I listened the other day as the talking heads on ESPN discussed the idea of Troy Tulowitzki earning this year’s MVP based on his 14 homeruns in 15 game performance.
Ok, it makes for good TV and to their credit most of them came to the conclusion that if he did it again (assuming that were even possible this late in the season), he should get it. Otherwise, they wisely said no, probably not. Last year, Fangraphs had a good article on “When Samples Become Reliable”. I think it’s relevant here.
But it’s more than just sample size too. It’s choosing the range of data, too. Let’s take a hypothetical example…
Say a player, call him Player X, goes hitless on Sunday and Monday. He then gets one hit on Tuesday, Wednesday and Thursday. Friday and Saturday he goes hitless again. We could say:
- Player X got hits three days in a row.
- Player X got three hits in the week
- Player X got three hits from Sunday through Thursday
Those are all correct statements but convey very different things. It’s just human nature that most of us would like to probably choose to focus on the positive and most likely concentrate on the fact that Player X hit safely in three straight games.
So yeah, Tulo did hit 14 homeruns in 15 games and that is indeed a notable achievement. But let’s look at the whole picture, guys. He has 26 homeruns up to this point. The fact that Tulowitzki did this at this moment in the season probably didn’t hurt. Had he done this in June, the boys on ESPN wouldn’t have been bringing up his name.