I start my daily fantasy baseball research on the macro level by looking at the Vegas lines and weather. I try to get a really broad sense of which teams/pitchers might have the most upside in a given day. Part of the reason that I do that as opposed to first looking at individual players is because a top-down approach to research makes sense given how correlated fantasy production is in baseball
Still, this is an individual game, so it’s vital to be able to sift through a vast amount of data for every available player. In terms of individual player research, I start with lefty/righty splits. The difference in how players perform based on handedness can be drastic. Here’s a look at OPS—On-Base Plus Slugging—for the top 150 batters since 2000.
You can see that right-handed batters as a whole are better against lefties than righties, while lefty bats are much, much better against right-handed arms than southpaws. It’s vital to understand how a batter or pitcher’s handedness and matchup will affect his performance in a given day.
I always want to have the most exposure to the players with the most favorable splits. Though that typically means playing batters against pitchers of the opposite handedness, that’s not always the case. Some guys are even splits players who can be started against the same handedness of pitcher. Mike Trout is one such example.
For pitchers, handedness still matters a lot, specifically in situations in which they’ll face a lineup with an unbalanced handedness of hitters. Managers often shake up their lineups to account for the handedness of the opposing starting pitcher, so sometimes you’ll get a southpaw facing an entirely right-handed group of hitters, which is of course generally less ideal (from the pitcher’s viewpoint) than going up against a balanced lineup.
So handedness splits are probably the most important factor I consider, but we still need specific numbers to analyze. For me, those stats are wOBA (Weighted On-Base Average) and ISO (Isolated Power) for hitters and xFIP (Expected Fielding Independent Pitching) and K/9 (Strikeouts Per Nine Innings) for pitchers.
Weighted On-Based Average is a Sabermetrician’s dream because it’s a really nice catch-all statistic that does a nice job of capturing overall offensive quality at the plate. wOBA is basically a superior On-Base Plus Slugging. OPS does a good job of combining both power and the ability to reach base, but it doesn’t weigh certain achievements in the proper way according to how much they’re worth to an offense. wOBA corrects for that, providing a really accurate representation of a hitter’s ability.
As with every stat I analyze, I care about wOBA splits—not overall wOBA. Some hitters have well above-average wOBAs against righties but can’t get a hit to save their life against southpaws, for example, and I want to know that.
DraftKings does a pretty good job of pricing players according to their overall quality, which leaves room for us to find inefficiencies in that pricing based on splits. If Player X is priced at $5000 because he has an overall wOBA of .350, that pricing might be accurate in a vacuum, but we’d still be overpaying for him against lefties if he has a .320 wOBA against southpaws (and thus underpaying for him versus right-handed arms, against whom his wOBA would be higher than .350).
A .400-plus wOBA is elite territory—only six players surpassed the mark in 2014—and anything above .340 or so is really good. The league average has dipped over the past decade, for obvious reasons (global warming, am I right?), but .312 or so is about average league-wide.
ISO—Isolated Power—is a very simple metric that calculates raw power by dividing extra bases by at-bats. I like to use ISO because so much of daily fantasy baseball success, especially on DraftKings, comes down to giving yourself as much access to home runs and extra-base hits as you can get. To win a big GPP, for example, you’re probably gonna need some dingers, and ISO will help you deliver those. Typically, I’ll be more inclined to use wOBA in cash games, but some combination of wOBA and ISO for GPPs.
And again, everything is broken down by handedness. I don’t care about overall ISO numbers—just ISO splits. The league-average ISO is right about .135.
While ISO and wOBA do a really nice job of capturing the majority of what you need to win, they’re not the only pieces of the puzzle. Neither metric accounts for stolen bases, for example, which can be a really large part of increasing your upside and safety in daily fantasy baseball.
Typically, I have one general rule of thumb when selecting batters: if he doesn’t have the ability to either go deep or steal a bag with regularity, then I don’t want to use him. Baseball is too volatile of a sport to rely on someone who hits for average, for example, since they still have big downside (zero points is realistic for any player in any game), yet don’t offer you very much upside.
Let’s move to the mound, where one of the most popular catch-all statistics—the wOBA for pitchers—is xFIP (Expected Field Independent Pitching). xFIP is a derivation of FIP, which calculates what a pitcher’s ERA would look like if he had normal results on balls that are put into play. The idea is that pitchers don’t have much control over what happens once a ball is hit, so FIP attempts to remove the “luck” and provide a number that’s more predictive of future ERA than past ERA.
xFIP adds another layer by also calculating how many home runs a pitcher should have allowed based on his fly ball rate—something that generally regresses toward the mean. So in effect, we’re trying to calculate what a pitcher’s ERA should look like based on how he’s pitched, attempting to account for the randomness of batted balls.
I like to exploit this when a pitcher’s xFIP is much different than his ERA. Earned Run Average is such a commonly used statistic and a big component of his DraftKings salary. When a pitcher has an ERA of 3.00 but a xFIP of 4.40, for example, that’s a sign that he’s probably going to regress, and thus could be overvalued. The league average xFIP is typically around 3.75 or so. Anything around 3.15 is great, while anything approaching 4.30 is considered poor.
While xFIP captures a pitcher’s overall ability very well, it doesn’t directly account for strikeouts. And if you’re playing daily fantasy baseball on DraftKings, you need to target high-strikeout pitchers. The strongest statistical correlation for pitchers in daily fantasy baseball is the link between strikeouts and 50/50 wins. The correlation between strikeouts and winning cash games is so strong that you could make an argument that the majority of your daily fantasy baseball research time should be allocated to predicting strikeouts for pitchers.
K/9 simply calculates the number of strikeouts a pitcher records per nine innings. It’s not the only metric that matters, but K/9 is a major piece of the puzzle. And since it’s somewhat independent of xFIP, you can use the two in conjunction without much overlap (as opposed to wOBA and ISO).