The advent of daily fantasy baseball has brought about new areas of focus in the way the fantasy industry analyzes hitters. As you would expect more research is put into single-game matchups and predicting short-term outcomes. One such stat that you can find in use on a ton of sites focused on daily fantasy baseball is Batter vs. Pitcher, or BvP. On the surface it makes perfect sense. What could be more relevant than the history a hitter has against a specific pitcher? Cut through all that lefty/righty stuff and drill right down to that one-on-one matchup. You hear announcers talk about how a hitter has done against that day’s pitcher all the time.
Before I tell you why BvP is pretty much worthless for the vast majority of fantasy players, I’m gonna tell you that I am certain that there are certain hitters who hit better against specific pitchers than they do against others. Some will consider this an opinion, but I having played baseball at a high level, I know it’s a fact.
The Research Behind Batter vs. Pitcher Numbers
I’m not going to get too deep into the research, but there has been quite a bit of it showing virtually zero correlation between how a batter has done against a pitcher and how he will do in the future. Here’s a piece that Derek Carty, a former cohort of mine, did for the FanDuel Insider. Dave Cameron also does a nice job of summarizing BvP stats in this piece over at FanGraphs. I’m not a statistical guru, but I trust these guys when their research shows that the numbers don’t lie. In a vacuum, with no context included, previous history has very little to to do with forecasting future performance. That seems to be pretty much a fact.
WHY BvP Doesn’t Work
Most of the analysis I’ve seen on this subject simply shows you the research and stops there. BvP doesn’t work. Period.
I’m here to explain why it doesn’t work. It seems so intuitive that seeing a hitter dominate a specific pitcher should mean something, that’s why too many people ignore the research.
But wait you say, didn’t I tell you earlier that I considered it a fact that some hitters excelled against certain pitchers? Yes, I did, but here’s why you can’t use that fact.
Let’s take 100 positive batter vs. pitcher matchups. Let’s say 20 of them are true examples of a hitter having “figured that pitcher out” and he will continue to dominate that pitcher. That means the other 80 are basically a result of sample size. Maybe a few bloopers have fallen in, or the wind was blowing out that day. Maybe they caught that pitcher on a few off days. Whatever the case, they’ve built some pretty impressive numbers that have no real meat to them. They are just the result of what a small fickle sample size can do.
What happens to numbers that are skewed by small sample size? In general they regress to the mean. That is, they are going to start looking more like what we should expect. So here you have 80 hitters who are very likely to experience extreme struggles against a pitcher that they’ve previously “dominated.” All that regression pretty much negates the cases where a batter truly excels against a certain pitcher. The research takes all hitters as a group, so when it’s all said and done, BvP is a very dubious basis for selecting hitters for your daily fantasy baseball lineups.
Do Batter vs. Pitchers Stats Have Any Value?
I’ve already shown you that BvP numbers are not a reliable indicator of future performance, but I’m going to be honest and tell you that I do check them out in some cases. Rather than look for instances where a hitter has done well against a pitcher, I look for the opposite; a hitter who has been dominated by that day’s pitcher. But I don’t just look at batting average. I’m more interested in an extreme strikeout percentage. A high strikeout percentage can still be a result of sample size, but much less so. You’re not just looking at the whims of BABIP; if a hitter strikes out half the time he faces a pitcher, he’s likely been more than just a little unlucky. So basically I’ll use BvP as a filter, eliminating hitters from my daily options. But even with this it has to be a pretty extreme example and I’ll still run it through a common sense filter.
As I said, I do think there are cases where certain hitters will experience more success against specific pitchers than their overall numbers would suggest. It’s just too hard to separate with all the sample size noise. I think BvP could be of some use in conjunction with high level scouting skills, but for most of us, it’s best to pretty much ignore the numbers and rely on other data sources… but that’s a whole ‘nother article.
Latest posts by Doug "RotoDaddy" Anderson (see all)
- 2017 Fantasy Baseball Auction Strategy - March 8, 2017
- Fantasy Points Allowed by Position for NFL Week 6 - October 12, 2016
- FanDuel Fantasy Points Allowed by Position for NFL Week 2 - September 14, 2016
Voted “Best Fantasy Draft Tool”!
“So-Called Fantasy Experts” Podcast!
Fantasy Blog NetworkGet This
Powered By: Fantasy Knuckleheads