I don’t think BvP is an illusion. I REPEAT: I don’t think BvP is an illusion.

This whole Batter vs. Pitcher thing has gotten out of control. It’s almost become a Democrat vs. Republican type of thing where the two sides feel like they need to buy into a whole plethora of tangential beliefs that come with their party’s central viewpoint. I’m pretty sure the correlation between BvP and streaky play supporters is stronger than that between liberals and those who like their steak cooked rare (I have no idea what the numbers are but I’ve had this theory for a long time that there’s a very strong correlation between how conservative you are and how charred you want your steak to be).

Let me start by saying that I think, in many cases, we aren’t arguing over the same thing when it comes to Batter vs Pitcher stats. The BvP supporters out there tend to argue that “of course certain hitters hit the ball well against specific pitchers” (even after accounting for their relative talent levels). That’s usually followed up with something like “if you’ve ever played baseball, you know you see the ball well out of certain pitchers’ hands” or “I talked to Vernon Wells and he confirmed BvP is real.”

Here I am using all these fancy numbers when I could just be listening to Vernon Wells.

But here’s the thing…I DON’T THINK BvP IS AN ILLUSION!

Kind of.

I think those who argue that certain batters should crush certain pitchers more so than others, even over the long run, are correct. Certainly there are batters who “own” pitchers or see the ball well or match up well in terms of pitch types or batted ball profiles or whatever. I think arguing against that is actually pretty silly. So in that sense, I do think BvP is “real.”

But again, I don’t think most of the anti-BvP crowd is arguing against that idea. Well, some of them are—some of them are taking the lack of predictive power of BvP and extending it to claim that BvP is completely illusory (not just the statistical value of BvP, but whether it “exists” in reality)—and I think that’s pretty stubborn. Absence of evidence is not the same as evidence of absence. But for the most part, the data-driven crowd that tends to reject BvP is saying “I just think this isn’t really useful to help me make better predictions.”

And for the most part, I’d agree with that. It comes down to sample size. When we’re looking at a batter who is 5-for-10 against a particular pitcher, we’re seeing a very small glimpse into “reality.” There’s a chance the hitter really “owns” the pitcher, but there’s a much better chance that his .500 batting average over 10 at-bats is due mostly to randomness. Maybe he had three bloopers fall in for hits. Maybe he actually crushed the ball to the warning track a couple other times. It doesn’t really matter because we’re seeing such a limited sample of at-bats that it will be very unlikely that any conclusions we draw will be correct; maybe they will, but again, most likely, we’ll be getting fooled by noise.

A BvP Analogy

Let me use an analogy that might help explain this better. Let’s say we want to represent a .300 hitter and a .250 hitter using two large jars of marbles. There are 500 marbles in each jar—30 percent black and 70 percent white in one and 25/75 in the other, with a black marble representing a hit and a white marble representing an out. You’re blindly given one of the jars and asked to pull out 10 marbles, then guess which jar you have.

How unlikely is it that you’ll be able to guess which jar you have based on pulling 10 marbles—or two percent of the total marbles in each jar? Pretty unlikely, right?

It’s not that each marble selection is completely useless information, either. It isn’t. If you use those 10 marbles to help you guess which jar you have, you’ll probably end up getting it right a tad over 50 percent over the long run—better than the 50 percent random expectation. So the marbles—or BvP in small samples—aren’t totally useless pieces of information; it’s just that there’s so much variance in the results that placing much emphasis on that data would probably be unwise.

So that sort of sums up my basic thoughts on BvP in a nutshell: probably not an illusion, but also probably not very useful in most situations (at least as we’re using it right now). In that way, I think both sides of the argument have made good and bad points.

BvP supporters are right that some batters probably crush certain pitchers and wrong that it’s an incredibly powerful tool for DFS decision-making (again, at least as it stands right now…more to come on that). The anti-BvP crowd is right that it currently isn’t very predictive, but wrong that the lack of predictive value means it doesn’t exist.

I’m going to try to show how we can get the best of both worlds: a version of BvP containing relevant data that isn’t completely noisy.

My Goals with This Course

My goals with this course are as follows:

About the Author

JonBales
Jon Bales (JonBales)

Jonathan Bales is the founder of RotoAcademy and author of the Fantasy Sports for Smart People book series.