Give Counsell More Starts

[DrWood]

I don't see why Weeks and Hardy, and Hall for that matter, need to have a day off every week. Unless they're clearly fatigued or struggling mightily, Counsell should not be starting any games. I can see if Weeks or Hardy need a game off every other week or something, but each of them once a week is a little ridiculous.

They need to get a day off because they are rotten vs. RH pitching.

 

[DrWood]

This is a very minor point to an otherwise very agreeable take but it is not completely irrelevant. It does tell us that he is playing well right now and we might as well use him while he's hot since he inevitably will be so cold at some point this season that he will play to his career numbers.

 

I've said this many times before, but a batter being "hot" or "cold" is almost 100% descriptive and has almost no predictive value. It's easy to imagine a way to test the "hot bat" theory and it has in fact been done many times (the most recent and best, IMO was in "The Book". The effect was so small (something like 5 points of OBP and 10 points) that it's basically negligible. People are just noticing the random bunching of hits and non-hits. A .250 hitter isn't going to go 1 for 4 every night. I've made spreadsheets that simulate the most consistent player in the world (one who has exactly the same probability of a hit every AB) and he's insanely streaky:

 

http://rluzinski.blogspot.com/2006/05/jenkins-and-consistency.html

 

And even if a guy can legitimately and significantly, raise his expected performance level for a short period of time, how the heck would you know? It's going to look exactly like the very typical normal variance all players go through. Coin flip luck. You don't make your lineup based on that.

I know the numbers. I also know that a good bit of what goes on in baseball is psychological rather than independent random trials. Changing things helps, sometimes, because it affects the players mindset.

[Thurston Fluff]

And even if a guy can legitimately and significantly, raise his expected performance level for a short period of time, how the heck would you know? It's going to look exactly like the very typical normal variance all players go through. Coin flip luck. You don't make your lineup based on that.

 

The way you would know is by being around the person all the time seeing their approach at the plate, batting practice, ect. The thing we do not have the abiltiy to see is when someone legitimately is seeing the ball well, has his mechanics, timeing ect all in order thus giving him the best chance for his abilities to succeed. That is the only problem I have with a completely stats related approach to who plays when. It doesn't take into account a lifetime of seeing how well someone is acuatally swinging the bat, reproducing the mechanics of their swing or pitching motion ect. Just because I am not good enough to see if someone is on a good streak due to luck vs actaully playing really well doesn't mean people more adept at it can't differentiate between the two. All the stats you wish to show me does not differentiate between the two. Numbers don't have eyes and they do not see physical evidence. If you can find a study where the manager says the guy is really in a good grove and his numbers are equally random as that of when the manager says he's playing at his usual level but getting better results due to random luck then I'll concede that point. Until the I don't think it's unreasonable to think that players could be like everyone else and have ups and downs in their professional life due strictly to just being in a groove so to speak. It is equally likely that someone immersed in a field of work all their lives would be able to see when someone was really effective due to tangible visable evidence or just getting lucky.

[rluzinski]

Whenever I see managers try to ride the hot hand, they seem to do it based on results and seem to ride it until the good results stop (or rather, a few days after it stopped). I won't say that managers don't have access to physical evidence beyond the results; I just see no evidence that they can use that information to predict hot and cold streaks.