On Thursday, I looked over the
relevant factors for evaluating batter matchups. Today, I’m going to do the pitchers, first going over each of the factors you need to consider in detail and then providing a short do-it-yourself formula. Doug Anderson recently posted a good article identifying
important factors for pitchers, so there will be a little overlap here. (Since there’s also some overlap between the factors that are important for hitters and pitchers, some text has simply been copied and pasted. It’s been italicized to make it easy to skip past.)
True Talent Level — As with hitters, true talent level is the first and most important thing to consider. It sets the baseline for the player, which we then layer all of these other factors over.
It should also be noted that it’s important not to confuse True Talent Level with a player’s actual performance. As I’ve shown before, a player’s salary is heavily influenced by his current season performance (more heavily on some sites than others), but a player’s actual performance can be far from a reflection of his true talent level. There’s big value to be found just by identifying players who are underperforming.
Opposing Offense — This is one of the most vital factors that you must consider when deciding who to plug into your pitcher spot on any given day. Ideally, you want to pick a pitcher that’s up against an offense that doesn’t score many runs and strikes out a lot. Unfortunately, strikeouts are correlated with higher run totals (since lots of strikeouts usually means lots of power), so these teams are on the rare side. When the two conflict, go for the team that doesn’t score many runs. You can afford a couple fewer strikeouts if it means fewer runs and getting the win.
Opposing Pitcher — The guy who’s pitching for the other team is a relevant factor. The runs he allows will influence the decision of the game, although its importance is often overblown by daily players. Unlike most of the other factors in this article, it only impacts one stat, wins (and losses, for sites that use it), and it is far down on the list of factors that influence wins most.
Supporting Offense — This may be built into your true talent level projections, but you should leave it out whenever possible, since team lineups change from day to day, deal with injuries/call-ups/trades, etc. The supporting offense is important because they need to score runs in order for the pitcher to win.
Supporting Defense — Not as important as the offense that’s backing up a pitcher, but the defense is still noteworthy. It’s the reason why Rays pitchers allowed so few hits on balls in play for years. It becomes less important the more strikeouts the pitcher generates (and thus the fewer balls in play he allows), but it’s still a medium-importance factor for all pitchers. Additionally, you can evaluate the outfield and infield differences separately, particularly when the pitcher is an extreme flyball or groundball hurler. Just be careful how you do your evaluations, though, because defensive metrics are riddled with noise.
Home Field Advantage —
Home field advantage is not an enormous influencer of player performance, but it is important. Suffice it to say, players on the road take a value hit.
Ballpark —
Everyone knows that ballparks play differently, but most people just consider them in terms of home runs or hits. Home run park effects are the most important and largest ones, but few realize that parks also affect things like strikeouts and walks. A lot of daily players will simply go to ESPN and look at their single-year, overall park factors, but take heed: single-year park factors can be highly deceiving. As a rule of thumb, you should be looking at a minimum of three years worth of data, as a starting point. They don’t do them the same as I do mine, but StatCorner uses a pretty solid method and a number of years when calculating theirs, if you’re looking for a reference guide. Another big pitfall some daily leaguers make is relying on “career in park” data. If a pitcher allows a lot of home runs in PNC Park, chances are that’s not going to continue since PNC Park is extremely favorable for pitchers. He’s just gotten unlucky. Ignore these numbers and rely on the park factors instead.
Weather —
In line with the effects of ballparks is the effect of weather. I’ve yet to do much extensive research into weather, but in general terms, wind blowing in is bad for power and wind blowing out is good for it (duh). Temperature, humidity, and such also play a part, but I’ll be delving into weather more in the future. And of course, weather needs to be taken into consideration because rain can lead to a game not being played. You don’t want a player in such a game in your lineup.
Catcher — The catcher’s impact on the running game is the main consideration for hitters, but for pitchers it goes well beyond that. A catcher who can frame pitches and call a good game can have a big effect on a pitcher. The baseball community is still in the process of quantifying these things, but Max Marchi has done an excellent job at Baseball Prospectus over the past couple years. A good (or bad) catcher can impact a pitcher’s ERA by 0.10 points or so.
Umpire —
Despite playing a part in the outcome of almost every single pitch, the umpire is rarely a consideration for daily fantasy players. Part of this is a lack of easily accessible data. Another part is not quite knowing what to do with this data. Just looking at the raw numbers is a recipe for disaster. I’ve developed a system for umpires that I’ll be talking about at some point in the future. Umpire effects are limited to strikeouts and walks, but since those are the two most important stats for pitchers, umpire effects are definitely worth considering. My results are still preliminary, but it seems that a good or bad umpire has about the same effect as a good or bad park does -- and a greater effect for certain batter/pitcher handedness combinations. A good umpire/park combo can improve a pitcher’s strikeout rate by 20 percent (which will also bleed into improved run prevention, naturally).
Platoon Advantage — This is less important for pitchers than it is for hitters. For hitters, they have a single opponent. For pitchers, they have a whole team of opponents, so getting much out of the platoon advantage is difficult. If anything, this is a factor you need to check off just to make sure that it’s not a negative. For a pitcher like Justin Masterson with a big platoon split, make sure that the opposing team isn’t stacking their lineup with opposite-handed batters. Unsure which pitchers have legitimate platoon splits? They’re more stable than those for hitters, plus you can predict them well using arm angle (the lower the angle, the more extreme the split) and pitch repertoire (sinkers and sliders generate the most extreme platoon splits).
Home/Road Splits — Home/Road splits should be ignored. They are based on small samples and ignore half of the player’s performance. Instead, focus on park factors and home field advantage.
Hot/Cold Streaks —
Lesser analysts love to use hot and cold streaks to prove their points or make their player recommends. Others will piddle around by saying, “Well, the sample size is small, but this happened so I’m going to talk about it anyway.” There’s no room for pussyfooting here. Hot streaks should be ignored completely, and cold streaks should only be relied upon if they’re dragging on for an inordinate period of time (over a month, ideally longer). It’s far more likely that a player is having mechanical, health, or psychological difficulties than that he’s somehow legitimately pitching better than he truly is.
As I’ve discussed before, there’s a big difference between statistics and sabermetrics. Any chump can look at data, but it takes an extra level of skill to decipher what the data means, whether it’s important, and how best to use it. You can see some of my previous work on
hot streaks and
cold streaks here.
Career vs. Team Data — This is the most useless piece of data you can possibly look at (assuming you’re not reading tea leaves or something), particularly for veteran pitchers. Team rosters are constantly changing, so how does it make any sense to evaluate David Price’s chances against the Yankees by including the data from when he faced Jorge Posada in 2009? Even if you only focus on the players currently on the team the pitcher is facing, you still are dealing with small sample sizes and a potentially unrepresentative mix of hitters (i.e. the pitcher has faced Eduardo Nunez 20 times but Alfonso Soriano just twice). Plus, pitcher talent levels change more quickly and frequently than those of hitters, making it less reliable when you’re going back more than a few years.
Additional Factors — There are additional trend-like factors that are likely to be important to some extent for pitchers, but which I haven’t conducted extensive enough research into it to really say for sure. Things like days of rest, pitch count in the last start and the like are all on my list for topics to explore in the future.
A Simple Formula To Get Started
As you can see, there are a lot of factors that go into figuring just how favorable a pitcher’s matchup is on any given day. I know that I didn’t exactly provide explicit details on how to evaluate some of those factors. For some, it’s because I’m still studying them myself. For others, it’s because to do them right (or even at all) takes a lot of work, but that’s why DFSEdge is here: to do most of that heavy lifting for you. If you’re a do-it-yourself kind of person, though, and want a simple formula for evaluating players yourself, pay particular attention to True Talent Level, Opposing Offense, Supporting Offense, Ballpark, and Home Field Advantage.