Fantasy: Batter-vs.-pitcher data can mislead
The Official Mini Fantasy Game of MLB.com, free and hosted at DraftKings.com, is gaining in popularity. After all, people are winning prizes by playing the free game.
As with any game of competing minds, players who partake in daily fantasy baseball look for every little edge they can get, searching far and wide for stats that will help them figure out who's likely to crush a 450-foot homer that night. One of the most popular stats is batter vs. pitcher splits (BvP).
BvP stats are simply one batter's numbers against a specific pitcher. For example, the Braves' Gerald Laird is 2-for-20 in his career against Phillies left-hander Cliff Lee. Daily fantasy players often study BvP splits to try to gain an advantage over their opponents, seeking out players who have hit or pitched well against specific individual opponents in the past.
The question, then, is: Do BvP splits lead to an advantage for daily fantasy players who emphasize them? In most cases, the answer is no. The reason? Sample size. Even if we knew nothing about baseball, we can test for the significance of BvP numbers by using statistics. What would such tests tell us about Laird getting two hits in 20 at-bats against Lee? They'd say the numbers are meaningless. The problem is that Laird hasn't faced Lee nearly enough times for us to discern whether his .100 average is the result of his inability to hit Lee or just bad luck. There's a good chance it's the latter.
If you think about the nature of BvP stats, it's kind of obvious that they're fragile. In Laird's battles with Lee, maybe he's actually hit the ball fairly well but still been put out; maybe unusual wind gusts held back a home run, or perhaps he hit a few line drives right at defenders. If Laird had just three more hits drop in against Lee -- making him 5-for-20 -- he'd actually check in slightly above his career batting average.
The point is that we generally can't trust BvP stats because the sample size is too small. We can't possibly know if the results are real or just noise -- the byproduct of randomness. To be clear, this doesn't mean that certain batters aren't worse against particular pitchers than others -- we'd expect that to be the case, actually. With that said, we need a sample size of at-bats much larger than 20 to ensure we're not being fooled by randomness.
So can BvP stats help you succeed in the Official Mini Fantasy Game of MLB.com from DraftKings? Yes, but only when a batter has a significant history against a certain pitcher, such as two players who have played in the same division for years. When you see 5-for-50 instead of 2-for-20, then you're starting to get into significant territory.
Remember, the Official Mini Fantasy Game of MLB.com from DraftKings is free to play. The game's top two nightly winners will receive 30-day MLB.TV Premium subscriptions. And even if you don't win, you'll still win. When you play the Official Mini Fantasy Game of MLB.com from DraftKings on any given day, you unlock free entries into three separate DraftKings contests: the MLB.com Regular Season Tickets Contest, the MLB.com All-Star Game Contest and the MLB.com World Series Contest. The prizes range from two tickets to a regular-season Major League game of your choosing to two tickets, lodging and airfare to the 2014 World Series.