MLB DFS: The Truth About BvP
Batter vs Pitcher data, known as BvP, is a hotly contested subject in DFS, and takes on the tone of political debates with people who are strongly on both sides. If by chance you haven’t heard the argument, the debate is whether or not a batter’s past performance against a specific pitcher is predictive of what is going to happen in their upcoming matchup. BvP truthers say that it is an important tool to use, while the naysayers insist that is unreliable data due to small sample size and the changing skills of players. Even though I am unlikely to change anyone’s mind, I am going to tell you the truth about BvP. And like most political debates, the truth lies somewhere in the middle. But first, it’s story time.
While I could never cut it as a little league baseball player, I was very good at tennis. I still play competitively in leagues and tournaments and there are several players in town that I compete against regularly. I’m going to talk about two guys that we’re going to call Mike and Joe. Mike, Joe and myself are all pretty good at tennis for a bunch of old men. If I were to make you a very sophisticated chart of all of our skills, you would have a hard time telling which one of us was the best tennis player. I could map out speed of serve, percentage of forehand and backhand winners vs errors, foot speed, volleying ability, any data you might want. In addition to that data, I could list common opponents that we have played against and how we have fared against them. Again, from all this data you would have no idea which one of us is better. Furthermore, if you were to watch a small sample of a few games here and there of Mike playing Joe, me playing Mike, and me playing Joe, you wouldn’t know who was winning. It’s all very close.
Mike and Joe play against each other a lot, and the matches are all close and unpredictable. But, here’s the thing: I have played against Mike once a week for about four years. He beats me nearly every time. In roughly 200 matches, I’ve beaten him about 10 times, and I win a set about once a month. Do you want to guess how many times Joe has beaten me at tennis out of about 50 matches? Never. I beat him every single time. Do you know why Mike beats me and I beat Joe? It’s part matchup based and it’s part TBvP, tennis batter vs pitcher. As soon as we walk out on the court, we all know who has the advantage, and it’s very hard to get over that barrier for any of us. It is worth noting that it didn’t take long in our matches against each other for a psychological edge to be gained. By about the fifth time I played Mike, I had lost confidence. For Joe, I could see in his body language during our third match we ever played that he was not going to be able to beat me, and it nothing to do with our tennis ability.
I’m sure you see where I’m going with this, but it is not nearly as simple as saying the TBvP is real. As I mentioned, if you just looked at any small sample of data, a few matches here and there or watched a handful of games, unless I told you who has the advantage, you would simply not be able to tell. We know it’s real, but you couldn’t know without being told or having a huge sample size of data. In baseball, nobody has a 200 at-bat sample size to go off of; we are dealing with generally very small numbers of plate appearances to try and analyze. So, here is the simple and ugly truth of BvP: it is absolutely real, but we do not know when it is real and when it is randomness of small sample size. Vernon Wells told us it’s real, Zach Miner told us it’s real, and I am positive that they know better than we do what happens in the minds of major league baseball players. Baseball is a mental game as much as physical, and confidence is crucial. If a player walks up to the plate knowing that he is going to beat this pitcher, of course he has a better chance of success than the guy walking up to the plate thinking, “I hate facing this guy, he’s so hard to hit”.
At this point, maybe you’re with me, and maybe you’re not. Let’s say you agree with my premise that BvP is real, but essentially undetectable from looking at the numbers alone. What do we do with this information? DFS is a constant struggle to find the very best play at each position amongst a very large group of options. Some days you have to settle for the best thing you can find, but ideally, you can find players at each position that have every angle in their favor. With that as the goal, and knowing that we can’t get into the minds of the players, and have only statistics to deal with, I would only consider using BvP history if it is backed up by the other matchup data. It is entirely possible that in some instances, the difference between BvP being correct or not is simply that a hitter succeeds against a certain type of pitcher, which would end up producing the same data whether it’s due to BvP or just a low fastball hitter succeeding against low fastball pitcher. One of the many incredible tools available here at RotoGrinders is the Inside Edge Batter vs Pitcher Type. On the starting lineups page, if you click on a batter’s name, it will show a 1-5 star rating on how that batter fares against similar types of pitchers as they are facing on that day. In a perfect world, you will find a batter that has good BvP numbers that are backed up by their numbers against other pitchers of a similar style.
There is another side of the BvP conversation that I don’t hear mentioned very often. If a hitter has strong BvP numbers against a pitcher, there are basically four options to consider:
1) BvP is real, that hitter has a specific psychological advantage over that pitcher.
2) That hitter is strong against the particular type of pitch that the pitcher throws. As in, hitter is a good changeup hitter, the pitchers best pitch is a changeup.
3) It’s completely random data due to a small sample size and is not relevant to what will happen next time they face each other.
4) It has nothing to do with either BvP or pitch type, but is simply a matter of one player is better than the other. This is the idea that I don’t hear a lot of talk about. Are we overthinking it to try and figure out why Ryan Braun has good numbers against Kyle Kendrick? Edwin Encarnacion, Jacoby Ellsbury and Miguel Cabrera all have great numbers against Jeremy Guthrie. Well of course they do, but maybe that is just because they are all elite hitters while Guthrie is an average at best pitcher.
So What Do We Do About Adam Jones vs John Danks?
We all know that John Danks is terrible against right-handed hitters, while Adam Jones is excellent against left-handed pitchers with a career .380 wOBA and .544 slugging percentage. We also know that Jones is a very good hitter overall, while Danks is a below average pitcher. Everything about this matchup screams to play Adam Jones. So, why is he 1-for-33 lifetime against Danks and what do we do with this information? Even the most outspoken BvP skeptic should be able to admit that it is possible that Danks has a real BvP advantage over Jones, either psychological or some hole he found in Jones swing that nobody else has found. And, even the most ardent supporter of BvP should be able to admit that it is possible that this is just complete randomness from a small sample size, and Jones could just as well go 4-for-4 next time they face each other. Here is what I suggest you try with data that is this extreme:
In cash games, you should be able to find a player similar to Jones ability and salary with a matchup that comes without a red flag like this. Maybe you think the BvP is complete nonsense, but why would you risk it?
In tournaments, you can hope that the BvP data scares off more people than it should. If it turns out the BvP was just a sample size anomaly, then you get a great right-handed hitter against a bad left-handed pitcher at lower than normal ownership levels.
Wrapping It All Up
There is a good chance that you had your mind completely made up about BvP a long time ago, maybe you use it and maybe you don’t. The best thing I can do as a fantasy baseball analyst is to try and help you improve your game in any way possible, and perhaps try some new things. So, with that in mind, I want to encourage you to look at the opposite of the BvP debate, and try to understand what it would mean for your game if the other side were true.
- If you don’t believe in BvP, spend a few days looking at it. You don’t have to change your lineups based on the data, but at least think about how you would play differently if you were open to believing in the extreme cases. It can’t hurt you to know that Adam Jones is 1-for-33 against John Danks or that Elvis Andrus is hitting .455 against David Price. You can choose to ignore the information, and you might be correct to do so, but it is useful to know that other players are looking at it.
- If you already use BvP, and use it to help set your daily lineups, try setting your lineups without it. Use the larger sample size historical data, and after your lineups are finished, go ahead and take a look at the BvP. If something doesn’t match up, then you can decide which information you want to value higher.
In conclusion, to summarize what I consider to be the truth about BvP:
It is real; some players have a psychological advantage over other players, and it is silly to think otherwise. BUT, just because it is real doesn’t mean that we can detect it from looking at the numbers. As is the case with most heated debates, the truth lies somewhere in the middle.