APBRmetrics

There aren't many things in life that I rely on because they work "pretty well". Thankfully.

Did rlee's account get hacked? Weird to such those kind of subjective statements on message board like this, from a person with 1,000+ posts

Saric also ranks 10th worst in WS/48, out of 240, for players with 1,000+ minutes

rlee wrote:Why even have an RPM if Saric (legitimately 1st or 2nd in upcoming ROY voting) is "ranked" so low? Re-name it UPM. How about a requirement that any inventor of a statistical ranking system has to have seen at least one NBA game before his/her system enters the public domain? Many of the names above Dario don't even come close to passing the giggle test.

This post really bothers me, even if it was made in jest.
1) it's low rigour; Saric being low in RPM does not refute RPM
2) there are many other places where one can "attack" rpm
3) it's an unnecessary and hostile attack on an individual
4) that specific individual has contributed more to the public advancement of APBRmetrics than almost anybody else in the past decade
5) that individual has dedicated so much to helping this community and this field, in addition to being an active member of this forum, who will read it.

Among players age 19-22, Saric ranks #40 in Win Shares.
Maybe at age 28 he'll be top 40 among players age 25-28.
That would seem to put him among the top 100 in the world.

To Amepersand, Jermeias and others, I apologize. I wrote that in a fit of pique and frustration. Where I am coming from (ironically) is: I have for a long time held out hope that some form of plus/minus would ultimately be a good tool. RPM may very well represent significant progress toward that goal. Its predecessor single metrics have always seemed to me to be so flawed (in what they measure, etc) as to be, at best, silly. Plus/minus, on the other hand, at least seeks to measure impact that matters. I suppose each generation of PM gets us closer to that ultimate goal.

What troubled me most was using Dario's RPM standing as evidence against what Hinkie did. I believe it constitutes no such thing. As re: the challenge of where I would put Dario's standing among PFs, if not 76th: I'm not really into ranking players - don't see that it has much value but I came up with a #. To see if I was at least in the ball park, I tried in the most unbiased way I could to survey NBA GMs, Head and assistant coaches, scouts, retired and active players, beat writers, team broadcast analysts. From the 12 responses I have received so far after promises of anonymity (divided essentially equally among the categories): the consensus was right around what I contemplated: somewhere in the 20-22 range.

I've seen almost every Sixer game this year. I'm especially puzzled by the flippant comments here regarding his future potential. They truly seems to ignore the impact of current context he performs in.

I would be happy to break down the elements of his game that impress me from a scouting perspective should any one be interested.

Again, my apologies for my overreaction and inappropriate tone.

RPM is sorta like religion. There are folks who believe, want to believe, don't believe, don't want to believe, want to think / talk about it or don't, think there is only one way to think about it or several, think people who think differently are idiots or just hard to understand, etc.

There have been at least 9 folks who have published some form of APM employed by teams. I don't know what the ratios of believers or considerers are to agnostics or atheists. Some seem to consider it a possible tool, some "a main tool". Some prominent analysts have said they don't use it. From thoughtful evaluation and / or fear of being labeled or having to explain / defend it to employers who may be unconvinced or hostile to the idea. Almost no GM (or none) has ever taken a strong public stand. One owner has. Yet at least 9 authors employed. Sloan has touched on it occasionally on panel discussions in a far less than adequate & thorough way, yet it has published APM papers and hosted APM speakers. If Sloan or hiring GMs & Owners wanted to give APM more credibility, they could. But they don't. To protect advantage, to avoid discussing and looking shockingly primitive in their analysis or expose their uncertainty or whatever.

Not even the authors are immune to flaw or critique. Cuban moved away from one author who got exposed not being cautious and nuanced enough. But he later hired another publisher of it.

ESPN publishes it, but did Dean Oliver or Ben Alimar or anybody else in leadership ever take a stand behind it or aired the nature and extent of their cautions about it? I am not aware of anything more than brief deflections' in low visibility places.

The easiest things to say are: there is theoretical error of uncertain size (but there is error in any stat or box score metric when compared to true talent or future behavior); one can't ignore other methods; and there is a point where science stops and art, judgment, experience take over. Those statements are fine, except when they are used to avoid saying whether RPM gets or should get weight in team decision making and how much.

Crow, it is funny. I'm not a believer in APM as presently constituted/used but I see its potential as it continues to get refined. As I said, unlike its predecessors, it is at least trying to measure what I consider to be the right thing: impact on the outcome of games. You know that I have often expressed admiration for your view of things generally and specifically. Generally & re: Dario: I do trust my eyes to some extent as well as the views of a range of folks deeply involved in personnel decisions. But I certainly do not dismiss metrics. Far from it. I think there has been and continues to be much value added. In fact, my favorite quote on the subject is from Ben Alamar: " 'Not a fan of analytics' Is like saying 'no thanks, don't want the extra info my competitors are using' and using analytics does NOT mean using ONLY analytics."


I don't think it is helpful to defend using a metric to establish a point by saying of course it is a flawed metric. (I realize that wasn't you who did/said that - just responding to another part of the discussion.)

I have always tried to consider both sets of arguments for and against APM / RPM. I am for more debate and decision making than I am strickly for defense.

I and others may make the mistake of too often & simply presenting a RPM estimate and appearing to take it as the final perfect word and the only word. I don't feel that way but I also don't feel compelled to cite all the caveats every time, including when RPM is most rejected or rebuked. Almost no one recites all the caveats of imperfect / potentially misleading knowledge when citing a box score stat or video clip or scouting report. To some extent, caution & further thought is an unspoken expectation for serious students.

I'll say it again, having the 4 factors of RPM would help get over the black box rebuke and get down to the real work of comparing scouting, box score metrics and RPM factors and making judgments based on how they agree & agree, uncover truth or color it.

We don't have current RPM factors for Saric; but using the rest that is available, one could try to identify where the biggest impact debates lie. At this point it appears Saric is a well below league average efficiency shooter / scorer and that is the biggest offensive issue. Turnovers, rebounds, fouling may be more near average but aren't especially strong. Usage is elevated, rebounding is a major part of his role and he makes more assists than a generic big. Shot defense is inferred from DRPM to be below average but check the player tracking data. It may be more responsible to say: RPM estimates Saris is similar to the average rookie (-3) and does not have enough evidence yet to rate him higher or lower. Using RPM for rookies is the least reliable use of RPM. I'll try to remember that. When he is a 3rd / 4th year player there will be more & better data of all kinds and hopefully rigorous analysis & judgment.

Has anyone compiled a table showing the mean and range for RPM estimates for rookies by draft pick and the average year to year progressions (of the best, middle and worst)?

I assume an average rookie goes from -3 in year to maybe -1 by year 3 or 4 but it might help to know more detail about the transitions. Is a lottery pick who is -3 in year 1 (by the end of the season) behind the norm for that level of draft pick? Probably, but maybe not by a lot. How often do they catch up or exceed? I dunno off-hand but I'd think it useful for teams to know and consider that. If you are doing everything you can / should you probably want to know the transitions for mixes of scouting evaluations, box score metrics and RPM. What are the range & norm when they basically agree and what are they when they don't agree. And of course what did you do to manage the player year to year? Stayed the same or changed to explore potential or minimize harm or in between?

Many who became decent or better shooters were well below average as rookies. His shot isn't "broken" so I would anticipate the same kind of improvement that we have often seen, especially given his reputation for hard work.. Some have labeled his passing as "genius" level. I wouldn't necessarily go that far but he certainly has the vision. I think his level of play will shock some here next season as he will be a nice fit with Covington. Simmons, Embiid et al, assuming the last two are healthy.

The 25-30th range on RPM for PFs includes guys such as Aaron Gordon, David Lee, Mirotic and Bjelica.
25-30th on BPM for guys 6-8 to 6-10 and over 750 minutes are guys like Faried, Leuer, Booker and JaMychal Green.

The average RPM for these two groups combined is about 0.5. The average BPM is just below 0.

RPM now has Saric at -2.4 (and improve since this afternoon??) and BPM has him at -2.4 as well.

I'm certain that anyone who watched last night's game would say that Dario's play was outstanding ( the 2nd best performance of any player); yet he had a PM of minus 11.

But then again Nurkic's monster game only earned him a plus 9.

RPM is of course far different than raw plus minus.

Saric did have a big individual night. If your future role is going to be as big box score numbers guy, then big nights as a rookie are probably important to try for & get, with hopefully greater efficiency & frequency as time moves on.

In first 20 games Saric had 15 plus points 3 times, 7 plus rebounds 9 times, and 4 plus assists twice. In second 20 games, 5 such scoring nights, only 8 such rebound nights and again 2 such assist nights. Not much improvement. But then... in last 24 games, 14 such scoring nights, 12,such rebounding nights and 9 such assist nights. Corner turned? Probably. Season stats masking it. Those watching everything were aware. I don't follow Sixers except from 1,000 feet in sky. I didn't catch on to the change til now. Always more layers to dig down to & thru. I can't do everyone fully in my spare time. Argument about a player is a clue to look further though.

I need to check efficiency changes too, especial shooting / scoring efficiencies. His minutes are up 40 plus % from early til now and that explains part of the increase in "big game" stats.
3 point fg% falling from bad to awful pre-allstar game compared to post. 2 point shooting up significantly. FTs up significantly. He is evolving.

I try to save RPM estimates every 20 games or so. The rate and direction of change from period to period
should be used with caution but it could be an important clue especially for rookies. Improvement is expected for lottery picks as the RPM prior for rookies diminishes in importance relative to actual performance. I'll check that later.

Subjectively I have the feeling Saric is going to be good, but there have been examples of guys with good counting stats where the advanced ones where more predictive.

While I will agree RPM is by no means perfect, it holds up pretty well for a catch all metric.

Saric being low in RPM does not refute RPM

the san antonio spurs currently lead the league in defensive efficiency at 102.9 pts/100poss allowed. the reigning 2-time DPOY kawhi leonard - the first non-C to win the award back-to-back in some 25 years - leads the team in minutes played, has played 1/8 of the team's total minutes played...

yet real plus/minus at espn lists his DRPM (0.84) at just 118th in the league, and just 26th among all SFs...

does this refute RPM?