RADIO
One of the most fundamental questions in fantasy sports is if a player's current performance is sustainable. More than any other sport, baseball has a slew of statistical measures that can be dissected numerous ways to analyze player performance.
PITCHf/x is a publicly available pitch tracking system that provides a lot of different data to help fantasy owners make this determination for mound breakouts and busts alike.
In this article, you will learn how to effectively use this data to give you an edge in your fantasy baseball league throughout the season.
Editor's note: Be sure to check out all our strategy articles on how to win your fantasy baseball leagues. Our Sabermetrics series - Part 1: BABIP for Hitters, Part 2: HR/FB%, Part 3: Batted Ball Distribution, Part 4: Plate Discipline; Part 5, Pull %, Part 6: Lineup Slot and Counting Stats, Part 7: FIP and xFIP; Points League Primer; Using SIERA to Win Your League; How Your Brain Messes with Your Drafts; and Why You Shouldn't Overpay for Saves.
How to Interpret PITCHf/x Data
The first data point, and the easiest to understand, is velocity. Generally speaking, a pitcher that loses fastball velocity is losing something to either an undisclosed injury or the aging process. Pitchers that gain velocity can expect to increase their production. For example, Seattle's James Paxton increased his average fastball velocity from 94.1 mph in 2015 to 96.7 last season, striking out more batters (18.9 percent K% in 2015 to 22.9 percent last year) as a result. The average major league heater was 92.6 mph in 2016, though of course a pitcher's established baseline is a better indicator of future performance. Other variables like movement and location also matter, but velocity is a good introduction to using PITCHf/x data.
Slightly more advanced is pitch mix, or what pitches a pitcher throws and how often he throws them. A pitcher may improve his production by abandoning a poor pitch or developing a new, effective one. This is a good stat to consult if a pitcher sees a sharp change in his GB% or K%, as a change in pitch mix could represent the change in approach that justifies the new number. If the change does not have a corresponding pitch mix shift, it may be less sustainable.
For example, consider Johnny Cueto. His GB% increased last year relative to 2015, 42.4 percent GB% to 50.2 percent. His K% spiked in the same time frame, from 20.3 percent to 22.5 percent. Are these numbers the result of random fluctuation, or did Cueto change his pitch selection to bring them about?
PITCHf/x tracks each pitch's individual results, so any change in pitch selection can be evaluated by comparing an offering's usage percentage and its performance, in this case GB% and SwStr%. The biggest change in Cueto's pitch selection was that he threw fewer heaters (30.6 percent to 23.2 percent) in favor of sliders (11.9 percent to 19.3 percent) relative to 2015. Cueto's fastball posted the lowest GB% in his arsenal with a rate of 38.1 percent. The slider produced a 50.3 percent rate, so additional sliders at the expense of fastballs would indeed be expected to increase Cueto's GB%.
Cueto's strikeouts are not supported by the switch however. His big K pitch is his change, which compensates for a low Zone% of 35.2 percent with a strong SwStr% (18.4 percent) and chase rate (43.6 percent). His heater sets this up well with a Zone% (55.6 percent) that gets him ahead in the count while also boasting a strong SwStr% (9.7 percent). The slider is weaker on both fronts (52.5 percent Zone% and 8 percent SwStr%), so we would expect more Ks from the fastball. Cueto's strikeout rate is due for regression.
The same type of analysis may be performed for a number of other stats, including FB%, LD%, BB%, HR/FB and even BABIP. There is no point in looking at a league average pitch mix, as every pitcher owns a different arsenal. All of these variables may be considered over a pitcher's complete repertoire to determine how good he is (or should be) without relying on any conventional metrics. This can be good for identifying sleepers, as pitchers that have one or two stand out pitches could break out by simply using them more often. Let's have some fun with our example and look at Clayton Kershaw.
Kershaw threw four different pitches in 2016: a fastball 50.5 percent of the time, a slider 33.3 percent of the time, a curve 15.6 percent of the time, and a change 0.6 percent of the time. The change was thrown 13 times over the entire season, so it may have been a misrecorded slider or a rare mistake pitch. At any rate, the sample size is too small to consider it in this discussion, leaving three offerings for our analysis.
His fastball registered a Zone% of 62.8 percent, explaining how he barely walked anyone (11 all season). Most pitchers who live in the zone like this get hammered, but Kershaw is not an average pitcher. The heater recorded an above average 9 percent SwStr% despite living in the zone, allowing batters to hit only .246/.269/.371 (AVG/OBP/SLG) against it. It was a good pitch, but not enough to make Kershaw his vintage self.
That is what the slider is for. It was only a strike 46.6 percent of the time, but compensated by making hitters chase it at a whopping 47.7 percent clip. That helped give it a SwStr% of 25.7 percent, absolutely obliterating the league's 10.1 percent SwStr% rate and explaining how Kershaw compiled 172 Ks in just 149 IP last year.
Kershaw also has a curveball. It was a strike even less frequently than the slider at 32.4 percent, and also posted a mediocre O-Swing% of 31.8 percent. This gave it a SwStr% of 14.6 percent--very good, but inferior to Kershaw's slider. Why throw it?
Sometimes, hitters actually put the ball in play. Batters managed a triple slash line of only .118/.118/.161 against Kershaw's curveball in 2016, compared to .138/.152/.191 against the slider and .246/.269/.371 against the heat. All three are well above average, and Kershaw's arsenal is an embarrassment of riches if there ever was one. He's fun to look at, but he can't be a baseline.
What is the baseline for this type of analysis? It depends on the observer, as there are almost as many ways to interpret this data as there are data points to consider. The league average O-Swing% was 3o.6 percent in 2016, and most good wipeout-type pitches need to beat this number substantially. The overall Zone% was 47.8 percent, including pitches like splitters in the dirt and high fastballs that were never intended as strikes.
The fastball will always be inferior in results to pitches that do not need to live in the strike zone, like Kershaw's slider, as pitches hit outside of the zone offer better results than offerings in the hitting zone when they are put into play. However, getting ahead in the count is necessary to make those pitches work as intended, making mediocre fastball results a necessity.
It is dangerous to generalize, but 2-seam fastballs and sinkers tend to stink for fantasy purposes. They're usually in the strike zone, but get hit harder than fastballs. They may post strong GB% rates, but also have high BABIPs and scary triple slash lines. Any sinker hit in the air was probably a mistake, so the HR/FB rate is usually high for the limited number of fly balls hit against them. Their SwStr% rates also tend to be poor. Overall, fantasy owners prefer a fastball or cutter to be the strike zone pitch in a pitcher's repertoire.
Personally, I like a fastball with a SwStr% of around 9 percent and a Zone% of at least 53%. Many pitchers succeed with a lower Zone%, but I can't stand watching walks. I then look for a wipeout pitch that offers a SwStr% of at least 15% and an O-Swing% of 40%. Ideally, there is a secondary K pitch, like Kershaw's curve, that prevents the 0-2 pitch from being too predictable. Only aces really fulfill all of these criteria, but I can dream, right?
Conclusion
To conclude, PITCHf/x tracks a lot of data of interest to fantasy owners, including average velocity, pitch mix and individual pitch results. All of this data may be used to predict who will break out or which breakouts can sustain their current performance. The next entry in this series will discuss how to deal with minor league stats, which do not include all of the advanced metrics discussed thus far. Projecting prospects has increasingly become a part of every fantasy owner's job, and there are ways to analyze them beyond a blind faith in homers and ERA.
