THE BOOK--Playing The Percentages In Baseball

Friday, September 04, 2009

Is it important to be a “non-results-oriented” person?

I was watching Pedro pitch the other night against SF.  Now granted, SF does not have a very good hitting team (not as bad as the current Mets) or perhaps even PIT, but Pedro allowed one run on a HR and struck out 9 in 7 innings.  By those numbers he pitched “brilliantly.”

Now, I am an extremely non-results oriented person.  Part of that is from my background in poker and blackjack.  For those professions or avocations, you have to completely divorce yourself from your results, especially in the short run, at least as much as you can.

Anyway, when I watch a pitcher in a game, I pay no attention to the score or even the game situation, other than how the pitcher pitches given the score and the game situation (inning, outs, runners, etc.).  That sounds a little like a contradiction, but basically I am evaluating the pitcher independent of the results of each PA and ultimately the score.  That is not 100% true.  The results of a pitch or a PA can give you some idea as to the effectiveness of the pitches, but that can be dangerous as well, for obvious reasons.

Basically, I am looking at pitch selection, command and location, stuff, and things like that.  Again, some of the results of each pitch and each PA can give you a clue as to the effectiveness of each type of pitch, but there is lots of fluctuation in those results so one has to be careful focusing on it.

Now I am not pretending that I can tell you from one outing what kind of pitcher Pedro is, but my assessment of him last night is not nearly as positive as the results (one run and 9 K) would suggest.  I saw decent but not great stuff in an 89-91 mph fastball and change ups which were decent although he missed his location with them a lot.  I didn’t see many other pitches, although I am not necessarily recalling too well right now.  His command seemed to be so-so and he appears to be trying to throw a lot harder such that his command is going to suffer (much like Carmona who appears to try and throw as hard as he can and has no idea where the ball is going by and large).  Maybe at this stage in his career, Pedro has to reach back and fire away in order to reach 89 or 90 in the gun, and consequently his command is going to suffer.

Anyway, it got me thinking that in order for a scout to be good at evaluating players, I think he really needs to be divorced from the results by and large, especially the results of the game. If I were hiring scouts, I would look for that - non results oriented people, at least with regard to scouting and player evaluation.  I’m talking observational evaluation of course.

I might even give them a test.  Take the same pitchers, mask their identity and let them watch 2 games for each pitcher.  In one game, their results are terrible, and in the other their results are great.  I would make sure that they appeared to have around the same stuff, command etc. in both games.  Then I would ask them to rate each pitcher in each game.  I would even tell them to not worry about how they did, game-wise, unless they thought that would help them in their evaluation.  I would be willing to bet that for some of those scouts, maybe even most of them, the results of the game would significantly affect their evaluation.  I would also bet that you could, theoretically, find some scouts who would be able to evaluate the pitchers mostly independent of the results and that they would be way more accurate in their evaluations.

Another experiment I would do is this: I would show them videos of 10 pitchers who had bad games and 10 pitchers who had good games.  Again, I would mask the identities of the players.  The 10 pitchers who had bad outings would be good pitchers of course and the 10 pitchers who had good outings would be bad pitchers.  I would ask them to rate each pitcher.  Again, I think you would get some interesting results…