THE BOOK--Playing The Percentages In Baseball

The other thing that I think it is missing (and I am not sure how big a deal it actually is), is that this study is treating all hits the same.  Moving a third baseman from his normal position to the right side of the infield may turn Utley’s would-be singles into outs, however the would-be outs or singles hit to where the third baseman *used* to be are now turning into doubles and triples.  What I am trying to get at is that Max’s analysis is focused on lowering lowering batting average, but at what cost to slugging percentage?

Still, great, great work.

Adam, agreed. 

In addition to that, it presumes that the player will not try to hit the other way (or even bunt the other way).

The better metric to choose, over BABIP, is a wOBA-based BABIP, where the numerator is:
0.9*(1B+RBOE+FCSA)+1.3*(2B+3B)
RBOE is reaching base on error
FCSA is fielder’s choice, with batter and runner safe

The denominator is all balls in play.

Thank you for the reviews on my first posts.
I completely agree on your observations.

For different purposes I had looked at where infielders usually play, but I’m not sure about gameday coordinates: can anyone tell where is home plate located in the gameday diagram? (and first, second and third base bags too...)
I’m using the coordinates suggested by Adler in Baseball Hacks, but I’m not sure they are 100% correct.

Max,
Approximate coordinates for the bases are as follows:
1B: (x=151,y=176)
2B: (125.5,152)
3B: (100,176)
HP: (125.5,203)

There are a few time frames in a few parks, mostly in 2005 and 2006 and few in 2007, where different coordinates were used.  There are some comments to a thread at StatSpeak that address this issue.  I’ll see if I can find that.

It was Brian Cartwright’s article:
http://statspeak.net/2009/01/defending-mannys-defense.html

Max and Mike - Home Plate coordinates vary from field to field and also for pre-2008 and 2008 data. I have normalized the coordinates and I am sending the numbers to be published in the Hardball Times.

Excellent news, Peter!  I look forward to it.

Nice work, obviously I’m a proponent of tracking the precise location of batted balls

A comment, though: I don’t think a fielder in short right field will work for a runner like Utley.  Balls fielded out there that can be thrown to first ahead of David Ortiz and Travis Hafner will arrive too late to get Utley. I haven’t looked at how many bloops he hits to short RF, but probably not very many.

Utley’s speed also makes the abandoned 3B position hurt more, as Utley on first (or second, or third) is more valuable to the offense than Ortiz on base.

Tango, I think your estimate of where the first baseman needs to be positioned is somewhat faulty.

First, even if your guess that a typical 1B runs the 100m in 15 seconds is correct, that does NOT mean that they run 20m in 3 seconds or that their top speed is 6.67m/s.

Second, on slow gounders (i.e. balls the 1B has to wait on), shouldn’t the pitcher cover first base?

Third, the batter is 90’ away, even if the 1B also has to stop and turn (to recive the throw) does that realy equate to 60’ of distance?

I don’t think this should be to much of a concern actually as long as your talking about shifts towards the first base line.

If a team wanted to shift towards the third base side that’s an entirely different matter, that would require the 3B or SS to throw “blind” accross the diamond to a spot and for the pitcher to catch said ball in stride and while not impossible I don’t think that this is practical.

I think that typically the pitcher, starting 60-ish feet away and falling of the mound usually CAN beat the hitter to the first base bag, but clearly does not have time to stop and turn.

Bjorn: I was trying to do is very quick.  Obviously there is start/stop considerations, which I noted.  Which is why I revised the 20m down to 10m.

Indeed, your point about the pitcher is great: the pitcher is about 20m from 1B.  So, that would be about as far as you would want you 1B to be positioned if he didn’t have to field a ball.

And so, having the 1B positioned 10m from 1B, to handle the start/stop and fielding the ball seems about right.

I don’t understand the problem.  I’m taking a WAG that the 1B can be shifted up to one-third of the way toward 2B.  Are you saying that’s too much or too little?  Or do you simply have an issue with the process of my educated guess?

I didn’t realize Utley was such a pull hitter. This may be just me, but it seems like he’s talented enough to hit the ball the other way enough to make them stop shifting (if they shift in the first place).

Tango, when I originally read it I interpreted your “guesstimate” as substantially more than a “WAG”. After reading it again I realise that I must have missinterpreted it and this made me come on to strong and for that I appologize.

My issue was more that the process of calculating it was very crude, but as mentioned above I now realise that this was probably intentional.

If (as I guess) your main point is that one should use a “constraint distance” on where the 1B can line up and not so much the exact value of the distance then I fully agree.

Presumably the distance is also quite situational, depending for instance on the speed, handedness, and “quick start ability” of the batter. And also obviously faster, more athletic 1B can play furter away vs ALL batters.