clock menu more-arrow no yes

Filed under:

Daily MLB Tip: Consistency Affects Optimal Style of Play

Jonathan Bales is the author of Fantasy Baseball for Smart People—a guide designed to help you profit on DraftKings. Warren Buffett once said, “Be fearful when others are greedy and greedy when others are fearful.” I think that’s pretty sound advice for pretty much all aspects…

If you buy something from an SB Nation link, Vox Media may earn a commission. See our ethics statement.

Jonathan Bales is the author of Fantasy Baseball for Smart People—a guide designed to help you profit on DraftKings.

Warren Buffett once said, “Be fearful when others are greedy and greedy when others are fearful.” I think that’s pretty sound advice for pretty much all aspects of life, but especially the most important one—DFS.

Neither risk-seeking nor risk-averse behavior are inherently bad; that’s determined by the context. There are numerous factors that determine when you should take on risk and when you should try to avoid it, including public opinion, cost, upside, and so on.

One of the characteristics that’s associated with risk in daily fantasy sports is consistency. The degree to which we can and should go against the grain is fundamentally related to predictability, which is governed by consistency. Remember, in a perfectly predictable world, there would be no incentive to ever be a contrarian.

Baseball is the least predictable of all sports in the short-term, which is a major advantage if properly harnessed. In such a system, I think it behooves us to take more chances and really embrace a contrarian style of play. Unpredictability, if properly leveraged, is our best friend.

But not all aspects of MLB play are as unpredictable as others. While you’re going to have a heck of a time predicting Mike Trout’s home run total over the next 10 games, you can probably get pretty close to projecting Clayton Kershaw’s strikeout count in his next 10 starts.

Pitching stats are way, way more predictable than batting stats on the nightly level because of the difference in the sample size of relevant events (pitches thrown versus pitches seen). Pitchers have so many more opportunities for their play to regress toward their particular mean, which results in the pay-for-pitching strategy to generally make sense in both cash games and GPPs (dependent on cost, which has been a bit different this year).

Since pitching is predictable and the elite arms are continually those that post the top DraftKings scores, it’s a risk to fade the top pitchers in favor of cheaper options. It’s not that such a move is never justified—I “semi-punt” one of my pitcher spots with some regularity—but doing so is “being greedy” when the long-term numbers suggest you should play it safe and pay for the top arms.

For me, pitcher selection is more of a “don’t-mess-this-up” type of deal than anything else. My goal when placing pitchers into my lineup is generally to avoid risk. I don’t really care all that much about value, and in fact, players like Kershaw will almost never be among the top values in a strict dollar-per-point sense because their cost makes it very difficult to “return value.” However, I’ve never been a fan of analyzing players in this way because it 1) ignores probabilities and 2) treats players individually and not as one component of a broader picture.

Regarding the first point, I don’t care as much about maximizing the median projection of my lineup as I do about either increasing the floor or ceiling, depending on my goals. I’d rather maximize the probability of cashing by constructing my team in a specific way than to use a fragile value system to theoretically maximize points.

Further, when you pay up for a player like Kershaw, it obviously forces you to save money elsewhere, and those cheaper players are very likely to return value. So for me, it isn’t about a single player’s projection or even their isolated probability of returning value, but rather the overall probability of the lineup as a whole performing as I’d like. There’s no way to properly analyze a lineup in a holistic manner if all you care about is each individual player’s theoretical value in regards to a median projection.

In my latest book, I made an effort to detail how I try to structure daily fantasy baseball teams that will benefit from chaos—antifragile lineups that excel when “things go wrong.” I stole this idea from Nassim Nicholas Taleb.

Antifragility is of course in opposition to fragility. A wine glass is fragile—it is harmed by volatility—while Skip Bayless’s income is antifragile; the more we talk about that asshole, even if it’s with complete disdain, the more he benefits.

In between those two extremes, though, is what Taleb calls the “robust”—things that neither benefit nor are harmed by volatility. A diamond is an example of something that’s robust; unlike a wine glass, it isn’t harmed when you drop it (but it also doesn’t benefit in any way). It’s indifferent to chaos.

My goal when selecting pitchers is to be robust. I want consistency. I want predictability. I want pitchers on whom I can rely, even if the value isn’t there in terms of projected points and cost. The reason is that, due to the day-to-day consistency of pitchers, there’s not as much of an opportunity to reap rewards from being antifragile. Yes, you stand to benefit if you fade Kershaw and King Felix on a day when they’re highly owned but manage to tank—that’s certainly antifragile—but the probability of that happening is low. Whereas even the best offenses in optimal situations can still be held to a run or two, the odds are much lower that the three most expensive pitchers will all turn in duds, for example.

I want my lineup as a whole to be antifragile, at least in tournaments, but that doesn’t mean I need to be antifragile at every single spot. This is a similar concept to balancing value with usage; yes, a completely contrarian lineup stands to benefit immensely if all of the players perform well, but as you forgo greater levels of value, you’re necessarily decreasing the chances of your lineup putting you in position to actually benefit. Even if you could guarantee a large GPP victory if all of your players return value, for example, it wouldn’t be worth the struggle if the odds of that happening are 1-in-one-million.

Thus, my pitcher selection is very much about “staying in the game.” I want a high enough floor from my arms to give my bats a chance to win it for me. In that way, I take a cash-game-oriented strategy toward choosing pitchers and a GPP-oriented approach toward batters, regardless of the league type; I try to be fearful on pitchers when others are greedy and greedy on hitters when others are fearful.