OOTP Developments Forums - View Single Post

pstrickert · 01-26-2009, 04:14 PM

For the last few years, I've tried to get a feel for historical leagues by simming the 1974 season repeatedly. I thought I would share a few observations and, hopefully, start a conversation about it. (NOTE: I use 1-year recalc with a couple of adjustments to the LTMs -- reducing SBA to .670 and increasing errors to 1.300. I could tweak the other settings, I suppose, but these two stats need the largest corrections. BA and ERA results compare favorably with real life.)

IRL, the Dodgers won the NL West by 4 games over the Reds. L.A. had 102 wins that season. The Pirates won the NL East with 88 wins. The fourth place finishers, the Expos, trailed the Pirates by only 8-1/2 games. So, it was a competitive division race. In the AL East, Baltimore (91-71) nipped New York by 2 games. In the AL East, Oakland (the eventual World Series winners) topped Texas by 5 games. California finished last at 68-94.

For the most part, the simmed 1974 season produces credible results. I do, however, see two anomalies: the Dodgers underperform and the Angels overperform. So, I've tried to examine these teams more closely, to see if there's something about them that gives the OOTP sim engine trouble. Or, in other words, if these results expose a flaw in the OOTP sim engine.

In the NL West, the Reds often win the division over the Dodgers. That's reasonable, since the Reds won 98 games in 1974. The problem is that the Dodgers often fall to third or fourth. I haven't spent enough time checking for problems, but I have noticed that the OOTP AI never uses Mike Marshall for 200+ innings in relief (as IRL). He seldom even reaches 100 IP. That, presumably, could affect the late inning results for L.A.

I have spent a lot more time looking at the California Angels. They had a decent pitching staff in 1974, but a below-average offense. In fact, they had the lowest SLG in the AL in 1974. When I sim, I notice that SLG is right where it should be -- at the bottom. So, that's OK. For some reason, though, the team steals a lot more bases (and more successfully) than IRL. This boosts their run production for the season. (IRL, they scored the fewest runs per game in the AL.)

Perhaps the biggest issue has to do with pitching, though. The staff ERA is usually lower than IRL when I sim. Often, the staff ERA is the lowest (or second-lowest) in the AL when I sim. Looking into this further, I noticed that IRL, Bill Singer (one of the SPs) spent a lot of time on the DL. Also, Rudy May, who had an 0-1 record with CAL in real life, was purchased by the Yankees in June. With NY, he put up excellent numbers as a SP. (He has terrific ratings, as a result, in the 1974 sim.) When I sim the 1974 season, then, Singer often stays in the rotation the whole year. May, too, becomes a main cog in the starting rotation. These two differences from real life make the "sim" California pitching staff a formidable one. It is enough almost invariably to put them in contention for the division title.

What do we do about that? I think it's OK to sim the 1974 season and ask, What if? What if California had these two pitchers in the rotation for the whole season? The staff would have rivaled Baltimore's and Oakland's as the elite of the AL. But I wonder if Markus should implement an option to use real-life transaction files (which, BTW, can be found on the Internet). I wouldn't mind additional options to use as-played lineups or most-common lineups (also available on the Internet). This would give the user more flexibility in historical league play.

I'd be glad to hear about your own thoughts and observations.

01-26-2009, 04:14 PM	#1
pstrickert Hall Of Famer Join Date: Dec 2005 Posts: 16,545	Historical League Observations For the last few years, I've tried to get a feel for historical leagues by simming the 1974 season repeatedly. I thought I would share a few observations and, hopefully, start a conversation about it. (NOTE: I use 1-year recalc with a couple of adjustments to the LTMs -- reducing SBA to .670 and increasing errors to 1.300. I could tweak the other settings, I suppose, but these two stats need the largest corrections. BA and ERA results compare favorably with real life.) IRL, the Dodgers won the NL West by 4 games over the Reds. L.A. had 102 wins that season. The Pirates won the NL East with 88 wins. The fourth place finishers, the Expos, trailed the Pirates by only 8-1/2 games. So, it was a competitive division race. In the AL East, Baltimore (91-71) nipped New York by 2 games. In the AL East, Oakland (the eventual World Series winners) topped Texas by 5 games. California finished last at 68-94. For the most part, the simmed 1974 season produces credible results. I do, however, see two anomalies: the Dodgers underperform and the Angels overperform. So, I've tried to examine these teams more closely, to see if there's something about them that gives the OOTP sim engine trouble. Or, in other words, if these results expose a flaw in the OOTP sim engine. In the NL West, the Reds often win the division over the Dodgers. That's reasonable, since the Reds won 98 games in 1974. The problem is that the Dodgers often fall to third or fourth. I haven't spent enough time checking for problems, but I have noticed that the OOTP AI never uses Mike Marshall for 200+ innings in relief (as IRL). He seldom even reaches 100 IP. That, presumably, could affect the late inning results for L.A. I have spent a lot more time looking at the California Angels. They had a decent pitching staff in 1974, but a below-average offense. In fact, they had the lowest SLG in the AL in 1974. When I sim, I notice that SLG is right where it should be -- at the bottom. So, that's OK. For some reason, though, the team steals a lot more bases (and more successfully) than IRL. This boosts their run production for the season. (IRL, they scored the fewest runs per game in the AL.) Perhaps the biggest issue has to do with pitching, though. The staff ERA is usually lower than IRL when I sim. Often, the staff ERA is the lowest (or second-lowest) in the AL when I sim. Looking into this further, I noticed that IRL, Bill Singer (one of the SPs) spent a lot of time on the DL. Also, Rudy May, who had an 0-1 record with CAL in real life, was purchased by the Yankees in June. With NY, he put up excellent numbers as a SP. (He has terrific ratings, as a result, in the 1974 sim.) When I sim the 1974 season, then, Singer often stays in the rotation the whole year. May, too, becomes a main cog in the starting rotation. These two differences from real life make the "sim" California pitching staff a formidable one. It is enough almost invariably to put them in contention for the division title. What do we do about that? I think it's OK to sim the 1974 season and ask, What if? What if California had these two pitchers in the rotation for the whole season? The staff would have rivaled Baltimore's and Oakland's as the elite of the AL. But I wonder if Markus should implement an option to use real-life transaction files (which, BTW, can be found on the Internet). I wouldn't mind additional options to use as-played lineups or most-common lineups (also available on the Internet). This would give the user more flexibility in historical league play. I'd be glad to hear about your own thoughts and observations.