APBRmetrics

xkonk wrote:

Crow wrote:This could be considered insensitive or offensive or just not interesting to some but I’ll throw this topic out there to see if any one else has opinions to share.

I saw this old article.
http://sports.espn.go.com/espn/columns/story?id=596975

Has there ever been a male vs female game....

That page won't load anything for me, so I can't speak to any of the content. But there's a 2007 book called Andy Roddick Beat Me with a Frying Pan that has a chapter on gender and sports. It starts with an anecdote about Artie Lange almost beating a professional women's basketball player 15 years his junior in a one-on-one game to 15 (he was up 14-10 but lost). Lange never played competitively past grade school. The author also beat a different pro 21-14. In terms of team play, it says that women's college basketball teams (such as University of Tennessee, so good teams) often scrimmage against men's IM teams (where the men obviously are not DI quality) and the games are competitive. Sonny Vaccaro is quoted as saying that a WNBA all-star team could be competitive at the low DI level or against an all-star team of the country's best 14-15 men. The chapter looks at a variety of sports and notes that even in games where physical ability shouldn't be an issue, like pool and darts, the gap shrinks but the best men are still better than the best women.

Thanks for the feedback. The link worked for me when I posted it. The link was changed apparently (?) http://sports.espn.go.com/espn/columns/story?id=5969758

Bobbofitos wrote:

xkonk wrote:

Crow wrote:This could be considered insensitive or offensive or just not interesting to some but I’ll throw this topic out there to see if any one else has opinions to share.

I saw this old article.
http://sports.espn.go.com/espn/columns/story?id=596975

Has there ever been a male vs female game....

That page won't load anything for me, so I can't speak to any of the content. But there's a 2007 book called Andy Roddick Beat Me with a Frying Pan that has a chapter on gender and sports. It starts with an anecdote about Artie Lange almost beating a professional women's basketball player 15 years his junior in a one-on-one game to 15 (he was up 14-10 but lost). Lange never played competitively past grade school. The author also beat a different pro 21-14. In terms of team play, it says that women's college basketball teams (such as University of Tennessee, so good teams) often scrimmage against men's IM teams (where the men obviously are not DI quality) and the games are competitive. Sonny Vaccaro is quoted as saying that a WNBA all-star team could be competitive at the low DI level or against an all-star team of the country's best 14-15 men. The chapter looks at a variety of sports and notes that even in games where physical ability shouldn't be an issue, like pool and darts, the gap shrinks but the best men are still better than the best women.

A friend of mine played on the Duke's woman's team (as a practice player). He said they routinely won - sometimes they lost, but basically a collection of guys with HS experience were favorites over one of the best women's programs in the country. This sorta jives with my own experience - any D1 men's basketball team would be prohibitive favorites against any woman's team.

That is why I think if there ever is a contest and it is to be "fair" the average heights & weights would have to kept within 10-15% and it would probably have be the D3 male talent or even good high school players to be close.

I forgot about the college or WNBA women vs modest talent male practice players scenario. That may be the only place the contest ever takes place.

With so much of even fairly simple analytics tuned out by or ignored or only modestly applied by players and coaches does the big data movement make any sense anytime soon for any directly on the court application? Should big data focus mostly on information for executives and owners for the decisions they make? If so, will coaches be any less upset and hostile if this information is used separate from them than when analytic practitioners and advocates try to involve them it is development and use?
Poo-pooing analytics probably shouldn’t make it go away or radically reduce its ambition. At least I don’t think it should. At some point folks who truly understand and believe in analytic products are probably going to force its use by order or take over the parts of the coaching job where its application occurs (lineup, rotation, shot distribution, maybe playcalling, in game adjustments). Maybe the traditional coach should mainly be a teacher of the fundamentals and a teacher of what else he is told to teach, off the court. Maybe some teams should try have one of their best overall qualified analytic staff members take a few turns as head coach in the pre-season or take a sabbatical to go do it full-time in the NBDL. Instead of hiring a low to mid-level assistant for a D league (probably unpaid), put one of your main long-term analysts in the head coach job. Go big or do you really learn that much? Game coaching is a different role that requires experience but I am not willing to completely concede without trial that traditional path coaches are clearly better than those with analytic experience paths or the only ones who can perform such a role. Although it might fall apart I am not willing to categorically say that is always best for one guy to perform all the current responsibilities of head coaches. By the way who would be considered the most non-traditional coach in NBA history? Phil Jackson might the most non-traditional traditional coach but have there been any non-player, more non-traditional head coaches?


I saw an article about the conference where John Hollinger said the Grizzlies’ analytics department when he arrived was basically his laptop. I didn’t hear the statement and don’t know the tone or intention but taking it at face value it makes it seem that either the Grizzlies didn’t get or use a lot from their previous analytic employee or consultant or they chucked it. I would find it hard to believe that they or anyone with a reputed analytics effort didn’t produce tens or hundreds of gigs of data and a file cabinet or a wall of them with analytic findings and reports. But maybe they didn’t or maybe parts of the organization didn’t while others (ownership?) did. If every team now has an analytic department what standard needs to be met to be considered a serious department? A utilized department? If a team does not have analysts producing reports of comparable sophistication of Sloan conference research papers are they really a strong analytic department? Can you still be a significant analytic effort with less than expert level technique? Is anything less than elite PhD level talent short of what the task calls for or what elite handling of the task calls for? How many of the analytic departments are at that level? Is it only about 20-35%? If the percentage is that low, how much deference should teams get for their efforts from outsiders? How many analytic departments are yielding impacts that one would expect from elite PhD level talent involvement? How many of the analysts who have moved on from teams essentially feel disappointed with the level of utilization of their work? How many of them are more right to feel that way than the people in the organizations who didn’t utilize their work product that much?


Only about 1/3 of the coaches in the RAPM study have coached 2+ teams by my quick count.


Better pro prospect TJ Warren or Rodney Hood? I have seen some contradiction in the draft evaluation I’ve seen (scouting based and number-based). I haven’t seen them play, so I am asking.


Will the Sloan conference ever have a panel exclusively devoted to the production, utilization, efficacy and issues with adjusted plus minus (perhaps across sports)?

As of a week ago, there are 32 players whose RAPM estimate and SPM estimate (at JE's site) were 4 or more points different. 2/3rds bench players. The 12 highest differences were all higher on RAPM and also for 30 of these 32 with high variances.


Here are the names of the high RAPM- SPM variance players:

More than 6

Patrick Beverley
Channing Frye
Avery Bradley

More than 4

Derek Fisher
C.J. Miles
Patrick Mills
Draymond Green
Vince Carter
Andre Iguodala
J.R. Smith
Derrick Rose
Iman Shumpert
John Henson
Evan Fournier
Jeff Pendergraph
Kirk Hinrich
J.J. Hickson
Landry Fields
Ryan Hollins
Tony Parker
Mike Conley
A.J. Price
Manu Ginobili
Dion Waiters
Taj Gibson
Chuck Hayes
DeMarre Carroll
Jose Calderon
Amir Johnson
Chris Andersen
Danny Granger
Matt Bonner

Overall the absolute differences were

2-4 points for 127 players
1-2 pts for 151 players
and 0-1 for 145 players.

This is not necessarily an error for either metric as they measure different things (SPM lacks shot defense).

Of the 28 2+ worse on RAPM than SPM it is about half and half interior and perimeter players and about half on offense and defense on average.

Of those 4 or more higher on RAPM than SPM about 2/3rds are perimeter, 1/3rd interior.



Calderon underappreciated? I heard that in an article I read recently but his RAPM is over 4 points worse than his SPM. Is he actually over-appreciated generally or at least by folks who go by boxscore based stats?

Has anyone studied analyze shooting efficiency as it relates to player and ball speed and rate of change of speed? I see so many layup attempts where it seems the player is going too fast and / or releases the ball too hard. I wonder if player and ball deceleration to some degree is coorelated with finishing efficiency to any degree. I'd think that trying to slow down a little at the very end could be helpful but would want data to look at.

I looked at a version of James Broccato's player projections fromDec. '13 and found 27 players with a +1 projectionor better inthe last 6 drafts or 4.3 per draft. That is pretty low for "impact" players even to a modest +1 level.

Scales aren't necessarily the same but I found about 100 +1 or better RAPM estimates on JE's site for this season with an average stimate for this group of about +2.4 Only about 40 at or above +1 on his ASPM. It looks like a large part of the additional value on RAPM is on average coming from additional defensive impact beyond the boxscore stats.

Whether one tries to build / improve by the draft or trades or free agency, the supply ofpositive impact playerrs is limited.

The +1 players on Broccato's draftee projections were pretty even distributed among PGs, wings and bigs and averaged less than +2.

Of the players in the league this season with +1 or better estimates on RAPM, on average they were almost right at neutral on ASPM for offense and defense on average. On average the boost they gave the team was coming outside the boxscore based ASPM. Food for thought about the need to look at both and maybe which to emphasize? Of course part of the difference between RAPM and ASPM may be RAPM "noise" but part may also be error in how ASPM estimates value at play level for players as opposed to the value derived for league averqage value of those actions.

I think there is very little chance Phil Jackson gets the Knicks a title in next 5 seasons. I think Steve Kerr might be an average or a bit better coach but I doubt he is the right / best choice. I don't think many players are going to make playing for Phil / Steve / the Knicks a priority. I think most or all GMs will be very reticent about helping them, especially Phil with his $12 million salary and perhaps his wiggle room about how much he has to be there and do the heavy work. The Knicks got there PR assistance and Phil will get a ton of money. Will they even get to a conference finals? I'd guess more likely no than yes.

The idea of a 4 point shot was apparently at least briefly considered by the NBA. Seems dead though.
If they at least considered that, I still think they should consider 3 points for a dunk to further stretch the floor between the long ball and the inside shot. 4 on 4 still seems to me to the best way to create more space and flow, even over a bigger court which seems unlikely. The court is just too crowded with people doing little to nothing at any given moment and offensive players simply trying to get themselves and their defender out of the way of the ball holder. If the NBA were serious about possible major changes they should try them for at least 10 games in the NBDL.

"3 points for a dunk"

Great idea. Lots of replay reviews to see if it was really a "dunk".
Others have suggested the equally great idea that a dunk should be worth just 1 point.

IMO, 3 pts for a dunk is a good idea to consider to balance off an over-valued 24+ foot jump shot (gets 50% more credit than any other shot when it is only slightly tougher to make on average than a 15 footer or a 5 footer for that matter.). It would return the game to a better balance of power / speed and shooting skill. 1 pt for dunk is a completely unfair concept and I have not heard anyone seriously suggest it. I can see almost no need ever for a replay to see if the shot was a dunk.