Carlos Martinez St. Louis CardinalsZach Gifford | THE INTREPID STL

In their first game after the non-trade deadline, the Cardinals turned to number one Carlos Martinez. He got off to a rough start before settling in, giving up three runs in the first and none after. Unfortunately, those three runs were enough, and the Cardinals dropped back to two games below .500.

This game continued two troubling trends. For one, the lack of run support for Carlos Martinez. The second, and the one I’ll focus on today, is Martinez struggles in the first inning. Through his August 1st start, Martinez has a 6.55 ERA in the first inning and 3.04 ERA afterward. We can’t really blame bad luck here, since he also has 6.29 FIP in the first inning and 3.36 FIP afterward.

Now, there is a leaguewide first inning penalty to take into account. Generally, you’re always facing a team’s best hitters in the first, so you’d expect a starter’s numbers to look a little worse in the first than they do afterward. Martinez, though, has the 19th largest difference between his first inning ERA and second+ inning ERA among 153 pitchers with 10 starts. His FIP differential is the 8th largest.

So, what’s causing Carlos Martinez first inning struggles? His raw stuff doesn’t look worse in the first inning than it does afterward: his pitch velocity, movement, and spin rate are all about the same in the first as they are from the second inning on. His Statcast Zone% is 49.0% in the first and 51.5% afterward, so maybe he has trouble locating early on (although his pitch heatmaps don’t indicate any dramatic differences). What about pitch selection?

Carlos MartinezZach Gifford | THE INTREPID STL

In the first inning, Carlos Martinez turns to a fastball on 64.7% of his pitches. Hitters can sit on the fastball and be right about two-thirds of the time. Sure enough, hitters own a .375 batting average, .475 slugging percentage, and .435 weighted on-base average (wOBA) against Martinez fastballs in the first inning.

After the first inning, Carlos Martinez drops his fastball usage down to 55.1% and drastically increases his slider frequency. The result is a nearly even mix of fourseamers, sinkers, and sliders, with a healthy dose of changeups mixed in. Martinez changes speeds by more than 8 mph 35% of the time in innings 2-9 compared to only 30% in the first, and that doesn’t movement into account. As a result, hitters are less able to sit on fastballs and it shows: Martinez allows only a .254 batting average, .395 slugging percentage, and .307 wOBA against his fastballs after the first inning.

Listen to any broadcast, and you’ll hear about the importance of establishing the fastball early in the game. Based on his pitch selection, it looks like Carlos Martinez tries to do just that. Based on his pitch selection in past years, it looks like he’s tried to do that every year he’s been a starter. It didn’t work well in 2015, when he had a 4.45 first inning ERA. It appeared to work great last year, when he had a 2.32 ERA. This year, it hasn’t worked at all – maybe hitters have caught on.

Looking at first inning FIP supports the notion that hitters are getting a better book on Martinez in the first inning. In his first eight MLB starts (spanning 2013 and 2014), Carlos Martinez allowed only a 2.25 FIP and 3.48 xFIP in the first inning. In 2015, his FIP jumped to 3.03 and 3.66. Still very good, but a little worse. In 2016, both numbers jumped again to a 3.79 FIP and 4.34 xFIP. Then this year, his FIP and xFIP have skyrocketed, to 6.29 and 5.12, respectively.

There’s a chance that Carlos Martinez struggles in the first inning this year are random. After all, we’re dealing with small sample sizes. There is, however, a possibility that Martinez is too predictable in the first inning. Maybe, instead of “establishing his fastball” early in the game, Martinez should work to keep hitters off balance from the start. With the Cardinals playoff chances slipping away, now would be an opportune time to experiment.

Photo credit: @cardinalsgifs