APBRmetrics

RPM estimates don't get / display avg. error estimates (neither do box score metrics) but are known to have them. Modest or big? Hard to say since we don't have them and modest and big are subjective scalings. One can ignore RPM and its trends and make an argument to justify that or you can look at the trends with interest, curiosity and some caution. Many GMs are probably in the first group. Maybe almost all of them.

If you do look at the trends with interest, curiosity and some caution, here is some of what you'll find for selected centers:

3yr marks and trend for beloved professional Al Horford are real similar to those of overweight, doesn't listen Jared Sullinger. Both went from about plus 1.5 RPM estimate in 13-14 to about plus 3 in 15-16. Overall RPM estimates that are similar can vary in meaningful ways at offensive / defensive level and at four factor split level but one characterization is that the Celtics just gave $25 million more to Horford next season than the Raptors are giving Sullinger and may get a very similar overall impact if RPM is credible and the players don't change much.

By RPM estimates Noah was great in 2013-14 and dropped to modestly above mean performance the last 2 seasons. Sounds plausible or "right" in conjunction with eye test. Phil Jackson is probably unaware and uninterested in the RPM trend and betting on bounce back by Noah or at least fan buy in for the high cost "try".

Dwight Howard's RPM estimates have gone from over plus 4 to plus 2 to zero. But Budenholzer apparently doesn't know or care. Lopez over plus 2 in 2013-14 but barely above zero in 14-15 and at zero in 15-16. Apparently not a problem for the Bulls. Very similar trend for Al Jefferson. Reputations linger, perhaps longer than RPM estimates or "realities", at least it looks that way for the buyers. The former teams may know better but again probably by eye test. Same deal with Diaw.

Bogut's estimates strong, going plus 2, plus 3, plus 3. Pachaulia has one of the greatest variations. Dallas bought on his great plus 4.8 14-15 estimate and let him go when he only delivered 1/3rd as much last season (but plus 1.6 was still good). By 15-16 estimates Bogut delivered twice as much as Zaza but we will see what happens as they change teams.

Pau Gasol, characterized as slow and perhaps "faded" by respected media members but targeted by the Spurs. RPM suggests Pau's impact went up dramatically from plus 1 in 2013-14 to plus 2.3 in 14-15 and to plus 3.7 in 15-16. Spurs hope they are right for $15 or whatever million / yr.

By RPM Biyombo moved from -4 in 13-14 to zero last season. So did Kanter. Learning the game better? Probably. Neither special by RPM but which improves from here? Both, one or neither? Big bets. Cole Aldrich went from -1.3 in 13-14 to -3.3 then plus 3.7. Did he suddenly figure it out? Was LAC just a great low minute context? Wolves make a modest offer and will find out what happens next.

Nene not loved that much anymore. RPM marks at plus 2.1, plus 0.6 and now plus 2.4. A bargain for Rockets even at his worst mark and a steal if he maintains and you believe his best mark.

Giving weight to RPM is a choice. If you consider it credible (relative to other stuff which isn't absolutely "right" / reliable either) then you can score the GM decisions by it and identify "smart" "winners" and losers. There are other ways to view things but whether they are more or less accurate is a judgment call. A lot of media panned the Mozgov signing and by RPM it looks bad unless he gets back to his best or gets better than that. Paying Horford $25 million more for Sullinger level impacts could be quite bad too from a resource allocation perspective, especially if his impact declines as he ages as is the norm. But maybe he won't decline just yet or perhaps could get even better moving from one "genius" coach to another touted as even better. If you want to puff your chest and say you got a player commonly referred to as the second best free agent realistically available and not discuss the impact / price ratio, I guess you can do that. It kinda fits with calling a coach who has yet to win a playoff series and won less than the Pythagorean expected wins every season of his NBA career a great NBA coach or the greatest coach in the land. Believe what you wanna believe. Maybe reality will catch up with optimism. Lots of judgments to judge again in future by various means and by different folk.

I side with BPM over RPM so using that:

- Indiana made a lot of flashy moves but it doesn't add up analytically. George Hill's career rates a lot better than Teague's, and Mahinmi had a better rating season last year than either Al and Thad have had for years.

- I like the Noah risk/reward as a player who in his prime finished top ten in BPM 4 straight seasons, while he has declined last few years and is in his 30s, it's still only been 9 years since his draft which feels early to be conclusively washed up. The chance of him putting up some more 3-5 BPM seasons seem fair enough to be worth trying.

- By BPM after Lillard, the key of Portland's success is rated as the Ed Davis and Mason Plumlee combination. It doesn't look like they agree with that right now since they appear to be talking about Aminu as their main guy at PF, just signed Ezeli and Leonard at C, and after McCollum contract likely kicks in next year are putting themselves in the ugliest tax situation in the league meaning when Plumlee is a FA in 1 year and Davis in 2 years they won't be in good position to sign them if they even keep them that long. Evan Turner's deal is yeesh. Not only rates as an average player in Boston as is, but it wasn't that long ago that Turner was unplayable, his deal with Boston in 2014 was signed on September 29th because his NBA career was hanging by a thread, and he wasn't all that young then, he turned 26 a month later

- Orlando's offseason doesn't look to have gotten them that far because they gave up a good player in Oladipo to get one in Ibaka. Adding Ibaka to a team with Oladipo and the rest of their roster in tact would have been progress. Not sure what this is.