Lance Lynn and Finding Relief in the Underlying Underlying Metrics

Lance Lynn was phenomenal in his final start before the All-Star break, going seven innings, striking out 11 batters, and allowing just three baserunners. By game score, his outing against Toronto was among the top 20 starts of the year. But if you prefer to focus on factors solely within a pitcher’s control — strikeouts, walks, and home runs — it wasn’t even the best of Lynn’s season. By FIP, his most dominant performance was a 16-strikeout gem against the Mariners a few weeks prior. Over seven frames, he walked only two and kept everything in the ballpark. Even more impressive, all but one of his strikeouts came on a swing and miss, as Lynn collected 33 whiffs on the evening, the most by any pitcher in a single outing this season. His 16 strikeouts were also the most in a single game this year; no other pitcher has topped 13.

Lynn has two more double-digit strikeout games this year, one each against the Rays and the Twins. Only eight other pitchers have four or more double-digit strikeout games, and only four have multiple starts of seven-plus innings, double-digit strikeouts, two or fewer walks, and zero home runs: Kevin Gausman, Zac Gallen, Logan Webb, and Pablo López. (The White Sox, being the 2023 White Sox, went on to lose all four contests, but that’s hardly a reflection of Lynn’s efforts on the mound.)

So why am I trying to prove that Lynn is still so good at what he does? Because the traditional statistics haven’t been kind to him lately. No one has given up more runs this season, and his 6.03 ERA is an ugly reminder that bad things happen to good pitchers. His 4.81 FIP is lower, but that isn’t saying much; his ERA ranks 63rd out of 64 qualified starters, and his FIP ranks 54th. It’s easy to look at those numbers and assume Lynn’s age is starting to show; he turned 36 in May, and his average fastball velocity has been dropping for the past four years. He’s a prime candidate for age-related decline.

But if you look past ERA and past FIP, the numbers paint a picture of a pitcher who remains as steady and reliable as ever. The difference between Lynn’s ERA and FIP is the second-largest among qualified starters, behind only Brady Singer. The difference between his FIP and xFIP is also the second-largest, just trailing Yusei Kikuchi. Consequently, the gulf between his ERA and xFIP is wider than a Little League umpire’s strike zone:

This is definitely a good sign, and while Lynn surely isn’t pleased with his performance thus far, at least his xFIP is right in line with his career average. Four years ago, he had his best season by WAR (6.8) and earned down-ballot Cy Young votes for the first time; his xFIP was 3.85. Two years later, he was a Cy Young finalist and came five innings away from an ERA title; his xFIP was 3.82. This year, his ERA is higher than Mariah Carey’s falsetto, and his Cy Young chances are slim to none. But his xFIP is 3.74.

Then again, xFIP has its limits. When a pitcher has a FIP higher than his xFIP, it tells you one thing and one thing only: His home run-to-fly ball ratio is above league average. HR/FB is highly volatile in a small sample size, and xFIP operates under the assumption that with time, it usually regresses toward league average. To quote the FanGraphs glossary:

The number of fly balls that go for home runs is very sensitive to sample size meaning that over the course of a season, the number of home runs a pitcher allows may be higher or lower than their true talent indicates. This is not to say pitchers aren’t responsible for the home runs they did allow, but rather to say that if you want to judge how well they pitched, xFIP will remove some of those fluctuations in HR/FB% and will give you a better idea.

I find xFIP philosophically fascinating. For all intents and purposes, it’s a version of FIP that uses fly balls as an input instead of home runs. Yet unlike strikeouts, which are inherently good for a pitcher, and unlike walks, HBPs, and home runs, which are inherently bad, fly balls are too complex for such a rigid dichotomy. Most of the time they’re good, but sometimes they’re very, very bad, and some pitchers allow the bad kind more than others. What xFIP presumes is that all fly balls are created equal, and though this assumption is patently false, it’s a useful metric nonetheless. Why? Because the goal of xFIP isn’t to reflect each individual pitcher’s true abilities, but to eliminate extreme outliers. It’s willing to be a little bit wrong about a lot of pitchers to avoid being entirely wrong about anyone. It works because the variance in the true talent of MLB pitchers is so much smaller than the potential variance in HR/FB over a small sample of innings.

To bring this conversation back to Lynn: the gulf between his FIP and xFIP introduces a dilemma. His 20% HR/FB is unsustainable, and no one would dispute that. The difference between his FIP and xFIP rightfully flags his HR/FB as a massive outlier, so he’s due for some serious regression. The big question, however, is just how much regression to expect. It’s hard to watch a guy give up 22 home runs in 18 starts and not begin to wonder if something is amiss, but it’s just as hard to watch Lynn’s dominant starts and think anything is wrong at all. Has he become more susceptible to home runs, or has he simply been the victim of exceptional misfortune?

One way to answer this question is to increase the sample size, which will help determine if Lynn’s HR/FB is anything more than a half-season fluke. As it turns out, it was below league average in eight of his first nine seasons. In 2022, it suddenly shot up, and this year it’s only gotten higher:

Over his last 64 starts dating back to May 2021, Lynn has given up about 400 fly balls. That’s the estimated number necessary for HR/FB to stabilize. He has a 14.4% HR/FB in that time, which isn’t nearly as dramatic as a 20% rate, but it’s still significantly higher than average. In contrast, his HR/FB was significantly lower than average over his previous sample of 400 fly balls. With that in mind, it starts to look like we’re dealing with a worrisome trend instead of a mere fluke. To make things worse, his HR/FB has been steadily rising since 2021:

But not so fast. The troubling trend all but disappears when you split his numbers between home and the road. Lynn spent his first six seasons at the pitcher-friendly Busch Stadium. In that time, his 6.5% HR/FB at home was elite, even for Busch Stadium standards, but his 11.7% HR/FB on the road was right around league average. Since he came to the American League in 2018, his HR/FB ratio has been a healthy 12.5%. When he joined the White Sox in 2021 and started playing at Guaranteed Rake Rate Field, a veritable home run haven, his HR/FB at home skyrocketed, but it was still par for the course on the road.

In other words, Lynn was better than average at preventing home runs in St. Louis, and he’s been worse than average on the South Side of Chicago. On the road, however, his HR/FB has always hovered around league average — until this year, anyway. That being the case, I’m not convinced Lynn has been trending in the wrong direction for his last 400 fly balls. His home run problem this season is new this season, so it’s the numbers from this season that demand our attention.

Take a peek at the number of hard-hit fly balls Lynn is giving up this season. This is the kind of contact that accounts for the vast majority (about 96%) of home runs. Therefore, if a pitcher is giving up too many homers, there’s a good chance he’s giving up too many hard-hit fly balls. But Lynn? Not so much:

How about pulled hard-hit fly balls? Pulled flies account for 61% of home runs, so if a pitcher suddenly starts allowing more pull contact, it could explain an increase in his HR/FB. As fate would have it, Lynn is, in fact, giving up more pulled hard-hit fly balls this year, but the difference is slight:

Clearly, an abundance of pulled hard-hit fly balls isn’t the problem. Instead, the problem is that Lynn can’t catch a break. Of the 16 such balls he’s given up, 15 have left the park. That’s a 93.8% home run rate, compared to the league average of 60%. It’s a similar story on hard-hit fly balls up the middle; he’s allowing home runs at twice the average rate. Given the league-average HR/FB on hard-hit balls pulled or up the middle, Lynn would have allowed nine fewer homers this year, bringing his FIP down from 4.81 to 3.68. That’s a massive difference.

Another interesting wrinkle to the story is that Lynn’s home run problem has only been an issue against left-handed batters and, more specifically, only against his fastballs. Against righties, he has a perfectly normal HR/FB ratio and an excellent 3.26 xFIP. Versus left-handers, however, he has the highest HR/FB in the league, and 15 of the 16 home runs he’s allowed have come on fastballs. The numbers are extreme against both his four-seamer and his cutter, but it’s the former taking the brunt of the impact:

Even stranger still, Lynn’s four-seam fastball has been downright nasty against right-handed hitters. Righties have a .196 wOBA, a .216 xwOBA, and a 40.4% whiff rate against it; it’s been his best put-away pitch against same-handed hitters, and he’s striking them out at his highest rate in years. There aren’t any significant differences in his fastball’s movement profile or his release point, and his velocity is only down by half a mile per hour. Neither Stuff+ nor PitchingBot thinks his fastball is any worse this year than last. Thus, while the hyper-specificity of Lynn’s home run troubles could point to a very peculiar problem, it could also indicate this is one big fluke. After all, if something really wasn’t right, wouldn’t you expect a couple more symptoms to show up?

When a pitcher can’t stop bleeding runs, it’s natural to presume he’s doing something wrong. Yet for Lynn, his 3.74 xFIP is a bright light at the end of the tunnel. As awful as all the home runs look, there’s no deeper indication that anything is seriously wrong. Lynn’s FIP is dreary, and his ERA is grim, but luckily, there’s relief in the underlying metric’s underlying metric. He should be just fine.