Since 2007, major league baseball teams have lost almost $3 billion in injury cost to players on the disabled list. A huge portion of that cost comes from players who get hurt early in the year and miss entire seasons at a time. When a pitcher tears his UCL and undergoes Tommy John surgery, it typically takes nearly 200 in-season days for him to return to the active lineup; most seasons last about 180 days. While shoulder injuries are now far more costly in terms of both time lost and money wasted paying a player who physically cannot contribute to the team, elbow injuries aren't going away. The best teams in baseball typically lose 400 days or less to players on the DL over the course of a season, which simply won't happen for a team where one player accounts for nearly half that total.
Losing any player for a full season and beyond is never a good thing, but when the pitcher who goes down is slated to anchor a staff -- Adam Wainwright, Stephen Strasburg, Scott Baker and Daniel Hudson all fit that description -- teams can lose far more than their shot at recognition. While there are certain red flags for elbow issues, finding a set that applies to both starters and relievers is more than trivially difficult. Pitch counts aren't particularly useful for starters, but for relievers they don't begin to yield usable information; consecutive days of use might be a better measure for relievers, but it doesn't even begin to make sense for starters. One thing that both starters and relievers have to deal with, however, is pressure.
That actual game pressure is something that's virtually impossible to simulate in a lab setting, which is why this game data may deviate from the data presented by some of the leading biomechanics experts. Time and again the research has shown that offspeed pitches don't tax a pitcher's arm any more than a normal four-seam fastball does, provided, of course, that the mechanics don't change between the two. However, add in mental fatigue from bad weather conditions, an umpire with a particularly tight strike zone, an opposing hitter who has been hitting everything that night, or a pressure-filled game situation in late September and even a pitcher with typically clean mechanics can slip into poor form to get a few extra ticks on his fastball or a sharper break on a pitch. It would be great to simulate these stressors in a lab, but there's just no way to make a set-up man feel overworked and underpaid relative to his teammates if he doesn't already feel that way normally, which is why the concept of a proxy variable comes into play.
Assigning a number to the amount of pressure a pitcher feels or internalizes borders on impossible, but there is an objective measure for the extent to which a game is on the line known as leverage index or LI. The stat, invented by Tom Tango nearly a decade ago, is useful for looking at which plays and players determined the outcome of a game. For the purposes of this article, it's primary use will be to determine when the most pressure is on a pitcher. While the correlation between the pressure a player feels and the actual pressure of the game situation might not be exactly 100 percent, it's as good a proxy variable as exists until players agree to submit to functional MRIs between innings (which will definitely happen six months after the robot umpires are first installed.) Using leverage index, it is possible to observe whether a pitcher is changing his approach -- if not his mechanics -- when the pressure is on.
Fangraphs breaks down each game, play-by-play, and sorts them by leverage, making it possible to cross check those exceptionally intense moments with MLB's Gameday PITCHF/x system to determine which pitchers are throwing an abnormal amount of secondary pitches. An LI of 1.00 is average, anything below that is considered low leverage, while high leverage begins at about 1.50; according to Fangraphs' glossary, just 10 percent of all game situations have an LI over 2.00. When Heath Bell came into Sunday's game against the Cardinals, the LI was a moderate 1.76, then fell to 1.13 after he struck out Carlos Beltran to open the frame. By the time the game reached its walk-off climax, the LI had soared to 10.78.
Those moments -- those top 10 percent -- are the ones that will show what changes a pitcher might make when the pressure is really on. One thing that was important to consider when picking a sample of players to study was each player having a variety of high leverage situations to look at, lest a particular day skew the results because a particular pitch was working better at that moment than another. Having a broad variety of high-leverage appearances to look at helped ensure that a one-day quirk wasn't going to produce a false positive.
The results are nothing short of astounding. There isn't anything complicated or hard-to-explain about it. A sharp pair of eyes in the dugout and the standard pitcher charts could replicate this work easily on a team-by-team basis, but this gives you access to the same information. Of the 16 pitchers sampled, just under half increased their usage of a single secondary pitch -- usually a slider, changeup, curveball, or two-seam fastball -- by 10 percentage points or more over their average usage. For reference, a control group saw an average increase of just a hair over one percentage point, so the double-digit increases really are a substantial deviation from the norm.
Another five pitchers increased their usage, but in a less dramatic way. For pitchers like Brian Wilson, who already throw their secondary pitch upwards of 40 percent of the time, a four or five point increase in secondary usage is more than a trivial increase. All told, the Tommy John patients studied who showed an increase in their secondary usage were: Baker, Carlos Carrasco, Rubby De La Rosa, Sergio Escalona, Cory Luebke, Ryan Madson, George Sherrill, Joakim Soria, Arodys Vizcaino, Wilson, and Blake Wood.
Like every single baseball truism or cliché, there are some examples who don't quite fit the pattern. John Lackey did the exact opposite of the players listed above, throwing his offspeed pitches less and using his four-seam fastball a lot more, though the fact that he was pitching through an already torn UCL may have something to do with his pitch selection. Jose Ceda threw fewer sliders and Danny Duffy kept his pitch selection remarkably consistent irrespective of leverage. Brett Anderson and Mike Pelfrey went to their primary pitches more frequently in high leverage situations, but because both tended toward offspeed pitches in all situations, neither threw more fastballs. Anderson increased his slider usage by five percentage points, but his slider was already his primary pitch; Pelfrey increased his sinker usage by a remarkable 29 points, but again, it was already his primary pitch.
Hard and fast rules are nearly impossible to come by when it comes to sports medicine, and having specific numbers to support those rules are even harder to find. Everybody is different, every training staff has different strengths and weaknesses, and there can even be an element of luck when it comes to injuries and recovery time, but that's why red flags that can largely be applied to pitchers in general have the potential to be so valuable. This isn't a theoretical construct. Eleven teams are missing key players and a keen eye may have noticed the trouble signs before the tendon tore. With any luck, teams will start keying into details like this and there will be more intervention-based DL stints and fewer players who end up on a super surgeon's table, looking at 180 days away from their team.
The takeaway from this isn't that every pitcher who throws a lot of sliders when the game is on the line is absolutely going to blow out his elbow, but it certainly should give a team pause when they notice a pitcher -- whether they're starting or relieving -- trying too hard to make a slider bite, a two-seamer cut, or a changeup drop too frequently when the pressure is high. Some pitchers may get away with it, but it's becoming increasingly clear that they're the exception to the rule.