Thoughts on 5-year Recalc Games

[Brad K]

I don't remember the exact numbers but 98 isn't an impossible performance for someone rated for 73. At least that's the analysis Syd posted.

[David Watts]

Originally Posted by Brad K

I don't remember the exact numbers but 98 isn't an impossible performance for someone rated for 73. At least that's the analysis Syd posted.

Just as long as it's not Roger Maris, I don't see why anyone would question it.

[Brad K]

Originally Posted by David Watts

Just as long as it's not Roger Maris, I don't see why anyone would question it.

Yea, that guy didn't "deserve" to hit 61, did he? And since overall he wasn't good enough for the HOF, keeping him out is no punishment. Something else had to be done.

[Brad K]

Originally Posted by Garlon

With 5yr recalc it adds up all the AB or IP in that span. In this game I set AB to 1100/700. Even if a player has a gap season and did not play as long as they had 1100 AB over those 5 seasons in the recalc period they get their full rating. The advantage of multiyear recalc is that you can get a good sample for player ratings.

Real Ty Cobb from 1911-1913 batted .407
OOTP 5yr recalc Ty Cobb 1911-1913 batted .408

Inputs from outside the evaluated years would result in output from the evaluated years being less like to match historic output. There are three rating events, occurring for the 11, 12, and 13 seasons, that each involved input from 5 seasons. That's 15 total season of input. Six of them, which is 40% of the total, are seasons not involving the evaluated years.

.407 vs .408 is a chance event. It happened contrary to the effects of 5 year, not because of it.

[Garlon]

You are not understanding that every player in the league is also getting a 5yr recalc and the results do not happen in a vacuum, they happen relative to the other players in the league with respect to the league totals while the modifiers keep everything in balance.

MLB BA by season

1909: .244
1910: .249
1911: .266
1912: .269
1913: .259
1914: .250
1915: .248

Notice that big spike in league BA in 1911. That is when they introduced a more lively baseball and Ty Cobb and Joe Jackson both batted over .400 that season. Then in 1912 they both nearly did it again with Ty Cobb batting .409 and Joe Jackson batting .395.

Consider 5yr recalc for 1912 which includes the 1910-1914 seasons. Ty Cobb batted .396 over that period, while the league batted .259. For the 1912 season we take the Ty Cobb 5yr recalc and divide by the 5yr league average which represents the rest of the players in the league, then multiply by the 1912 league BA of .269. This becomes (.396/.259) * .269 = .411. In OOTP he actually batted .404 in the 1912 season, which is very close to expectation based on how recalc works.

[Brad K]

While Auto-calc will make the league output match historic regardless of input this does not apply individually to each player.

So a player is rated dead average for the evaluation year. All fine. But if on five year he under performs or over performs the average he doesn't get an average rating for the evaluation year.

[Brad K]

Originally Posted by Garlon

Consider 5yr recalc for 1912 which includes the 1910-1914 seasons. Ty Cobb batted .396 over that period, while the league batted .259. For the 1912 season we take the Ty Cobb 5yr recalc and divide by the 5yr league average which represents the rest of the players in the league, then multiply by the 1912 league BA of .269. This becomes (.396/.259) * .269 = .411. In OOTP he actually batted .404 in the 1912 season, which is very close to expectation based on how recalc works.

You happened to pick a guy whose performance was close to the same relative to the league for those five years. There's your answer. It's not that 5 year recalc gives better results.

[Garlon]

As I said, the players are relative to each other and with respect to the league average. The post was about Ty Cobb, who historically won 12 batting titles between 1907 and 1919. He was 2nd in the league in BA in 1916.

You then said his performance was due to chance. I think you are just not familiar with Ty Cobb.

[Brad K]

Originally Posted by Garlon

As I said, the players are relative to each other and with respect to the league average. The post was about Ty Cobb, who historically won 12 batting titles between 1907 and 1919. He was 2nd in the league in BA in 1916.

You then said his performance was due to chance. I think you are just not familiar with Ty Cobb.

You tried to use Cobb as an example of how good 5 year recalc is. It's an exception not the rule.

[Brad K]

Originally Posted by Garlon

With 5yr recalc it adds up all the AB or IP in that span. In this game I set AB to 1100/700. Even if a player has a gap season and did not play as long as they had 1100 AB over those 5 seasons in the recalc period they get their full rating. The advantage of multiyear recalc is that you can get a good sample for player ratings.

Real Ty Cobb from 1911-1913 batted .407
OOTP 5yr recalc Ty Cobb 1911-1913 batted .408

It just happens to work with Cobb.

If you're sure it's universal, put the games' *.dat files where they can be downloaded.

[Garlon]

I stated that 5yr recalc provides a good sample size for player ratings. If you want 1yr recalc it is there, if you want 3yr recalc it is there, if you want to double weight them it is there, if you want to use precalc modifiers it is there. That is 10 ways to set up your league right there. Set up your league however you want. This thread is an example of 5yr recalc with the parameters stated in the first post.

[Brad K]

Why are you doing this? Again. Please don't.

The fact is for most players the output will not be as close to historic with 5 year as with 1 year. Just say it. People being able to play however they want is irrelevant to this.

[Matt Arnold]

Nowhere did Garlon claim this was the definitive way to run a historical sim. Heck, he didn't even claim if the results were good or bad. He ran a sim using very clearly defined settings listed out, and posted the results for anyone to see. I mean, I don't think Garlon was intending this to be anything close to "the definitive historical" setup, since it's pretty a-historical for McGwire to spend his career on Seattle (and still hit 69 HR!), for Mickey Mantle to be the big figure in the Red Menace, or the White Sox to be the dominant team in the 1920s.

There's lots of very interesting variations that people can run in their sims. They each have their advantages and disadvantages. That's why we leave things open for people, to run their sims in the way that they want. None of the ways is inherently better than another, since everyone will also wildly disagree even on which is the most accurate.

The above is meant for some fun, to click through, see which teams hit their dynasties when. Seeing which teams dominated for a stretch, and who were the key players for them. Kevin Brown leading the Expos to 2/3 World Series. The Blue Jays somehow winning the World Series in their inaugural season (We either need some work on the expansion draft, or they just got crazy lucky). Imagining what storylines you might write about the ghost of Honus Wagner basically creating Babe Ruth, as they overlapped for 4 years on the White Sox, and then Ruth effectively having his breakout batting season in Honus' last career year. Willie Mays desperately trying to get to 700 HR, only to retire at 698. You'd certainly have some questions about how McGwire could have a 62 HR season, be 12 short of Aaron, and decide to retire. And obviously the fun of having Rickey Henderson retire with 1234 SB, I'm guessing ending there because Rickey thought that was a fun number.

[Garlon]

Matt, regarding the expansion draft this isa how I set it up:

Season Rounds Protected Players
1876 16 10
1882 16 11
1892 17 11
1961 28 14
1962 28 14
1969 28 14
1977 30 15
1993 35 16
1998 35 17

I take the number of players total in the league and divide it by the total number of teams that will be in the league after expansion, and this gives me the number of rounds. This way the expansion teams have the same number of players in their organization as the rest of the teams.

I set the protect total low so that teams can protect their starting lineup, rotation, and a couple of relievers. This way any good players they have on the bench can get a starting role on the expansion teams.

I do not do any auto-protect either.

The computer will sometimes tend to protect younger players while letting some of their starting players go unprotected. The expansion teams are pretty good at grabbing those players though. I think it is a matter of the computer trying to protect too much of their future players rather than their current starters. I could see them protecting their top 1 or 2 prospects but they should be protecting their starting players.

My intent was for those expansion teams to be respectable and not necessarily end up in last place, but they are able to grab plenty of unprotected talent.

[Brad K]

There is a lot of words there defending against things I never said but none addressing what I did say.

My statement that five year recalc does not result in output closer to historic output has not been refuted. The desire to ignore the subject is as good a confirmation of its truth as saying it.

[Matt Arnold]

Originally Posted by Brad K

There is a lot of words there defending against things I never said but none addressing what I did say.

My statement that five year recalc does not result in output closer to historic output has not been refuted. The desire to ignore the subject is as good a confirmation of its truth as saying it.

There's lots of very interesting variations that people can run in their sims. They each have their advantages and disadvantages. That's why we leave things open for people, to run their sims in the way that they want. None of the ways is inherently better than another, since everyone will also wildly disagree even on which is the most accurate.

If you like the way you play, that's fantastic. We want everyone to enjoy whatever way they like best, and feel is the most accurate (if that's what you're aiming for). Don't put down other people's play styles, and don't hijack threads.

Mod Note: At this point the hijack has been resolved with a thread split.

[jksander]

Originally Posted by Matt Arnold

Nowhere did Garlon claim this was the definitive way to run a historical sim. Heck, he didn't even claim if the results were good or bad. He ran a sim using very clearly defined settings listed out, and posted the results for anyone to see. I mean, I don't think Garlon was intending this to be anything close to "the definitive historical" setup, since it's pretty a-historical for McGwire to spend his career on Seattle (and still hit 69 HR!), for Mickey Mantle to be the big figure in the Red Menace, or the White Sox to be the dominant team in the 1920s.

There's lots of very interesting variations that people can run in their sims. They each have their advantages and disadvantages. That's why we leave things open for people, to run their sims in the way that they want. None of the ways is inherently better than another, since everyone will also wildly disagree even on which is the most accurate.

The above is meant for some fun, to click through, see which teams hit their dynasties when. Seeing which teams dominated for a stretch, and who were the key players for them. Kevin Brown leading the Expos to 2/3 World Series. The Blue Jays somehow winning the World Series in their inaugural season (We either need some work on the expansion draft, or they just got crazy lucky). Imagining what storylines you might write about the ghost of Honus Wagner basically creating Babe Ruth, as they overlapped for 4 years on the White Sox, and then Ruth effectively having his breakout batting season in Honus' last career year. Willie Mays desperately trying to get to 700 HR, only to retire at 698. You'd certainly have some questions about how McGwire could have a 62 HR season, be 12 short of Aaron, and decide to retire. And obviously the fun of having Rickey Henderson retire with 1234 SB, I'm guessing ending there because Rickey thought that was a fun number.

In other words, why piss in anyone's sandbox, right? lol ... that's what I love about OOTP, there's so much great room to experiment and mess around with fiction and alternate realities.