THE BOOK--Playing The Percentages In Baseball

I tried to address the selection bias as best I could. This article isn’t really about the effect of size on value, because players of all sizes are probably of pretty equal value overall, but more about where each group derives its value from. There are some sample size concerns with the lanky and pudgy groups, though they both had a good number of PAs (I’ll check the exact number when I get home).

I’m not sure I agree with you about modifying the selection criteria, but I’m not sure I disagree either. I’ll have to give it a little more thought.

In part two of the series, I’ll be addressing causation.

Of course there is going to be a large correlation AND causation between certain aspects of hitting and size/weight.  That should be obvious.  I use height and weight to determine the mean that a player’s weighted stats must be regressed towards in doing my projections.  That is critical.  For example, if a 5’9” 170 pound player hits 25 HR in his first year and a 6’2” 210 pound player does the same thing, their projections are and should be vastly different.  And it really does not matters what the causation is in terms of doing the projections.  As far as Francouer is concerned, it matters a lot that he is a big guy.  His power numbers need to be regressed toward that of big guys.  The reason why he is porbably NOT going to be a good hitter is simply that his minor league and 2006 performaces have been terrible and he obviously has no strike zone judgment.  When you watch him and see that he will swing at just about anyhting at just about any count, it helps alot in terms of regressing his performance as well.  He is a perfect example of where “scouting” (and I simply mean watching him) can help in projecting a player without a whole lot of historical PA, although including his MLE’s, he has quite a bit already, which is why it is not all that hard to project Francouer.

I am not sure where David is going with this article, but I don’t see anything particularly wrong with it so far.

Too bad Tony Gwynn just missed helping Puckett haywire the short-n-fat group. Gwynn’s 5’11”, missing the cutoff by an inch, and bb-ref lists him at 199 lbs., which seems a little charitable, at least for later-career Gwynn.

I’ve tried answering some of the questions I received (including those posted here) on my blog:

http://stats.mostvaluablenetwork.com/general/does-size-matter/

Part 2 is up:

http://www.hardballtimes.com/main/article/does-size-matter-part-2/

Very interesting look, and highly recommended.

I think it is important to note that while David used “Crawford” and “Reyes” as his illustrations, he actually is using those names as stand-ins for guys of that type.  We know far more about Reyes than the average MLBer of his body type.  Therefore, it is more accurate to say “a guy of this body type will forecast to this many HR”, or “if we knew nothing else about Jose Reyes except his height, weight, and previous HR rates, this is our forecast”.

The problem is most people don’t make this distinction.  Once they see the label “Reyes”, the immediately pollute that piece of data with a whole bunch of their own data.  Rightfully so, since you should use all the data at hand.  But, it doesn’t invalidate the general point that the article is making.

Browsing the BJ Handbook for Cubs players, I notice that F Pie is listed at 170 lbs. I had no idea he was so light. I casually looked thru the book for hitters who were active in 2007, with at least 1000 ABs, and listed at less than 180 lbs. There aren’t very many of these guys, probably around 16 or so. Most hit fewer than 10 HR per 550 ABs. Pie’s best comp in this group, interestingly enough, is perhaps C Patterson--the last great Cubs CF prospect.

The projection systems seem to be fairly up on Pie. But if he has low walks, and hits only 10-15 HR per season, then he’d better have a high BA to be a good regular--even with good defense in center. But, Pie is also not a contact guy--he has a power swing, and mediocre SZ judgement, and thus strikes out fairly often. No reason to expect a high BA from him.

So, I now suspect that Pie will be the next C Patterson--a guy with plus physical talent, but in a way that does not really come together in the game of MLB.

You make him sound like Endy Chavez, Michael Bourn, and the young Juan Pierre.  I have no idea if you are right or not.

My point however is that all these guys, if they are top fielding CF and poor hitters would be overall league average.  A good fielding CF (+1 win), with the positional adjustment (+0.5), and likely plus baserunning (+0.2), can be -1.7 as a hitter (which is horrible) and still be overall league average.

Pie was the best baserunner the Cubs had last season (not that that’s saying much), at least when I ran the numbers:

http://otherfifteen.blogspot.com/2008/03/look-at-baserunning.html#comments

Unfortunately, his defense and baserunning are probably underrated compared to his offense; doubly so since the Cubs don’t seem to understand what he brings to the table on offense. They see a speedy guy and start thinking Juan Pierre, when as a hitter he’s a lot more like Soriano in style.