Sonics hire Statistical Analyst
Dave · December 30, 2004 at 8:17 am · Filed Under Off-topic ranting
Okay, two consecutive non-Mariner related posts. I promise my next one will have to do with baseball.
The P-I has an article today on the Sonics hiring of Dean Oliver, one of the more prominant statistical analysts of the NBA. I do find it interesting how the NBA has been much more friendly to “math geek” analysts who never played the game than MLB has.
Comments
48 Responses to “Sonics hire Statistical Analyst”
Dean Oliver is author of the book “Basketball on Paper” which is a really good basketball stat analysis book, though I found it duplicated much of the work Hollinger had already done in the “Basketball Prospectus” books.
Whoa, DMZ! >’s are your friend there, bud.
Oh, just my browser.
“I do find it interesting how the NBA has been much more friendly to “math geek†analysts who never played the game than MLB has.”
It’s a good question, worthy of some pondering. I would probably start by saying that the same conditions that have made baseball more conducive to that way of thinking now than when James first started writing exist in basketball: Guys with financial backgrounds (like Wally Walker) running/owning teams; the success of baseball teams with statistics; Moneyball, etc.
At the same time, I think part of it is that stat guys are less threatening in basketball because the sport is so much more difficult to quantify than baseball. That weakness is, in some ways, actually a strength … I think basketball guys tend to be more accepting of traditional views, perhaps.
While there’s not a lot of denigration by the NBA, there hasn’t been a lot of money either. Oliver is the only guy making a living (and a paltry one at that) off working for an NBA team in this capacity, and Hollinger is the only person I can think of making a living off of writing about NBA stats … I guess you can throw Roland Beech of 82games.com in there as well, though he makes most of his money on the football side.
Of course that makes about two and a half more guys than were making a living on this four years ago. …
OT: Why no post yet on Randy going over to the dark side?
Perhaps because he hasn’t been traded yet, so attempting analysis of the players involved would be speculative at best, and largely futile.
I think Kevin’s point about the difficulty of quantifying basketball stats makes a lot of sense. Additionally, or perhaps alternatively stating the same thing, I would think (I haven’t read Oliver’s book or Basketball Prospectus), both stats guys and old-school minds like Sund probably recognize that due to the greater team involvement in basketball, things like “chemistry” and “feel” really do play a greater role. Whereas it is easy for us to determine that it is idiotic for Winn to bunt Ichiro to second base when the Mariners are down 3-1 in the sixth inning and have nobody out, and we can immediately tell if Winn properly executed the bunt or not, and whether the Mariners scored (and can quantify the chances of more runs having scored, etc.), it is harder to quantify, for instance, the wisdom of a baseline pass being stolen by the other team when the Sonics attempt to go inside to Rashard rather than keep him out on the perimeter. Since the in-game execution would seem to have more variables, which at least seem much more difficult to quantify, coaches would feel less threatened and less second-guessed (notwithstanding criticism of putting certain lineups on the floor, running certain plays, etc., but even those decisions are not, at least to my knowledge, subject to quantification that sportswriters, casual fans, and the like are typically aware of).
Once upon the time, I remember reading the Mariners were looking for a ‘stat geek.’ Did they ever hire someone?
Part of it, too, I think, is that Bill James, who brought statistical analysis to the nation’s consciousness in the 1980s, was fairly obnoxious about it and probably alienated insiders as much as he attracted outsiders to his important and groundbreakingt work … to wit, he made it clear in numerous “Baseball Abstract” essays over the years that he thought the tools he has created, uncovered and worked with were superior measuring sticks of performance to anything being used in any major league front office at the time.
He wasn’t a bit shy about naming front-office execs (Gabe Paul, Buzzie Bavasi, Dan O’Brien, Dick Wagner, early John Schuerholz, etc.) he thought were morons. He had no hesitation about identifying iconic baseball players he thought were ruining their teams (Pete Rose, Enos Cabell, etc.) in singularly acidic style.
It almost doesn’t matter that James was largely right (and continues to be, as a senior adviser to the current World Champions) … the parallel legacy of those “Abstract” blastoffs — next to inspiring the current generation of performance alanlysts — was to alienate virtually everybody INSIDE major league baseball and create a bunker mentality against truth as an insitinctive defensive response.
That legacy has carried on today to some extent, thanks to the perceived smugness of “Moneyball” and the frankly snide scribings of the Baseball Prospectus crew, among others. (I’m not saying they’re wrong to take the tone they do … only that it makes few friends in front offices to do so.)
Even though performance analysis makes incremental progress in baseball with each passing season, there are still many front offices populated by defensive, offended guardians of tradition and history, who will stubbornly adhere to anachronistic strategies as a hardheaded response to all the cheap-seaters calling them doofuses for doing so.
As far as I’ve been able to tell, that hasn’t been the case in basketball or football … hence, in my view, the reason for their comparative openness to performance analysis.
Funny how KP has this conversation spill over to a baseball site:)
As I said elsewhere, it’s tougher to quantify basketball, especially on the defensive end, as it’s more of a team effort to play defense where players are double-teaming, playing help defense and have to defend based more on who’s on the floor with you.
That is to say, when Tim Duncan or Ben Wallace is behind you on defense, you’re going to play better on-ball defense because of the risk you can take knowing the lane is protected. You can’t really quantify the quality of defense based on those factors in an individual context. You probably can do it in a team context, or possibly an addition by subtraction using +/- statistics though.
Yes.
You should read the books and see what these guys are doing before you say it can’t be quantified. They’re doing good work.
Or, to put it another way, basketball is a much more analog game than is baseball. It’s far easier to quantify a game that is a series of mostly discrete events than one that tends to be continuous.
“You should read the books and see what these guys are doing before you say it can’t be quantified. They’re doing good work.”
Maybe I should read the books.. but I’ll qualify it with saying it can’t be properly quantified in my opinion based on suggestions I’ve read. I’m certainly open to finding out new methods to which they approach it.
I just find it hard to judge an individual in a team sport where the players he’s on the floor with and the system of defense being played can affect how much the man he’s guarding scores.
Then read the books. Re-iteration of your opinion that a mountain probably can’t be climbed while there are people waving to you from the summit doesn’t do any good.
I’m simply searching for discussion and an idea of what these methods are while I’m here. I’ll go back to my cave now, thanks.
The way that statistics are presented in mainstream NBA sources (for that matter, in just about any publication with NBA stats) has always led me to believe that the prevailing methods are ridiculous, more ridiculous than baseball stats have been since the early 20th century. It’s hard to see how anyone has paid the slightest attention to analysis when the rate stats are still presented per-game rather than per-minute. It’s like putting a speedometer in a car that measures miles per day.
Jim
You, too, should check out these fine books, as they offer innovations like per-possession effectiveness.
DMZ-
Who’d they hire? Or is this a state secret, like how “$92 million” becomes the official Mariner budget figure?
I find basketball profoundly boring, so I have little more than glimpses from the occasional game on the TV in the bar or sportscenter highlights to go on, but my impression is that there are fewer “old guys” hanging around the NBA teams. The owners seem to be younger (though the only ones I could name are Schulz and Allen) the coaches seem to be younger, and I’m guessing the GMs are younger too (or were prior to the latest Beane-led youth movement). There certainly don’t seem to be Don Zimmers and George Steinbrenners tottering about the courts (though Lennie Wilkins looks like he’s getting up there). Perhaps that, combined with the more hip-hop, wide-open dynamic of a game less burdened by its own self-importance as “America’s Pastime” means people are less worried about guarding sacred traditions and more open to new ideas.
ESPN reports that Randy has been traded to the Yankees for “Vazquez, the deal would include catching prospect Dioneer Navarro, at least one other prospect and about $8 million going to the Diamondbacks.”
Re 21:
I believe it when I see Johnson in a Yankee Uniform. This is the 10th or 11th time the trade has been completed.
Somebody should create a pseudostats award–they are proliferating almost as quickly as real stats. (Can I be hires as a pseudostats geek?).
Examples would be:
*”Productive outs” – (I propose tracking the antithesis “destructive total bases”)
*The IBM NBA player rankings, which almost always would pick someone absurd like David Robinson or Dikembe Mutombo as the MVP every year.
*Rolaids Relief Man
*Wall Street Journal’s MAX rankings (“we use a computer”) which always try to find the nation’s best college teams and usually pick some team like Syracuse in football and Stanford in basketball.
*and, of course, the BCS
Aside from the BCS, what’s weird about these “stats” is that what instantly tells you they’re wrong is usually your gut…
All Hail the BCS.
Praise the BCS.
The BCS knows all.
The BCS is in good shape.
All Hail the BCS.
You can’t spell “BCS” without “BS.”
BCS is keeping Boise State, Louisville out of their rightful spots in the big bowls. Why the heck should PITTSBURGH get a shot?!?! In fact, since Miami (FL) and Virginia Tech left the BCS, they got nothin’, so the Big East shouldn’t even get a guaranteed spot.
As much as I’d love everybody to purchase the Prospectii/Forecast and BOP, if you want a taste of what Hollinger and Oliver are doing without running out to the bookstore, it’s possible via our friend the Interweb.
Much of Oliver’s work is at his website:
http://www.rawbw.com/~deano/
while Hollinger has some old columns and links to his work at CNNSI.com on his:
http://www.alleyoop.com
I would say the problem with the BCS in the context of this discussion is that not even the most radical statheads are suggesting that computers be allowed to make personnel and on-field decisions. The human element, fallible and faulty though it may be, must still have the final word. The people who have let the BCS take over college football decision-making have decided that people aren’t good enough. And that sort of I-Robot thinking is not only wrong but dangerous when taken to its logical extreme. Computer-generated data and analysis should always reinforce human thinking, never the other way around.
Eh, it’s all moot. The BCS should be theoretically gone at the end of this year, with the AP flat dropping out, right?
DMZ, you have any specific book references for me to read in regards to basketball statistical analysis, that can be read by a layman with only vague recollections of how to do complex statistics.
I know that there are statistics to be found in the game, but tell Al McGuire (though deceased) or any great “feel” coach and they will agree that statistics pale in comparison to “reads.” I can fall for a statistical belief in baseball far easier, it is a more simplistic game in its movements.
Basketball has far more contributing factors. Constant flow that leads to reads and reactions that happen at a far greater rate than on the diamond. And while you can create equations for parts of the mix, you’ll never get an equation that also adds in the human factor.
I just don’t see it happening. But i am smart enough (and want an edge over the competition) to take a deeper look at the statistical aspects of the game.
Oh.. and on a somewhat connected sidenote– my aunt bought me “Curve Ball” by Jim Albert and Jay Bennett for Christmas. Where does it fit on the stat-world listing of “great books?”
tyler and others looking for more on statistical analysis in basketball, take a look at this page … it’s got links to Hollinger’s books (DMZ having already posted the one to purchase Basketball on Paper) as well as Hollinger’s and Oliver’s Web sites, amongst many others, so you can get a smaller taste of what they’re doing if you’re not quite ready to run out to the bookstore.
I think I’ll do some pioneering performance analysis in the game of Scrabble. My first project: Developing a metric for “VORT” — value over replacement tile.
Tile value is too situational. Early in the game, that J or X can do great things for you, but at the end when there’s only one tile left, you’d generally prefer it be an N or A.
I don’t think basketball and football are sufficiently interesting sports to want to analyse. Hockey, on the other hand…
And I thought I had too much time on my hands…
You think you have too much time now, just wait until your gf gets back and reads the regifting thread…
She’s back. But I “borrowed” her computer, at which time I put a nice block on this website. She’ll never know. All she’ll get is a 403 Forbidden.
“DMZ- Who’d they hire?”
I would sure appreciate knowing this, too.
people are less worried about guarding sacred traditions and more open to new ideas.
I was surprised that I didn’t see more of this in response to the question of why the NBA is friendlier to math geeks. I just don’t think the NBA is anywhere near as threatened by something that pounds at the walls of “tradition.” In the NBA, tradition is sneered at. In MLB, it is sacred.
I also agree that someone like Paul Allen or Mark Cuban would be far more likely to embrace statistical analysis than would George Steinbrenner.
anyone know of a good sonics blog?
So, it’s a secret who the M’s hired?
While it may be true that Bill James’s aggressive and blunt style might have alienated some baseball insiders, I think it’s at least as true that he never would have got an audience at all without that style. We read Bill James because he was entertaining as well as enlightening. I think the same thing is true of Baseball Prospectus. I’m not saying anyone needs to put a premium on obnoxiousness; but a lively style is necessarily going to put some noses out of joint, and a polite, no-toes-treaded style will likely put people to sleep.
Maybe the stat geek the M’s hired is… DMZ? 😀
I have no idea whatsoever, I’m just stirring the pot.
‘Stat analyst’, I mean.
So here’s the thing. Until he’s announced, or there’s some kind of nod, I’m not going to say anything. This is the curse of inside knowledge –sometimes it’s all frustration.
I will only say this: it is certainly not one of us three.
Is he likely to be announced in the foreseeable future? Just curious.
And by the way, it’s a shame that it’s not one of you three. One would think *some* progressive major-league front office would’ve taken an interest in you guys by now.
Re Adam’s comment in #23 about “productive outs” and “destructive total bases”… classic! 😀 Fortunately the M’s should be “destroying” themselves somewhat more in 2005. I’ll be cheering for every devastating home run.
Wasn’t it BoMel who said that a home run wasn’t always good, because it could suck the momentum out of a rally? I guess now he’ll have Troy Glaus practising sac bunts and ground balls to the right side.
I think James can be plenty entertaining when he’s not being critical, but I also question how much being critical had to do with James getting noticed.
Having admittedly not read any of the Abstracts, my understanding is that James started out with the assumption that people just hadn’t bothered to think about the kinds of things he was thinking about and only later realized they just didn’t care and changed his style.
It’s worth noting that James never wanted to work for a team, whereas Oliver did. Criticizing people ain’t the best way to get them to give you a job.