His best years were with bad Don Nelson teams where they usually scored the most and gave up the most points. He was always in foul trouble and barely played 30 minutes a game. Yes he had a high FG% @ the rim, rebounded well and blocked some shots, but unstoppable offensive force? If 14pts per 36 minutes is unstoppable then I guess yes. :grin: Last year even when started he never played 20 minutes and probably played 10x more minutes during garbage time than he did in crunch time. #11? No freaking way.
OK, so you basically have no idea what RAPM is. Nobody is saying he is the 11th best defender in the league. However, he is pretty good at that end and a good rebounder. His drtg last year was 98, by far the best on Warriors. When he was on court the opponent shot 45.9% EFG. When he was off? 49.0% EFG. They blocked 3% more shots when he was on the court etc. He got his oversized contract because he was (about the only) plus defender and rebounder on those Nelly squads. The statistics above suggest he still can play some D, although as Carl Herrera said, the sample size isn't that big compared to most players on that list. Why he wasn't playing more? Well, his Offensive RPM was basically the worst in the league. Maybe that had something to do with it?
I agree Biedrins sample size is too small that luck probably plays a big part in his very good rank. If I remember correctly Defino had a crazy high +/- the first third of the season and as he got more crunch time and first team minutes his +/- dropped down to earth. I assume RAPM can swing the same way to a smaller degree. I can't explain why Millsap has such a good RAPM but he only got two years $19m for a reason. I've lived in Utah since his rookie season so I know he's a high motor all around player who does a little of everything, but isn't great @ any one thing. Showed great promise, but never lived up to the potential. Hasn't really improved his game much year to year during his time in Utah. Has a few SportsCenter moments and games during each season, but can be very average way more than he should. Has games when he disappears on the boards, including 10 games last season with 3 or less rebounds. Shoots 65% @ the rim and 40% between (3' < 3PT). Takes more shots between (3' < 3PT) than shots @ the rim. Taken more long FGs(16'<3PT) than between (3'<10'). 64% of his shots were assisted. Seems to score a few buckets per game on tips. Improved 3PT shooter, but very streaky. A good defender, but he isn't a lockdown defender that's going to stop Duncan, Dirk, Melo, LeBron, Faried, ZBo , Boozer, Davis or even Lee. Considered underrated by many here, but I would rather have a good bench player and/or picks instead of Millsap.
As I think more about RAPM, I start to have some doubts. Take Asik for example, he's a superstar according to last year's ranking. This year though, I don't even need to look at the updated number to know it'll be a lot worse. RAPM has two parts: One is the prior performance carried over into the equation, and Asik should benefit from last year's ranking. Another part is extracted from this years plus/minus(with some math), and the early result would put him as one of the worst even on our own team. If he continues to play in Rockets uniform, his RAPM would look much more mediocre. The similar thing happened for Lin, who's RAPM is superb for a 22/23 years old in 2011-2012 season but last year was much worse due to recovery and overlapping role with Harden. Since it ultimately comes from minus/plus numbers, Asik/Lin's RAPM number fluctuation seem to be heavily related to the roles they played in the system. Remember Asik is the ONLY defensive anchor we had last year, and Lin was the ONLY capable passing PG Knicks had during Linsanity. On today's rockets team, they are not unique anymore. Milsap's RAMP number may be due to the fact that his defense is way better than Carlos Boozer(carried over) and AJ in the Utah days. Overall, I think one should take multiple years' RAPM number along with aging curve and fit/role to form judgement about a player.
How a player is used certainly changes both his raw plus minus and his RAPM. If Dwight Howard was asked to play the PG and shoot the 3 pointer all the time, he would probably look incompetent and his stats (box score, plus minus, whatever else is out there) would look terrible. With Asik, his RAPM numbers have been good to great both during his time in Chicago and last season in Houston. Combine this with other info-- like scouting, watching him execute the game plan, etc.-- and one can be fairly confident about the type of player that he is and conclude that this year's bad plus minus is an aberration caused by 1) spending time in the ill-fitting twin tower lineup, 2) small sample size, and 3) perhaps lethargic play due to unhappiness. In any case, I agree that RAPM or any other number requires a significant sample size before we can draw conclusions from them. And it is only one piece of info out of many. But it is nevertheless a useful piece of info.
RAPM basically takes the strengths and weaknesses of +/- and exaggerates them, including needing a large enough sample (essentially someone's entire career) before being able to draw the proper contextual conclusions from the numbers. People who look at RAPM year-to-year and just simply compare the numbers as some sort of ranking are doing the exact same thing as people who look at +/- for a given game and then use that to evaluate the players' performances.
This is pretty much backwards. RAPM is an attempt to smooth address the issue of raw +/- being dependent on which other players you happen to be on the floor with by looking at a larger sample size and adjusting for the variable of the other players. It does not 100% succeed since there is a margin of error like all stats and luck is still a part of it. But it is still better than raw +/- and is certainly not the same as looking at the +/- for a single game. For example, Cole Adlrich had one game last year when his +/- was +40. The number is extreme and is not all that meaningful. But if you look at a guys' RAPM over an entire year or over several years, there is something informative to be learned. At the very least, it allows you to ask questions and look at other sources of info (watching tape, talking to coaches, etc.) to figure out why a player's numbers seems surprisingly high or low-- Is the stat in erorr due to luck or another factor or does it reflect a reality that we didn't notice before? Every stat, box score, advanced or whatever, requires a large enough sample size. You can't use one big plus minus game to justify a conclusion any more than you can use Tracy Murray's 50 point game to say that this is the level you expect him to play at.
http://stats-for-the-nba.appspot.com/teams/BOS.html <style type="text/css"> table.tableizer-table { border: 1px solid #CCC; font-family: Arial, Helvetica, sans-serif font-size: 12px; } .tableizer-table td { padding: 4px; margin: 3px; border: 1px solid #ccc; } .tableizer-table th { background-color: #104E8B; color: #FFF; font-weight: bold; } </style><table class="tableizer-table"> <tr class="tableizer-firstrow"><th>Name</th><th>MP</th><th>Off. On</th><th>Def. On</th><th>Off. xRAPM</th><th>Def. xRAPM</th></tr> <tr><td>Brandon Bass</td><td>2442.82</td><td>1.023</td><td>1.060</td><td>-2.43</td><td>-0.18</td></tr> </table>
It appears that the Utah Jazz home court advantage is the reason why Millsap is ranked so high in RAPM. Utah Jazz record during Millsap years: Total 324-234 (58%) Home 209-70 (75%) Away 115-164 (41%) During the prime DWill+Boozer years(2006-2009) they were 133-31 (81%) and (+10) @ home and 73-91 (45%) and (-1.2) per game on the road. Millsap only played 24 minutes per game so he was able to be a high energy guy off the bench. As Millsap became a starter his +/- dropped but still was much better @ home than on the road. Thoughts? Comments?
