Wes, Westbrook, and the Coat-Tails Effect
Apr 2, 2015 15:02:27 GMT -6
20s Navidad and Heebs like this
Post by eric on Apr 2, 2015 15:02:27 GMT -6
Inspired by a discussion in shout today. On/off is a great tool because it sees everything, which obviously includes many things that are missed by the box score. If Nick Collison sets a great screen and frees up a teammate for a wide open shot, he doesn't get an assist. If he boxes out Roy Hibbert and a teammate snares the board, he doesn't get half a rebound. If he fights over Hibbert to front the post, forcing a difficult pass that sails out of bounds, he doesn't get a steal. On/off is a not-so-great tool because it sees too much. NBA teams don't have randomized lineups, so some players are more likely to play with each other or to replace a specific player in a given lineup. Does Nick Collison have great on/off because he does all the little things, or because he replaces Kendrick Perkins, Ruiner of Everything? Does Russell Westbrook have great on/off because he is a great player, or because he plays a lot of minutes with Kevin Durant?
One way to look into these questions is to look not at overall on/off but specific player combinations. www.nbawowy.com is the site I use for this, and it is great. I'm going to run down several pairings that are of interest to me. Of course, there are going to be hidden correlations here too just like for any other on/off measurement. For example, the LeBron era Heat would frequently have Chris Bosh anchor bench units while the substitution patterns for Wade and LeBron would be more in tune. Thus a Bosh/Wade comparison would overstate the worth of Wade because his minutes had a higher proportion of LeBron. But let's see what we can see anyway:
Unfortunately both of the last two seasons have super tiny sample sizes for the Thunder pairings so 2013 is the most recent read we can do. It's really interesting to see which pairings have emergent properties where having both is better than either alone, and it's not always the obvious ones. Lillard plays great offense, Matthews plays great defense, put them together and you get the best of both worlds (Gogeta). Parker plays great offense, Duncan plays great defense, put them together and you're worse off than if you just played Duncan (Veku). Sometimes the obvious lack of synergy is reflected in the numbers: Westbrook holds Durant back, no doi.
Wes is clearly a valuable player, +11 per 100 is championship material. But he's not more valuable than Lillard, clearly. He's just too limited on the offensive end from a creative standpoint.
One way to look into these questions is to look not at overall on/off but specific player combinations. www.nbawowy.com is the site I use for this, and it is great. I'm going to run down several pairings that are of interest to me. Of course, there are going to be hidden correlations here too just like for any other on/off measurement. For example, the LeBron era Heat would frequently have Chris Bosh anchor bench units while the substitution patterns for Wade and LeBron would be more in tune. Thus a Bosh/Wade comparison would overstate the worth of Wade because his minutes had a higher proportion of LeBron. But let's see what we can see anyway:
2015 Trail Blazers
110 - 110 = +00 Lillard w/o Wes
099 - 107 = -08 Wes w/o Lillard
112 - 101 = +11 both
2015 Warriors
107 - 102 = +05 Klay w/o Curry
115 - 101 = +14 Curry w/o Klay
117 - 099 = +18 both
2015 Spurs
107 - 099 = +08 Duncan w/o Parker
111 - 110 = +01 Parker w/o Duncan
108 - 103 = +05 both
2013 Thunder
114 - 101 = +13 Durant w/o Westbrook
109 - 103 = +06 Westbrook w/o Durant
115 - 103 = +12 both
2014 Pacers
104 - 100 = +04 Hibbert w/o PG13
104 - 102 = +02 PG13 w/o Hibbert
107 - 099 = +08 bothUnfortunately both of the last two seasons have super tiny sample sizes for the Thunder pairings so 2013 is the most recent read we can do. It's really interesting to see which pairings have emergent properties where having both is better than either alone, and it's not always the obvious ones. Lillard plays great offense, Matthews plays great defense, put them together and you get the best of both worlds (Gogeta). Parker plays great offense, Duncan plays great defense, put them together and you're worse off than if you just played Duncan (Veku). Sometimes the obvious lack of synergy is reflected in the numbers: Westbrook holds Durant back, no doi.
Wes is clearly a valuable player, +11 per 100 is championship material. But he's not more valuable than Lillard, clearly. He's just too limited on the offensive end from a creative standpoint.
