Jerry, Evan or others, any interest in computing a 5-10 year playoff only RAPM?
How good is a 5-10 year playoff only RAPM at predicting player and team performance in the next playoffs or say getting to the conference finals? How good is it compared to using regular season RAPM?
For the 9 seasons that we have prior-informed RAPM the champ's top 2 stars have combined for at least +9 on 7 occasions with the average probably near +10. Detroit's top 2 in its title season were about +8. The Spurs top 2 in 2003 were +7. Both may be low in part because of being in the early part of the period of prior informed ratings.
Which teams have top 2s who have combined for at least +9 this season? Dallas +12, LAC +9, Miami +10, Chicago +9, OKC barely below +9 (riding Collison and then Durant and not Westbrook), San Antonio +10 (Ginobili and Bonner, not Duncan & Parker). Boston, Portland and Orlando might be a bit short at +8. Denver, Houston and Memphis at +7. Atlanta at +6. Lakers at +5. Others might be in that range. One could also look at top 3s or 4s.
How frequently does the champ have the best top 2 or one of the 2-3 best top 2s on RAPM? I haven't calculated that yet but it might be worth looking at.
Which metric correlates best for predicting title winners (in contrast to the focus on overall predictive quality) based on prior year team numbers or current regular season team or top player numbers? Seems like it would be worth the time if anyone wanted to do it.
More with RAPM
Re: More with RAPM
Evan has the 4 factors of APM and other related data at his site thecity2.com.
On the adjusted factor for eFG% impact, Shaq in Cleveland and Boston at the tail end of his career still had almost twice the estimated impact of anybody else.
But correlation isn't necessarily causation. I am skeptical that K Perkins had the 7th biggest impact on eFG%. Almost anyways playing with Boston's big 3 and KD in OKC may have given an overly friendly estimate.
Kemba Walker with the 10th worst mark in the league on this factor and in the bottom 25 on overall A4PM. But almost everybody in the league and outside of it is going to continue to use boxscore metrics exclusively and think somewhat or a lot better of him and almost completely or completely ignore APM and his estimates on it. Call me when K Walker is actually leading his team to a good result. If that ever happens I am pretty confident that his RAPM won't be bottom 25 in the league as it is now.
Only 2 PGs make the top 20 on beneficial impact on own turnovers- Paul and L Williams, J Terry just outside.
Westbrook with very little estimated impact on any factor and only +0.5 overall. But he gets max contract and multiple All-Star status and lots of attention... largely because of his extremely high usage and dunks I guess. Team record too, whether it is because of him or not is hard to say for sure but this metric doesn't think it is.
Toronto, Houston, Orlando, Cleveland, Washington and Phoenix near the top (along with a few others) on number of players employed during the 2.5 year period with a negative A4PM (at least 8 and as many as 13) and worse than a -1 A4PM (at least 8). All with consultants, four of which have experience calculating APM. But it doesn't seem like reading publicly available APM is getting a lot of or "enough" weight in player evaluation. Maybe they are spending a lot of time and giving a lot of weight to stuff being sorted by other sophisticated technical programming languages & techniques or traditional scouting or whatever. Not working well for the majority of these 6 and seems unlikely to lead to a title or even conference finals appearance for any this season, like last season. Orlando had it working earlier then changed personnel significantly and it stopped working as well as before.
Chicago, Memphis, OKC and Dallas are among the teams with the fewest negative RAPM players employed in the period. By informed choice or otherwise.
Fesenko with the best estimated impact on opponent eFG%. How many of the teams thinking about hiring him know this or give it weight? For all the guys who teams employ with lousy performance, lousy metric estimates on boxscore stats or even RAPM, is no one interested in a big man with the best estimated impact on opponent eFG%? Overall he is estimated at +2.82 on A4PM. I'd hire him even it is to just mostly sit on the bench.
On the adjusted factor for eFG% impact, Shaq in Cleveland and Boston at the tail end of his career still had almost twice the estimated impact of anybody else.
But correlation isn't necessarily causation. I am skeptical that K Perkins had the 7th biggest impact on eFG%. Almost anyways playing with Boston's big 3 and KD in OKC may have given an overly friendly estimate.
Kemba Walker with the 10th worst mark in the league on this factor and in the bottom 25 on overall A4PM. But almost everybody in the league and outside of it is going to continue to use boxscore metrics exclusively and think somewhat or a lot better of him and almost completely or completely ignore APM and his estimates on it. Call me when K Walker is actually leading his team to a good result. If that ever happens I am pretty confident that his RAPM won't be bottom 25 in the league as it is now.
Only 2 PGs make the top 20 on beneficial impact on own turnovers- Paul and L Williams, J Terry just outside.
Westbrook with very little estimated impact on any factor and only +0.5 overall. But he gets max contract and multiple All-Star status and lots of attention... largely because of his extremely high usage and dunks I guess. Team record too, whether it is because of him or not is hard to say for sure but this metric doesn't think it is.
Toronto, Houston, Orlando, Cleveland, Washington and Phoenix near the top (along with a few others) on number of players employed during the 2.5 year period with a negative A4PM (at least 8 and as many as 13) and worse than a -1 A4PM (at least 8). All with consultants, four of which have experience calculating APM. But it doesn't seem like reading publicly available APM is getting a lot of or "enough" weight in player evaluation. Maybe they are spending a lot of time and giving a lot of weight to stuff being sorted by other sophisticated technical programming languages & techniques or traditional scouting or whatever. Not working well for the majority of these 6 and seems unlikely to lead to a title or even conference finals appearance for any this season, like last season. Orlando had it working earlier then changed personnel significantly and it stopped working as well as before.
Chicago, Memphis, OKC and Dallas are among the teams with the fewest negative RAPM players employed in the period. By informed choice or otherwise.
Fesenko with the best estimated impact on opponent eFG%. How many of the teams thinking about hiring him know this or give it weight? For all the guys who teams employ with lousy performance, lousy metric estimates on boxscore stats or even RAPM, is no one interested in a big man with the best estimated impact on opponent eFG%? Overall he is estimated at +2.82 on A4PM. I'd hire him even it is to just mostly sit on the bench.
Re: More with RAPM
I'd like to know what the positional averages are for offense, defense and total RAPM if possible. I think that'd be interesting to see.
Re: More with RAPM
I believe someone, maybe Jerry or someone else here, computed them previously or maybe it was a journal article but I don't immediately recall exactly where. I could probably find them but don't feel like searching for them at the moment.