Bigfoot
Dating back to the days of the indigenous people of the Pacific Northwest there have been tales of an ape-like, bi-pedal creature that roamed the woodlands. Stories of Bigfoot, as the creature would come to be known, permeated the region and eventually the nation. For years, the mythology surrounding Bigfoot survived as nothing more than a tall tale, with little evidence to support its existence.
All of that changed in 1967, when the highly scrutinized Patterson-Gimlin film footage of an apparent Bigfoot was released. To the pro-Bigfoot crowd, this footage constituted legitimate proof that the mythological beast was real. What proof could one need other than the visual confirmation provided by this film?
Yet, 49 years later, there has yet to be a shred of reputable scientific data to support the existence of this creature. In the eyes of many, the lack of such data suggests that these tall tales are nothing but stories. With such a divide between the pro-Bigfoot and pro-science communities, it’s unlikely that anything other than the discovery of a living Bigfoot will close the gap. The conflict between the two sides raises a very important question, one that we deal with in many different realms of our lives: Is it possible (or reasonable) to believe in an idea that appears to be unsupported by any underlying data?
BvP: What Is It?
If you mention the term BvP in the company of baseball aficionados, be sure to have some popcorn on hand, because a debate that would put Skip Bayless and Stephen A. Smith to shame is about to ensue.
What is BvP? It stands for “Batter versus Pitcher.” Put simply, it’s how a specific batter has fared versus a particular pitcher. It’s a divisive topic within the baseball community as a whole, but likely even more so within the world of MLB DFS. Supporters on each side of the spectrum tend to have very staunch stances on the topic.
The crowd that fails to accept BvP as a legitimate metric tends to rely upon the fact that the samples we have to work with aren’t of a sufficient size to have statistical relevance. Contrarily, the pro-BvP crowd believes that, despite the small sample size, we can see that players can possess the propensity either to have an inordinate amount of success or failure versus a particular pitcher.
Setting up a Sample Population
One of the inherent problems with determining the validity of BvP surrounds the lack of sufficiently-sized samples. There has been plenty of research surrounding the idea of sample size. The graphic below from FanGraphs breaks down the minimum requirements:
Unfortunately, we’re just not going to get anywhere near these totals, and that’s a sticking point for the anti-BvP crowd.
Still, let’s peruse the data that we have. Looking at the BvP matchups from the past calendar year, I’ve elected to analyze only players who have had a minimum of 40 at-bats against a pitcher, meaning that the pitchers and batters have (generally) faced each other on at least 10 separate occasions.
Utilizing this at-bat requirement, we’re left with a sample population of 54 batter/pitcher matchups, with a number of players appearing more than once. With the idea that we are looking for outliers, we’ll further reduce the list to include only players who have had extreme success — let’s say a .350 batting average or higher.
Only 21 batters remain, but three of them appear multiple times, a fact that prompts this question: “What if BvP does exist, but not necessarily in the way that we assume?”
BvP: Batter vs. Pitcher or Batter vs. Pitcher Profile?
There’s an aspect of BvP that is unquantifiable, which I won’t even attempt to address in depth: The mindset of the batter and pitcher. Athletes are human and if they routinely struggle against particular opponents then are likely to allow these failures to impact them. There’s no way for us to know which players may be predispositioned to these thoughts and so it’s difficult to include this element in any sort of analysis.
Instead, I’ll focus on variables that we can quantify. A very basic understanding of the game speaks to the idea that players will generally perform better against a specific handedness. Could the same be true for different pitch velocities? Are some batters better against hard throwers and others against soft tossers?
Let’s examine how batters perform against certain pitcher profiles (based on handedness and average velocity), using our Plus/Minus metric to gauge performance.
Handedness and Average Velocity
In evaluating BvP, I’ll be utilizing our Trends tool, as it now contains advanced data (including pitch velocity).
To get an idea of a player’s baseline performance, we’ll examine his Plus/Minus against the specific handedness of the opposing player in the BvP matchup, and then we’ll compare that to his Plus/Minus against all pitchers with that handedness and comparable average velocity.
Let’s start with a caveat: This data dates back to only the 2015 season, so we may not have enough coverage to make definitive statements about all members of the population. However, with 15 of the 21 samples showing a greater Plus/Minus against pitchers with a comparable profile to their BvP opponent, there does appear to be evidence that these batters are finding success against pitcher’s beyond just those their opposing BvP matchup.
In Conclusion
The BvP phenomena as it is traditionally examined may not be something we can ever fully analyze, if for no other reason than a lack of plate appearances between any combination of batter and pitcher. Expanding the analysis to include the makeup of the pitcher as we’ve done here may allow us to gain a larger sample size with which to measure the batter’s performance.
If the batter is routinely finding success against pitchers of a comparable makeup — and you could look at additional factors such as pitch selection, arm angle, etc. — there is likely something to the high BvP performance. If, though, a batter doesn’t do well against comparable pitchers, then the high BvP performance (such as Posey vs. Kennedy) might be an outlier, and perhaps over time we would be likely to see player regression.
Because the sample sizes still aren’t likely to be of a sufficient size to satisfy the stat crowd, and because I haven’t included a strict “head-to-head” component in the test, it’s possible that the findings won’t impress either side. But working somewhere within this middle ground — looking at pitcher profiles instead of just the actual pitchers themselves — appears to be a strategy that could lead to future success, no matter which side of the debate you fall on.
