Ratings Recalc Advice

[fauteuil7]

I'm starting a league in 1901 using the unmodified Lahman database and would like to get results that do not directly mirror history - maybe have some unexpected players do well while having some big names flop. I'm intrigued by the "recalculate player ratings" feature and was wondering if anyone was using it, and if so, if they have it set at 1, 3, or 5 years, and what kind of results that gives.

[redsox45]

fauteuil7,

Try using 3 year recalc. The game is set to default with it where the results have been the most reliable.

You could also try my 1901 Historical Template which is optimized for accuracy.

Will

[redmarkYankees]

If you want an alternative history, turn recalc off. With recalc, players who had injuries cut their careers short have their ratings plummet with recalc; if you have injuries on, players who historically didn't have their careers cut short, do. Turning recalc off, players get imported according to the option you select (career, peak, etc), but then develop with the OOTP engine, giving some variability.

In my current test reaching 1984, Tony Conigliaro tops the HR table (with Babe Ruth only 9th), Joe Torre and Babe Herman are in the top 5 VORP, 'Smoky Joe' Wood avoided injury and tops the pitching VORP table, while Bob Grim (who started only 60 games IRL) is a 337 game winning twice Cy Young award winning HOF'er.

[Ksyrup]

I really don't like the recalc option personally, but one thing I'm struggling with is how to deal with Babe Ruth. Right now, I'm just starting in 1919 or 1920 so that he's a hitter and his pitching career is documented, but is there some decent way to deal with him if I'm not going to recalc? Also, I tried 1 sim from 1901 with all of the default era settings the game imported, and Crawford had nearly 30 HRs. Is this unusual, or are the HR totals off for those of you simming from early 1900s?

[redmarkYankees]

Originally Posted by Ksyrup

I really don't like the recalc option personally, but one thing I'm struggling with is how to deal with Babe Ruth. Right now, I'm just starting in 1919 or 1920 so that he's a hitter and his pitching career is documented, but is there some decent way to deal with him if I'm not going to recalc? Also, I tried 1 sim from 1901 with all of the default era settings the game imported, and Crawford had nearly 30 HRs. Is this unusual, or are the HR totals off for those of you simming from early 1900s?

I had the same; ideally, his pitching ratings need boosting a little once imported to avoid him getting dumped (having minors instead of just a reserve team also helps). I think 1918 is the year he pitched very little, so would be the time to switch him to a hitter and boost his ratings - not yet all the way, though, as he will hit 50HRs a couple of years too early.

edit: I've not had that HR issue, may just be a random blip.

[statfreak]

Originally Posted by redmarkYankees

With recalc, players who had injuries cut their careers short have their ratings plummet with recalc

I believe this is an inaccurate statement. From what I read on the main page, players will develop using the OOTP engine if /when no statistics exist in the database. So, players with injury-reduced careers or went to war and missed a couple seasons will develop using the OOTP development engine even with recalc on.

[redmarkYankees]

Originally Posted by statfreak

I believe this is an inaccurate statement. From what I read on the main page, players will develop using the OOTP engine if /when no statistics exist in the database. So, players with injury-reduced careers or went to war and missed a couple seasons will develop using the OOTP development engine even with recalc on.

From what I've seen so far, your statement is certainly accurate for war-players (which I wasn't discussing) and players who suffered a quick 'terminal' injury (but even then, perhaps only when it was in the offseason or late season).

Many injured players don't simply have "no statistics" in the database - they have 1 (or more) years of reduced appearances (and usually poor stats). What seems to be happening is that if one of these years coincides with a recalc year (every year, if set to yearly recalc), then their ratings are affected. I suspect also that the 'adjust for less than 200 ABs, weaken for less than 50 at bats' from the import setup may be applying on recalc, so that severely injury hit years see the player nerfed entirely: so much so that he is then cut and retires at about the same time as IRL, despite the way in which recalc in the absence of stats is supposed to work.

[rasnell]

The most significant plus about recalc is that it keeps single season and career leaderboards within realistic numbers like no previous version we've ever had of OOTP.

It does a great job of catching the real swings with such players as Bernie Carbo and Johnny Bench who had some odd swing seasons right in their prime due to injury or illness.

It keeps Babe Ruth within real limits and you won't see unreasonable .400 hitters or 61+ HR until the right times and eras.

However, if you want true discrepancies and a real alternate universe, turn recalc off and allow the automatic league modifiers to at least keep league totals in check. But you could end up with more .400 hitters than you want and Ruth cranking 80 HR in his monster season.

What's great is that you have so many options and settings to test for yourself until you get the flavor that you like.

[redmarkYankees]

Originally Posted by rasnell

However, if you want true discrepancies and a real alternate universe, turn recalc off and allow the automatic league modifiers to at least keep league totals in check. But you could end up with more .400 hitters than you want and Ruth cranking 80 HR in his monster season.

So far with recalc off, I've found the overall numbers to be ok and few monster seasons, but just from different individuals. In my last test, Rickey Henderson developed power in additional to AVG, steals and longevity and 2nd all-time VORP; while A-Rod took talent hits in his rookie year and was retired by after 3 seasons.

This is ok from one perspective; what I'd like to do is also be able to keep recalc, but avoid it's impact on some injured players.

[statfreak]

Originally Posted by redmarkYankees

From what I've seen so far, your statement is certainly accurate for war-players (which I wasn't discussing) and players who suffered a quick 'terminal' injury (but even then, perhaps only when it was in the offseason or late season).

Many injured players don't simply have "no statistics" in the database - they have 1 (or more) years of reduced appearances (and usually poor stats). What seems to be happening is that if one of these years coincides with a recalc year (every year, if set to yearly recalc), then their ratings are affected. I suspect also that the 'adjust for less than 200 ABs, weaken for less than 50 at bats' from the import setup may be applying on recalc, so that severely injury hit years see the player nerfed entirely: so much so that he is then cut and retires at about the same time as IRL, despite the way in which recalc in the absence of stats is supposed to work.

so recalc is functioning the way it should... I mean, those that use recalc want results similar to real life and they get that. When players are injured or ineffective, recalc accurately models that. To prevent this, either turn recalc off or modify the database removing injury-shortened seasons.

[redmarkYankees]

Originally Posted by statfreak

so recalc is functioning the way it should... I mean, those that use recalc want results similar to real life and they get that. When players are injured or ineffective, recalc accurately models that. To prevent this, either turn recalc off or modify the database removing injury-shortened seasons.

Hmmm. I didn't say recalc wasn't functioning the way it should; I said it was nerfing (some) injured players - i.e. those that don't have a simple, sudden, terminal injury. You said that was inaccurate, but then agree that it is accurate but go on to tell me it's functioning correctly, when I didn't say it wasn't...

Now, where I think this isn't working as some historical what-if simmers might want/expect is that:

a) (some) players who were injured IRL get nerfed in game, without an injury, and will retire at about the time they did because of the injury, due to catastrophic ratings hits
b) if you have injuries ON, some players who didn't get injured IRL do get injured in game (obviously)

So, the only real options currently are historical replay, with injuries OFF, which should be very close to IRL - OR recalc off, with rather random results (e.g. Arod has 3 months in the majors, ratings fall, retires).

What I personally would like to have is recalc ON, with roughly historical results (less random drops for actual stars, less bit-part players becoming all-time greats - or Rickey Henderson developing power, hitting 700 HRs and 2nd all time in VORP), but without the players who were injured IRL being injured (or nerfed regardless) in game. Then I can turn injuries ON, and have Gehrig play for 20 years and get the all-time hits record, while Babe Ruth is injured in his rookie year and never heard from again.

I agree that (with the current setup, at least) the only way to do this is to adjust the database files. I was making the point that the assertion that IRL injured players will be ok in OOTP2007 on the in-game development model is not infact accurate, for many of them.

[Ksyrup]

I'm in 1995 simming in recalc mode with Redsox45's template. The results have been excellent.

HOWEVER, my biggest issue with the recalc option is that the changes that occur in the game based on what happened IRL have no context in the game. Guys switching from SP to RP, guys switching positions, etc. - none of it makes sense given what it occurring in the game (or if it does, it is purely coincidental). I've been discussing this at FOFC since before the game came out, and now I have some real examples of my concerns. Consider the following (from my post over at FOFC):


"Mike Maddux came up with the Phillies as a starter. IRL, his first year was the only year he was exclusively a starter, and he sucked. From that point forward, he was a reliever and spot starter.

In my sim, he also sucked his first year as a starter, but they kept him as a starter in 1987 and he went 11-10 with a 2.93 ERA. The next year, he became a reliever and didn't start a single game, seemingly based solely on the recalc which made him a reliever. However, the 1988 Phillies lost 92 games and certainly could have used a starter who the previous year had an ERA under 3, instead of throwing a ton of starts at a few crappy pitchers."


I haven't had a chance to dig into the results of my sim yet, but I'm curious about what happens with the Eckersley/Smoltz switches, or positional switches for guys like Biggio and Yount. The AI has no idea these guys are going to be switched, so what if the Astros AI signs an all-star second basemen the year before Biggio switches? What if the Braves sign a closer before Smoltz switches? The game should take this stuff into account before deciding whether to make the switch.

If it results in poor performance, at least that would be a contextual reason to move the guy - if Smoltz falters as a starter because during those couple of years he's only got the stamina of a reliever, then fine, move him after he demonstrates he's no longer a quality starter. But moving guys for no reason - especially when it might conflict with what the AI has done with the team - is my biggest issue with the recalc option.

[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.

[statfreak]

Originally Posted by redmarkYankees

Hmmm. I didn't say recalc wasn't functioning the way it should; I said it was nerfing (some) injured players - i.e. those that don't have a simple, sudden, terminal injury. You said that was inaccurate, but then agree that it is accurate but go on to tell me it's functioning correctly, when I didn't say it wasn't...

Ah, I read too much into your initial post. I think we are both saying the same thing. I like that it works that way, but see why recalc could be a cause of concern for some people...

[Ksyrup]

When I get up to present day and want to continue a career from an historical sim, am I correct in assuming I need to disable the recalc option (or does it do it automatically)?

[statfreak]

Originally Posted by Ksyrup

When I get up to present day and want to continue a career from an historical sim, am I correct in assuming I need to disable the recalc option (or does it do it automatically)?

you shouldn't need to disable it. Since the game won't find any stats to use, it will just use the internal development engine... actually, I am not sure how long it takes the game to access the database to check if any stats are available to use in the recalc process, so disabling might see a small performance gain...

[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.

[fauteuil7]

Point taken, but one wrinkle I like to work in is to have AI teams use 5-man staffs with useful bullpens from the first year of play, so I have my ammy draft in March to get some extra pitchers for each team. I know it brings players in a year early, but it makes it easier to compare pitchers' accomplishments over time. I suppose that's what I'm getting at, and have been talking around. I would like to be able to look at Ty Cobb's stats and Tony Gwynn's stats and have the era they played in not be a factor. Same for Ruth, Koufax, and Randy Johnson. I don't know if this would be possible without heavily editing the historical database.

[rasnell]

Originally Posted by statfreak

I believe this is an inaccurate statement. From what I read on the main page, players will develop using the OOTP engine if /when no statistics exist in the database. So, players with injury-reduced careers or went to war and missed a couple seasons will develop using the OOTP development engine even with recalc on.

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.

[rasnell]

Originally Posted by statfreak

you shouldn't need to disable it. Since the game won't find any stats to use, it will just use the internal development engine... actually, I am not sure how long it takes the game to access the database to check if any stats are available to use in the recalc process, so disabling might see a small performance gain...

Correct. Player recalc is so fast that you won't notice.