A few links
Dave · February 24, 2010 at 4:27 pm · Filed Under Mariners
Here’s a cornucopia of links for you guys.
I did a Q&A with Patrick Sullivan over at The Baseball Analysts. I don’t say anything too earth-shattering that I haven’t said here, but I think it’s a fun interview.
I also did a phone interview with HotStove.com, where we spend 30 minutes talking about the M’s off-season and how different roster stuff will shake out.
And, on a slightly more random note, those of you who want to see Dan Wilson play hockey should check out the Thunderbirds game on Saturday. He’ll be competing in a celebrity all-star game that takes place after the T-Birds game. The press release is here if you want more information.
Good stuff, but I think you go a bit overboard on your jokes during the phone interviews.
I’m still reeling from the Jim Bowden video.
I’m sure everyone will enjoy this, too. It’s so rare for writers for national magazines to actually understand what guys like Dave and Derek (and for that matter, Jack and Tony) do. As he would say, glove slap, Tango (quoted in the article, too).
Thanks for the link to the SI article. Very nice summation of just how the worst run producing team in baseball won 85 games. I certainly forwarded it to my friends.
Sadly, we probably wouldn’t get to see Dan in goal though
Well, if Dominik Hasek can still play goal over in Europe at forty-something, I think Dan could probably do it at the Hockey Challenge.
That’s the most conciliatory mention of chemistry I think I’ve ever read from Dave. Love the call-out of the pre-written narrative from the beat if Bradley struggles. It’s like an ugly storm cloud on the horizon I’m just hoping a gust of wind will push the other way. If things go poorly with/for Bradley, I am probably going to have to completely avoid the beat writers as they beat the dead horse over and over and over.
Speaking of chemistry, here’s another link.
From the NY Times, a story about how teammates touching each other more frequently may have a positive effect on their teams’ success. So there are ways to quantify the effects of team chemistry!
Seriously? Someone did a report that says that good teams, which win, celebrate more than bad teams, that lose? And it’s getting published?
Academic journals. They’re the best.
To be fair, the article portrays the researchers as being aware of that possible logical fallacy, and consciously trying to avoid making inaccurate conclusions along those lines.
I don’t care what metric they used to judge performance. Of course you’re going to find a correlation between success and celebratory touching. Inferring the latter causes the former is just beyond stupid.
I see your point, and agree, but from the description of the study, I inferred the researchers were interested less in the results of “celebratory” touching, than in the way touching as a teammate-to-teammate affirmation of fraternity may affect subsequent team success.
In any case, it’s clear camaraderie (or at least a willingness to put ego aside for the common good) is far more essential for a winning basketball team than a baseball club. I just found it interesting that scientists are actually examining how team chemistry can be built (hugging?) and trying to measure its possible effects. Perhaps similar research could be done for baseball in the future.
There are is one sense in which Dave’s reaction is correct, but one in which it is wrong.
The wrong part: duh. Successful players celebrate more. I assure you, the researchers are aware of this. Correlation does not imply causality.
But there are many statistical techniques which correct for this problem. In econometrics, the problem is called “simultaneity bias” or “endogeneity” and the typical solutions are called “simultaneous equation methods” or “instrumental variables” methods. They’ve been around for decades and are designed specifically for this sort of situation, where we know that good plays ==> more celebration and more touching, but we also want to estimate the extent to which more touching ==> more good plays.
In other social sciences, Structural Equation Models (SEM) are often used. Biostatisticians have similar techniques as well as others, including inverse probability weighting.
So, to that extent Dave’s knee-jerk reaction is off-base.
However, although the NY Times article did not give enough detail about how the researchers corrected for simultaneity bias, it doesn’t sound like they used very sophisticated methodology. Moreover, even if one does use techniques such as instrumental variables, one is dependent on having additional special variables (the instrumental variables) which have to meet specific statistical criteria in order to be valid. If they don’t exist or don’t meet the criteria, then you’re stuck and cannot hope to disentangle the causality.
The article hasn’t been published yet so we don’t know the details, but I’m dubious that the researchers have adequately corrected for simultaneity bias.
So I wouldn’t be surprised if the research is faulty — but not for the reason that Dave gives. The researchers already know that there will be an automatic correlation. They have attempted to correct for it, and there are ways in which this can be done. Have they done so successfully? Can’t tell, probably not, but we don’t know yet. To dismiss the research out of hand by calling it “beyond stupid” is itself, if not beyond stupid, at a minimum ignorant.