Lost in all of last night's late-inning festivities was an underratedly good second start by Luis Severino. He looked very much the part of a 21-year-old making his second start in the early innings, but righted the ship and finished up with a 6-inning, 2-run no decision performance. The outing was not as flashy as his debut in terms of results, but in showcasing Severino's ability to work through trouble and still provide length it may have been a better outing. The prevailing thought after the game was that Severino made some mid-game adjustments to fix things and that's how he was able to survive. That was the narrative I was selling in my game recap, so I figured it was worth looking into a little more to see what those adjustments were. Severino's 6 innings split up nicely into 2 equal 3-inning samples. A comparison of the counting statistics shows a drastic difference in results from the first sample to the second. Over the first 3 innings, Severino gave up 2 runs on 6 hits, 1 walk, and 1 hit batsman, and threw 60 pitches. Over the final 3, he gave up no runs on 1 hit, no walks, and no hit batsmen on 37 pitches. It goes without saying that a pitcher is going to throw more pitchers when he has to face more batters, but surely there had to be something he did differently that contributed to the turnaround, right?
There's his pitch location plot for the first 3 innings. If the blues and blacks look too similar to you, that breaks down to 28 4-seam fastballs, 23 sliders, 8 changeups, and 1 cutter. The first thing that jumps out to me here is how much Severino was missing to both sides of the plate with all of his pitches. His changeup was all over the place, which wasn't forcing Cleveland hitters to respect it, and he missed up and in the middle of the strike zone with just many fastballs and sliders as he did way off the corners.
Now there's innings 4-6. That's 19 fastballs, 15 sliders, and 3 changeups. You'll notice that Severino really tightened up his command on the corners of the zone. Only a handful of pitches missed way off the corner, which helped Severino get more strike calls and throw more strikes (59.4 strike % innings 4-6 compared to 58.3% in 1-3). Severino also ditched the changeup in favor of a few more sliders, not a bad decision considering how poorly he located it in the first 3 innings.
I didn't watch the game live and I admittedly don't know enough PITCHf/x stuff to be able to make meaningful determinations on pitch release point differences, so I can't comment on any mechanical changes that Severino made in the final 3 innings. From what I see in the pitch plots, it appears that he found his command and that may have helped him turn things around. But there doesn't appear to be any discernible difference in approach or pitch selection from the first 3 innings to the final, Severino didn't throw a significantly higher number of strikes from sample sample, and he didn't significantly increase his strikeout or swing and miss totals. I hate to say it, but those final 3 innings could have just been nothing more than good BIP luck.
Keep in mind that this is just 1 game and 1 quick look from a very high level. If I had the time to break it down into 2-strike performance and results against lefties versus righties, maybe I could turn up something more concrete. But from this high level, there isn't much to support the idea that Severino made some big change to change his results from the first 3 innings to the final 3. There are definitely some small changes and those changes likely had some positive impact, but in a 1-game sample it's difficult to say how impactful those small changes were. What's important is that Severino stayed in the game, limited the damage, and gave his team a chance to win.