Quote:
Originally Posted by Matt Arnold
If you're playing with morale and personalities on, then some players might get a bonus for the day for high morale. But those are relatively small factors.
I would say in the OP analysis, they might need to be a little more careful about un-intended relations. For example, the odds of someone throwing a shutout 7th inning after throwing non-shutout innings 1-6 is very low, since if they throw non-shutout innings 1-6, almost certainly they have either been pulled from the game. Similarly, if someone throws a shutout 8th inning, almost certainly that could be because they haven't turned the lineup over as much earlier in the game and gotten fatigued. Never mind some survivorship bias at play as well - if your data had 80% or so of 7th innings with shutouts from starters, that's going to be very biased for the guys who have the talent to make it that far.
As for the point of how many good/bad games a pitcher has, I think you might actually be slightly surprised by actual data in that regard. For example, in 2024, the worst SP by qualified ERA this last year was Corbin, and in 8/32 starts of his this year, he allowed 1 run or less. The 2nd worst, Mikolas, had 6/32. 3rd worst was Canning, who had 9/32 starts with 1 run or less. And all 3 ended up with ERAs north of 5.
Now, granted, the top arms this year seemed to be at 50% or more for 2 ER or less appearances. But we also live in a much more sheltered world than in 1965, where you comparison stands. I'd be willing to bet that the Sale or Burnes of the day instead of getting pulled after 6 innings of 1 run ball might last another couple innings and maybe give up a second run.
How much that would be, it's hard to say. IRL, even Koufax in 1965 had 1 run or less in 19 of his 43 starts. Gibson was 12/38. On the other end, Buhl was the worst qualified SP without a bunch of relief games, and he had 7/32 starts with 1 run or less. Simmons only had 4. Fisher had 6/36. So yeah IRL they seemed to have less for the worst guys. While we do our best, our usage isn't always a perfect match for some of those earlier eras. We certainly might be more aggressive at pinch hitting for a guy before he starts to struggle, or be stricter at pitch counts which they wouldn't have really cared about back then nearly as much.
I forget where I heard it or the exact quote, but I think in a lot of cases, the difference between a bad pitcher and a good pitcher isn't what they do on their best days, but what they do on their average or worst days. Chris Sale wasn't great last year because he necessarily had more great days that other guys. He was great because he just didn't have the bad games that others did. He "only" had 50% of his starts with 1 run or less allowed. But all but 3 of his starts he allowed 2 or less.
|
Very interesting data.
My first thought (and this is related to the question by Rain King) is that going deeper in games was very common in the 60's and before. That is reflected in OOTP by longer stamina, and shows up in how long the AI keeps starters in. 73% of the starts in my data went at least 6 innings and 53% went at least 7 innings.
That means the odds of throwing a shutout through 6 or more innings is more common in 1965. And, they happen for Bad Pitchers, too. Of the 114 Bad Pitcher starts, 14 (or 12% of them) resulted in 9 innings of shutout being pitched. So, any bias in later innings for better starters is likely much less than it is today. (Also, BTW, my "Luck Only" expectations are based to only those who make it to later innings, so if there is a bias toward better pitchers it affects those %'s as well, and can't explain why the %'s are higher in the "Actual Data" column.)
Your brief review of actual MLB data seems consistent with my expectation - everyone has "good days" some of the time, and better pitchers may have them somewhat more frequently than poorer pitchers.