holfresh wrote:

dk7th wrote:

holfresh wrote:

mreinman wrote:

holfresh wrote:

mreinman wrote:do you think that Kemba's disgusting TS of 46 had anything to do with that? Again, you are being way to black and white.

Don't really remember but I thought they shot the same fg?. not sure..again, team win total is part of the formula..

their efficiency numbers were not close. You need to look closer at the numbers before making such arguments. This was quite obvious.

Also, I don't look at WS, I look at WS48.

Either one, I think Butler was higher in both..I guess dk7 went to bed..

here is the breakdown of how win shares are calculated. there is little here to suggest that a player's win shares are entirely dependent on what team he is playing on. the formula clearly demonstrates an individuals relationship to the league as a whole, which in effect nullifies "what team he plays on" as an independent variable. in fact, as i said earlier, it seems you have the matter reversed: the total number of wins is reflective of the collective win shares of the individuals comprising the team. this is the reason why the author took pains to point out how closely win shares can or should mirror the total number of wins for that team. if that were not the case then the formula would be flawed.

http://www.basketball-reference.com/about/ws.html

as to the justification of my methods of predicting wins for the knicks this season, if i were only using win shares for one season for each player being replaced and then doing the same for the players who replaced them, you could have a point. however, i took a simple average over their entire careers to add wins to the 17 win team of last year to yield a minimum of 27 wins for this coming season, not including anthony or our two rookies. even a bad anthony is good for 4 wins. the rookies is anybody's guess, since they have no nba track record.

I have to admit I didn't research this in depth using different teams as I should but my contention is that in this formula and it actually tells you in the title what he is doing:
4. Calculate marginal points per win. Marginal points per win reduces to 0.32 * (league points per game) * ((team pace) / (league pace)). For the 2008-09 Cavaliers this is 0.32 * 100.0 * (88.7 / 91.7) = 30.95.

The .32 is a variable unique only to the Cleveland Cavaliers who won 66 games that year...That number would be .09 for the Wizards in the same year who won 19 games or for the Bulls,.20 who won 41 games that year...So the formula set forth a subtle thing that can be missed or maybe intentionally put in that manner so folks like you will miss it...But this isn't a fixed variable across the board for all players on all teams but a value unique to players of that specific team...I could be wrong, but I will do the math later and I'm willing to bet I'm right...

Looking at the entire formla, it's the only way they could make WS of each player when added, equal to that of the total team wins...There are no other value that corresponds with team wins...

the figures of .92 and .32 are constants, not variables as you would have them be-- what would be the point in that?otherwise the formula would have no objective validity, which is what you want.

i can see why you want to make them variable from one team to the next in terms of their wins, because creating a circular argument in this way undermines the utility of the statistic.