Pitchers - early retirement theory (early history)

[redmarkYankees]

I'd be interested in input from others with some significant early history sim stats to look at. I think too many pitchers may be retiring early because of the low 'Stuff' ratings.

The theory would be that while the 'stuff' rating is by design very low for early pitchers, it is being over-used in whatever calculations are used by AI GMs in determining pitcher roles and/or by pitchers themselves in determining when to retire. Changing the pitcher ageing figure to 0.900 in my latest sim appears to have had no impact on improving pitcher longevity.

Looking through retired quality SPs, a number appear to have very suddenly retired, or lost rotation spots, following a very small fall in their stuff rating (often in their early/mid 30s, but I think it even happens on random small falls at any age) - but only from 4 to 3, or some similar insignificant amount when they still have movement/control of 17-20.

For instance:

Pud Galvin - a Rookie of the Year, and 5 times Pitcher of the Year: sometime early in 1886 his 'stuff' rating dropped from 5 to 3 (only just noticed SISA updates with scouting off are only 1st Jan, so impossible to tell retrospectively exactly when a ratings boost/hit occurred). His movement was still 20, his control was still 20. Regardless - he was traded for a very average hitter, and was barely used in relief. He was 29. In the following year, he appears to have been demoted to the reserves, not throwing a pitch in the majors. He retired, still with movement and control at 20, before his 31st birthday.

[JWay]

so is there a way to test it, or will the database need to be reworked?

[Lastplaneout]

Would increasing Stuff lead to unrealistically inflated strikeout totals in Deadball-era leagues?

[JWay]

well you could counter it too with league modifiers though couldn't you?

[Qrusher14242]

I've seen this too. Frank Viola had a great career in Minnesota from 83-87. Then he became a FA and signed with Oakland. In 88 he had another good year with Oakland and then the next year his ratings dropped 10 points. He was prompty dropped out of the rotation(at 29) and had a terrible year in relief and retired.

[redmarkYankees]

Originally Posted by JWay

so is there a way to test it, or will the database need to be reworked?

It would need changes to the era_modifiers.txt file, to test and apply to rookies year after year. Alternatively, could perhaps be tested with purely fictional players by experimenting with the stuff modifier in game.

One thing I don't understand - if the era_modifiers.txt is applying a strong reduction on 'stuff' ratings (e.g. approx 0.4), isn't it doing that to imported players who should already have low stuff ratings based on their stats? If the whole point is to generate ratings which mirror real life stats, doesn't a heavy modifier 'double-count' that effect? Also of course the heavy additional values for movement/control.