how many pitcher vs. batter at bats are significant?

[sam h]

The quality of the individual batters and pitchers, and their righty/lefty splits, are obviously more important. So for practical purposes this doesn't matter much - ie, it shouldn't be a decisive factor very often in making pinch-hitting or bullpen decisions.

However, that doesn't mean that you can't tell anything from small samples of data. It doesn't have to be "statistically signficant" for you to draw an inference. It's just not that useful except maybe in very extreme cases.

[Jack of Arcades]

[ QUOTE ]
[ QUOTE ]
A lot more than 24.

[/ QUOTE ] nonsense. anyone who played baseball will tell you that if they faced a guy 24 times, they knew where they stood against him, and i think more often then not this will be reflected in 24 AB's

[/ QUOTE ]

But that doesn't mean the statistics will reflect that.

[sam h]

[ QUOTE ]
But that doesn't mean the statistics will reflect that.

[/ QUOTE ]

You will not be able to rule out the possibility that deviations from some standard (say the guy's BA overall) are not the product of random chance. But that doesn't mean the data are necessarily useless.

[andyfox]

I don't think it's meaningless. I got a guy who's 2 for 44 against somebody, I'll give him the day off the next time he's pitching.

[andyfox]

Damn straight.

[Sooga]

[ QUOTE ]
[ QUOTE ]
0-24 against a variety of different pitchers I'd imagine is a lot easier to do than 0-24 against a single pitcher. Plus Jones is only a .260something hitter. 0-24 isn't impossible with that kind of an average.

[/ QUOTE ]

This makes absolutely zero mathematical sense. 0-24 should actually be much more common against a particular pitcher, because there are particular pitchers that extremely difficult to hit. But obviously 0-24 against various pitchers happens more than 0-24 against a single pitcher, because there are so many strings of 24 consecutive at bats.

[/ QUOTE ]

Yes, but pitchers that are extremely hard to hit (i.e. those you could easily go 0-24 against) are few and far between. The whole reason for the opposing team to change their pitcher during a game is to try to lessen your chance of getting a hit as much as possible, so I don't think it's unreasonable to think you could go 0-24 through a particularly rough stretch of pitcher matchups. And even though a hitter rarely faces any one pitcher 24+ at bats, i just have a feeling that over history, 0-24 slides over a span of several games happen at a higher rate than 0-24 slides against a specific pitcher. Perhaps it has zero mathematical basis, but it's just a feeling.

This isnt poker, its baseball. Pitchers have different styles of pitching that make it much more difficult or easy for a hitter to hit the ball well. 24 AB's against a pitcher is a big enough sample size. The room for chance is much less in baseball, a pitcher is generally capable of doing x when he wants and a hitter is generally capable of doing y when presented with a certain pitch

[Triumph36]

I was thinking more about this today - I can see it now in our sabermetric future:

"Coach, why am I out of the lineup?"
"You're 0 for 24 against this guy."
"But my stats haven't even regressed to the mean yet!"

24 at-bats is not long enough for we, the observers of statistics, to make a judgement. However, it's plenty of time for batters and managers to make a decision, and this is where statistical analyses break down - we are not dealing with randomized events; a deck of cards, the location of an electron - we are dealing with people. Batters are aware of who they can't hit, and pitchers are aware of guys who can't hit them. Saying these things don't have any effect means one of two things - it's unquantifiable, so it's unimportant, or all that stuff cancels out in the end so it doesn't show up in the statistics. I have a hard time believing either position.