Quote:
Originally Posted by joefromchicago
I completely agree. Right now, you can sorta' do that by making the rotation size equal the staff size. Obviously, that only works for staffs that have six or fewer pitchers, since OOTP's max rotation size is six. But, as I mentioned in my post, the league rotation size is also connected to pitcher stamina, so a bigger league rotation size means that pitchers recover from starts more slowly. That leads to ahistorical results where the pitchers at the bottom of the "rotation" get too many starts.
|
Wanted to pull this back... I think there are two specific, actionable things here:
1. Allow the AI to use rotation sizes other than the one set in the game. In the fictional 1970 "replay" I'm doing, I'm getting good mileage from having the game set to a 4 man rotation and then setting many/most teams to 5 man rotations with the schedule as written (which, from the advent of air travel and well into at least the 70s and maybe even the 80s, summer schedules were absolutely insane at times - 8 games a week (including a doubleheader), an off-day every 3 weeks at times) and then flipping them into a 4 man due to injury, the time of the season (the schedule is clearer in April and September, and of course with the latter you'll often want to consolidate your rotation anyway). I of course have to do all of this manually but the stamina actually works.
2. As a stretch goal, eliminate that set rotation size altogether and have it become one of the things that's approximated in a config file (a number like 4.5 for example) and then perhaps tweaked along with the rest of the settings in the Opening Day sim. Even if you don't do that, I think it would be a good idea to have this set to some number other than "4" or "5" and figuring this out by either the top 10% in games started or (ideally, since this data is available on BBRef) the ratio of starts on short, "normal" (in BBRef terms that's 4 days rest exactly) or "long" rest.
To edify the latter part of the second point, take a look at these numbers from BBRef for a couple of different eras:
Code:
short 4 days long
41.94% 25.81% 32.26%
36.42% 37.04% 26.54%
7.41% 58.02% 34.57%
1.23% 51.23% 47.53%
Those are percentages from 1950, 1970, 1990, and 2010. I don't want to necessarily lock anyone into a formula but I think you can definitely see where 1950 should use a rotation in the low 4s, 1970 in the mid 4s (there I think the fact of a little over a third of the starts coming on exactly 4 days rest is more telling than the increase in the short rest/long rest ratio due to the train vs plane factor I mentioned above), then something right around 5 for 1990 and definitely over 5 for 2010, when roughly half of all starts come with 4 days rest and roughly half with more than that amount.
I'd have gone back further than that but a. box score data isn't available before I think 1915, and b. box score data is also not available for Negro Leagues (additionally I suspect pitchers were used differently in those leagues given radically different schedule sizes and as such even if we did have that data it ought to be used separately).