Ian Anderson’s First-Inning Issues: Fact or Fiction?

Ian Anderson had a rough go of first innings this year. In 24 games, he compiled a 6.38 ERA, and this isn’t some case of a pile of seeing-eye singles doing him in. He allowed 2.3 home runs per nine innings, walked 14.2% of the batters he faced, and generally let the offense do what they wanted. His numbers worked out to a 5.29 FIP, which hardly seems like a fair representation of his skill. In every other inning he pitched this year, he was comparatively excellent: 0.9 home runs per nine, an 8.8% walk rate, and sterling run prevention numbers (2.93 ERA, 3.71 FIP).

I hate to use the word “narrative” because it’s mostly a lazy baseball writer crutch, but it’s unavoidable here: the narrative that Anderson is vulnerable in the first inning and bulletproof afterwards has been omnipresent in his playoff starts. When he gave up a first-inning home run to Corey Seager in Game 2 of the NLCS, it was something obvious to point to. First inning? Must just be Anderson’s unique flaw, a magic spell that makes him terrible until he gets a chance to grab some sweet dugout pine.

If you can’t tell from the way I’ve described it, I’m skeptical. Splits like that feel too hand-wavy, too post hoc ergo propter hoc, if you’re into being pretentious like I am. Every pitcher has to be worst in some inning, by sheer chance alone, even if their true talent never wavers. If I had my druthers, I’d just ignore the whole thing and go back to watching dugout celebrations. But because the first inning is the first, rather than the third or the sixth or any other random number, I thought I’d do an investigation into how much we should believe it.

Let’s start, as many good investigations do, by considering the initial conditions we’re working with. The first inning is unique for several reasons, but the principal separating factor is that it’s the only inning where the batting team gets to choose who hits. Take Seager, Anderson’s most recent nemesis. In 2021, he has 73 first-inning plate appearances. His next highest total is 49 (a tie between the third and fifth innings, if you must know). Contrast that with a relatively weak starter, like Kevin Newman, who batted .226/.265/.309 for the Pirates in 554 plate appearances. His appearances are more spread out — he came to the plate 72 times in the fifth and eighth innings, but only 54 times in the first.

Put another way, teams aren’t dumb. They put their best hitters at the top of their lineups, and that means opposing pitchers face a stiff test right from the jump. Teams produced a 107 wRC+ in the first inning in aggregate this year, the highest for any inning. That’s despite facing starters rather than middle relievers. It’s also despite facing those starters when they’re at their freshest and for the first time through the order. The top of lineups, as it turns out, are mostly filled with good hitters.

Of course, that’s not enough to explain Anderson’s problems. The tops of lineups are good, but not that good. Teams scored at a roughly five-runs-per-nine-innings pace in the first inning this year. Anderson allowed runs at a 6.4-per-nine clip. Meanwhile, he was better than the average starter after the first inning. Is there really something here?

Before you decide one way or another, consider this. Alex Cobb was the best starter in baseball in the first inning this year. He allowed a piddling 0.50 ERA, with a 1.78 FIP to match. That’s elite production, obviously, but despite showing those good hitters who’s boss, Cobb was only okay after the first. He had a 4.54 ERA thereafter, and the peripherals all agree. He gave up more walks, ran a lower strikeout rate, and surrendered 0.6 extra homers per nine innings. Is Alex Cobb some secret master of the first inning, an opener who should pitch the first four times a week and then hit the showers?

Don’t answer too quickly, because I’ve got more information. In 2020, Cobb was the 18th-worst starter in the first inning. He pitched to a 5.40 ERA and a 5.79 FIP. He gave up nearly two home runs per nine innings (he didn’t give up a single one in the first inning this year). He walked nearly as many batters as he struck out but then improved markedly thereafter, times-through-the-order penalty and all.

On the one hand, that should be obvious. Cobb only faced 44 batters in the first inning last year. He only faced 72 this year. Anderson made more starts, but he still only faced 113 first-inning batters. Focusing on these subsets of a pitcher’s output will naturally create more outliers. In 100 plate appearances, a hitter can be Babe Ruth or Alcides Escobar. A better question than “did Anderson get lit up in the first inning?” is “will Anderson continue to get lit up going forward?”

First, let’s answer a similar but different question: do pitchers with very bad or very good first-inning marks in one year carry those characteristics over to the next year? I took data from the 2016-19 seasons (2020 is such an outlier that I’d just as soon not consider it) to find out. I took each pitcher who threw 20 innings in the first inning and 100 innings overall and looked at three paired years: 2016-17, 2017-18, and 2018-19. I then compared their first-inning ERA to their overall ERA in each pair of years.

The pitchers with the best relative first-inning ERAs were great at shutting down the opposition. In 2016, for example, the 20 pitchers with the lowest relative ERAs had an ERA 1.38 points lower than their overall mark if you only look at first innings. In 2017, however, their ERAs were actually 0.27 points higher in the first inning compared to their overall line. The same is true on the other side of the distribution; the 20 pitchers with the worst relative first-inning performance were a whopping 2.05 worse in the first. In 2017, those same pitchers were 0.27 points better in the first inning.

The same effect held between 2017 and ’18 (the “best” first-inning pitchers in ’17 were roughly average in ’18, as were the worst), and also in the 2018-19 pair. In aggregate, there’s no correlation whatsoever — the r-squared between relative first-inning talent in year one and year two is 0.003. In other words, none of the variation in second-year variation is explained by first-year variation.

I’ve tricked you a little bit. I feel quite certain that the overall phenomenon of “first-inning specialness” is just noise. In aggregate, what pitchers do in the first inning relative to their overall performance in one year says nothing about what they’ll do in the next year. In other words, first-inning performance isn’t predictive. Just because that’s true in the aggregate, however, doesn’t mean it’s true for every player.

For example, what if Anderson came up in the first inning throwing 85 mph, then ramped it up to his normal 95 thereafter? He’d likely put up terrible first-inning numbers, and he’d roundly deserve them. All I’ve proven with the part above is that we shouldn’t use Anderson’s high first-inning ERA as evidence that he has a real first-inning problem. Let’s look into the specifics to see if there is any evidence.

Fastball velocity? That’s a big fat no. Anderson averages 95 mph on his four-seamer in the first inning, 94.4 mph thereafter. His secondary pitches are correspondingly faster. That’s not strange — most pitchers are fresher in the first inning — but we’re looking for reasons Anderson might be worse. This seems unlikely to be the culprit.

Movement? All three of his pitches moved roughly the same in the first inning and in the rest of the game. That holds true whether we’re talking total movement or breaking it up directionally. His spin rates are also steady. His release point is quite consistent as well. There’s nothing clearly wrong with the shape of his first-inning pitches.

Location? It’s not a meatball issue. He throws 22.4% of his curveballs over the heart of the plate in the first inning, as compared to 23.5% thereafter. He throws an equal 28.7% of fastballs over the heart in the first inning and all other innings. Changeups? It’s 21.7% over the heart in the first inning, 21.2% afterwards.

Nor is it an issue of Anderson losing the ability to throw competitive pitches. He’s given up a few free balls on his heater, but not many: 28% of his first-inning fastballs are in the chase or waste zones, as compared to 27.2% afterwards. That’s basically two fastballs all year, and his changeup is similarly on point. If there’s a problem, it’s bouncing his curveball. He’s “wasted” 24.5% of them in the first this year (18% in other innings).

I’m somewhat at a loss as to what’s truly vexing Anderson, which suggests there might not be much to find. If I had to guess the most likely reason, I’d say that he’s been less in control of his curveball in the first, which gets him behind in counts slightly more often. The big problem seems to be home runs, with walks close behind, and pitching behind in the count can lead to both of those. One throwaway stat that supports this: in 1-0 counts, he threw a pitch in the strike zone only 41.2% of the time in first innings this year. In all other innings, that number is 60.1%. Whether it’s mechanical or mental, he seems to nibble more in first innings.

For the most part, though, I think it’s all a crock. Why is Ian Anderson bad in first innings? Mainly because batters are trying their best to beat him throughout the game, and sometimes they win. The next time Anderson makes a start, whether it’s in this series or the World Series, I’ll be watching his first inning and wondering whether he’ll give up a home run. It won’t be because of some mysterious lack of first inning talent, though. It’ll be because the guy at the plate drives a Benz, too.