82). Each player can at a basic (not accounting for OT) maximum play 3936 minutes a season. So, if Oden played 41 games at 24 minutes a game, he'd play 984 minutes in a season.

So, you'd take the WS/48min, divide by 48 to get WS/min, then multiply by 984 to get his Wins per Season (he'd have 4.428 playing 24 min/game for 41 games). Do that for all 10 players, and you have your estimate for wins per season.

 

[BlazerCaravan]

Okay, I actually did it by populating a spreadsheet an calculating it. This has the advantage of letting me use it as a toy. I got 61 wins out of it.

 

[oldmangrouch]

I know, I'm a horrible cynic, but....

I wonder what the "win share per 48" for Roy looked like *before* Oden got hurt.

 

[BlazerCaravan]

I know, I'm a horrible cynic, but....

I wonder what the "win share per 48" for Roy looked like *before* Oden got hurt.

Impossible to tell, unfortunately. But, given our record at the time, probably lower.

However, as an optimist, seeing Roy and Andre working together without Blake to interlope, the entire "Roy can't function with [Player Name Here] in the lineup" argument seems to have been countered pretty handily.

And really, they can still get the same WS/48 as this year... they just have to win more games than this year.

 

[MARIS61]

I doubt our team next year will resemble this team at all.

It's clear KP is about to be de-throned and that means someone else (probably that idiot Miller) will re-shape the team to leave their mark.

Our window is closing this summer, and we'll be starting our 3rd rebuild of this decade.