Ratings Recalc Advice

[rasnell]

Originally Posted by fauteuil7

On the point of how recalc deals with starters becoming relievers (and vice versa I suppose, which may be important since so many starters get imported as MR's) this just proves the point that one of the top priorities for a patch, or if it's too involved for a patch, should be to get rid of pitchers' roles.

I agree that OOTP has a hard time figuring out roles. However, I think the important thing is that it really does look at an entire organization and endurance ratings and put pitchers in their proper places on the roster. Despite the oddity of a MR starting and a SP in the bullpen, if you look at the overall team, ratings and endurance, I think the AI is doing far better than previous versions.

What do you think when you get down to the specific results?

[swampdragon]

Originally Posted by rasnell

You're right if they have no stats. But if they have reduced stats, player recalc will reduce them. Remember, though, that the default three-year recalc is best because it smooths out one bad year out of three and averages them into something pretty darn realistic in our hundreds of years of beta-testing sims.

I wonder what would happen if the recalc didn't use the "adjust" and "make bad" options for reduced playing time, and instead compensated by giving less weight to the year in question.

[MizzouRah]

Originally Posted by Comedian2004

Another important issue to keep in mind with the historical simulations is the date of the ammie draft. Using the default, players actually come into the league a year earlier than they did in the pros.

I usually set the draft to take place the next year when I 1st set up the league. By default, Joe Smith will be imported in 1968 and could play that year in the pros. Since Joe never played until 1969 in real life, your 1968 season is tainted by his and everyone elses (rookies) time in the bigs in 1968.

You can set the draft to take place after the all-star game, but then that is before all of the FA signings. I like ti set it the next year, so in 1901, you will have no draft, but in 1902 you will have one of players that came into the league in 1902. Plus, by having the draft in June, they will not get a complete season in their 1st year.

Note: I use the Lahman database.

Call me stupid, but please explain this further.

Thank you!

[rasnell]

Originally Posted by swampdragon

I wonder what would happen if the recalc didn't use the "adjust" and "make bad" options for reduced playing time, and instead compensated by giving less weight to the year in question.

You can turn the make bad option off during the initial import, which gives you a really interesting alternate universe. You truly can have no-name players turn in some truly historic performances.

Almost any idea you can think of, there's a way to modify or customize it within your own game settings.

[rasnell]

Originally Posted by MizzouRah

Call me stupid, but please explain this further.

Thank you!

What WebComedian is saying is that if you leave historic drafts at the default date, it brings rookies into the draft one year ahead of their debut. If you move the draft date back to the post-season, like Nov. 1, it brings their data in from their real first year of stats.

[MizzouRah]

Originally Posted by rasnell

What WebComedian is saying is that if you leave historic drafts at the default date, it brings rookies into the draft one year ahead of their debut. If you move the draft date back to the post-season, like Nov. 1, it brings their data in from their real first year of stats.

So is that what most are doing? I think I'll try that, thanks!

[KCRoyalsFan]

when players are imported into their league, are their initial rookie ratings based on the next 3 years of stats if the base is set to 3 or are they based on their career?

[KCRoyalsFan]

bump, surely someone knows.

[redmarkYankees]

Originally Posted by KCRoyalsFan

when players are imported into their league, are their initial rookie ratings based on the next 3 years of stats if the base is set to 3 or are they based on their career?

Initial ratings appear to be based on the stats from the single year of import (having compared players with an identical debut year, one with increased stats in years 2 and 3, one with zeroes), with a small random element.

edit: OTOH, there are some players whose import ratings suggest it's the 3 year thing. It's difficult to see quite how it works, consistently.

[statfreak]

Originally Posted by swampdragon

I wonder what would happen if the recalc didn't use the "adjust" and "make bad" options for reduced playing time, and instead compensated by giving less weight to the year in question.

I think I assumed something that wasn't happening... using a 3 year recalc you seem to be saying each year is imported by OOTP and then it adds the assigned ratings and divides them by 3... so a injury shortened season has as much effect as a full 162 game season. If that is how OOTP works, well I scratch my head at that design decision...

Wouldn't it make more sense and be more accurate to import 3 years of stats and base the ratings on the cumulative total of those three years?!? So if a player had 600 at-bats in year 1 and 3 and batted .300 and then was injured and only batted .100 in 60 at-bats in year 2, he would get a rating based on his 366 hits / 1260 at-bats = .290 average instead of say ratings of (9 + 1 + 9) / 3 = 6. I realize this is an extreme example, but I am shaking my head as I grasp how the recalc actually works...

[KCRoyalsFan]

Originally Posted by statfreak

I think I assumed something that wasn't happening... using a 3 year recalc you seem to be saying each year is imported by OOTP and then it adds the assigned ratings and divides them by 3... so a injury shortened season has as much effect as a full 162 game season. If that is how OOTP works, well I scratch my head at that design decision...

Wouldn't it make more sense and be more accurate to import 3 years of stats and base the ratings on the cumulative total of those three years?!? So if a player had 600 at-bats in year 1 and 3 and batted .300 and then was injured and only batted .100 in 60 at-bats in year 2, he would get a rating based on his 366 hits / 1260 at-bats = .290 average instead of say ratings of (9 + 1 + 9) / 3 = 6. I realize this is an extreme example, but I am shaking my head as I grasp how the recalc actually works...

I think this is how it works. I have seen pitchers who only pitched a couple of games, but had outstanding numbers and be rated ridiculously high...at least potentials. Maybe not actual ratings.

[swampdragon]

Originally Posted by statfreak

I think I assumed something that wasn't happening... using a 3 year recalc you seem to be saying each year is imported by OOTP and then it adds the assigned ratings and divides them by 3... so a injury shortened season has as much effect as a full 162 game season. If that is how OOTP works, well I scratch my head at that design decision...

Wouldn't it make more sense and be more accurate to import 3 years of stats and base the ratings on the cumulative total of those three years?!? So if a player had 600 at-bats in year 1 and 3 and batted .300 and then was injured and only batted .100 in 60 at-bats in year 2, he would get a rating based on his 366 hits / 1260 at-bats = .290 average instead of say ratings of (9 + 1 + 9) / 3 = 6. I realize this is an extreme example, but I am shaking my head as I grasp how the recalc actually works...

I confess to not being sure how it really works, but my impression is that the years are weighted independently and not cumulatively.

[Matches]

It seems like they should be taken cumulatively. But I'm having trouble figuring out if it does it that way or not. My sense is that a lot of guys are rated a bit high for their first year. I am using a 3-year recalc, for example, and Stan Hack with a 111 contact rating according to the editor. The real Stan Hack had 178 ABs as a rookie and hit .236. This right off the bat suggest to me that first-year players, if you use a multi-year recalc, are including the second and third years. This actually kinda bugs me because it means first year players are often stronger than it "feels" that the should be.

Hack hit .350 in his second season in just 60 at bats which means the ratings for this year should have been adjuted downward as well. He hit .289 in his third year. Taken independently and then averaged, it doesn't seem to me like those years would result in a relatively solid contact rating of 111.

That's just one example I happened to notice, though. For every one of those there seems to be a Hal Trosky, who got a contact rating of 60/250 even though he hit .330 in his second real season.

[edm]

Originally Posted by Matches

It seems like they should be taken cumulatively. But I'm having trouble figuring out if it does it that way or not. My sense is that a lot of guys are rated a bit high for their first year. I am using a 3-year recalc, for example, and Stan Hack with a 111 contact rating according to the editor. The real Stan Hack had 178 ABs as a rookie and hit .236. This right off the bat suggest to me that first-year players, if you use a multi-year recalc, are including the second and third years. This actually kinda bugs me because it means first year players are often stronger than it "feels" that the should be.

Hack hit .350 in his second season in just 60 at bats which means the ratings for this year should have been adjuted downward as well. He hit .289 in his third year. Taken independently and then averaged, it doesn't seem to me like those years would result in a relatively solid contact rating of 111.

That's just one example I happened to notice, though. For every one of those there seems to be a Hal Trosky, who got a contact rating of 60/250 even though he hit .330 in his second real season.

This is a very nice new feature for historical accuracy, but for the future, It might be a good idea to add a separate recalc modifier for first year players. That way we could set it to one, while other players use the default 3 (or 5)