So you don't use autocalc because "someone told me not to"? Did you actually try it?

[Sweed]

Originally Posted by locuspc

But the autocalc is what makes this result in unrealistic numbers. If players stopped eating their wheaties and large numbers of players without much power entered the league, you would see total home run numbers drop. You wouldn't see the players who still had power suddenly start hitting 80 homers a season, which is what the autocalc forces.

This just isn't true. HR numbers would not drop because AC is tailoring the LTMs to have the number of HR you have "requested" with your LT setting. Adding a "large numbers of players without much power" would simply change the distribution even more.

If you have data that shows by adding large numbers of low power batters the total HRs do go down, while using AC, then... that would be evidence that AC is not working as intended.

[locuspc]

Originally Posted by Sweed

This just isn't true. HR numbers would not drop because AC is tailoring the LTMs to have the number of HR you have "requested" with your LT setting. Adding a "large numbers of players without much power" would simply change the distribution even more.

If you have data that shows by adding large numbers of low power batters the total HRs do go down, while using AC, then... that would be evidence that AC is not working as intended.

I'm talking about what would happen in real life, in the bolded part you quoted. I agree that it would not happen in the game. In the game you would see the players who still had power suddenly start hitting 80 homers a season, as I discuss in the very next sentence. That's precisely the point that I am trying to make, that autocalc diverges from what would happen in real life in a bizarre and unrealistic way. You can say that's AC "working as intended" but it does not make the final outcome any better.

[MathBandit]

Originally Posted by locuspc

I'm talking about what would happen in real life, in the bolded part you quoted. I agree that it would not happen in the game. In the game you would see the players who still had power suddenly start hitting 80 homers a season, as I discuss in the very next sentence. That's precisely the point that I am trying to make, that autocalc diverges from what would happen in real life in a bizarre and unrealistic way. You can say that's AC "working as intended" but it does not make the final outcome any better.

I mean, it is exactly what would happen in real life if the league realized offence was down because of a lack of Wheaties in the minors and decided to juice the ball like they did in 1987, and again in 2019, to make it so every Tom, Dick, and Harry can hit 30 HRs.

[koohead]

Originally Posted by locuspc

So when you're using 2021 numbers with a talent distribution that doesn't look like 2021, autocalc ends up multiplying the HR values a lot, to try to get the top of the distribution to make up for what it's missing in the middle and low end, which then makes those top enders put up once-a-generation numbers every year.
.

Sorry, I need to go back to this earlier post because there is a fundamental gameplay issue that I want to call out. If I used 2021 LTs, I want to see, or expect to see, 2021 results....regardless of talent distribution. If for some reason the league evolved in some weird way that the style of the game itself changed, then the LT's need to be changed with it. Correct that you cannot control talent pool creation, so that tells me that over a 3/5/10 year period, the DISTRIBUTION of talent will work itself out to be equal. My assumption would be that over this time period, the yearly variances in speed/power/control/arm strength end up flattening out. So yes, if I kept the same LT's over this 10 year period, I may see an outlier here and there, but over time the overall talent in the league should be fairly constant and thus those outliers are just that...outliers. Someone played the game at the right time in the right place. That seems pretty realistic to me.
If you want to simulate league evolution, then I would expect that you would need to manually adjust your LTs every few years to simulate human or societal evolution

[Sweed]

Originally Posted by locuspc

I'm talking about what would happen in real life, in the bolded part you quoted. I agree that it would not happen in the game. In the game you would see the players who still had power suddenly start hitting 80 homers a season, as I discuss in the very next sentence. That's precisely the point that I am trying to make, that autocalc diverges from what would happen in real life in a bizarre and unrealistic way. You can say that's AC "working as intended" but it does not make the final outcome any better.

Fair enough on the you meaning real life.

You're correct in saying AC working as intended still leaves an unwanted outcome. Too me you're point would be "more" correct if you said "even though AC is working as intended I am still stuck with an unrealistic outcome".

The trouble, in my eyes is, that the unrealistic outcome is not because of AC. It is in spite of it working correctly. The divergence from real life is not due to AC. It is due to the mix of player ratings. The fix is not and will not come by doing anything to AC. It will come when, as Lukas noted, they come up with a way to tweak player creation without breaking another area. His post let's us know they are aware of the issue and trying to come up with a solution. So there is at least that to take away from your posts reporting the issue.

[locuspc]

Originally Posted by koohead

Sorry, I need to go back to this earlier post because there is a fundamental gameplay issue that I want to call out. If I used 2021 LTs, I want to see, or expect to see, 2021 results....regardless of talent distribution. If for some reason the league evolved in some weird way that the style of the game itself changed, then the LT's need to be changed with it. Correct that you cannot control talent pool creation, so that tells me that over a 3/5/10 year period, the DISTRIBUTION of talent will work itself out to be equal. My assumption would be that over this time period, the yearly variances in speed/power/control/arm strength end up flattening out. So yes, if I kept the same LT's over this 10 year period, I may see an outlier here and there, but over time the overall talent in the league should be fairly constant and thus those outliers are just that...outliers. Someone played the game at the right time in the right place. That seems pretty realistic to me.
If you want to simulate league evolution, then I would expect that you would need to manually adjust your LTs every few years to simulate human or societal evolution

If you have 500 guys hitting 12 homers each, that's 6000 homers. If you have 50 guys hitting 120 homers each, that's 6000 homers. Those will look exactly the same to the league totals, it can't tell the difference between those two different ways of reaching 6000 homers. You can't correct that by changing the league totals.

[md40022]

I've ran auto-calc twice in my entire life.

The first time was after the first completed season in this year's OOTP game. The very following season the entire league wide BABIP broke the all-time MLB record by a mile. The season concluded with a league wide BABIP of roughly .325.

The 2nd time I ran it was after completing the 2nd season. League wide BABIP proceeded to drop to .271..... a .050 point swing league wide. Complete insanity

In both scenarios I ran it manually on opening day as was suggested by users on here. My opinion is it completely over compensated both times that I ran it. In my first season on here league wide offense seemed a little low. I don't have the energy to look up the stats, but it just seemed a little low. Ran auto-calc for the first time ever after that season and BABIP jumped to .325 which shatters MLB records by a mile. Total over compensation. Ran it again after that season and it drops league wide BABIP to .271 which is like the lowest it's been in 100 years or something.

I'm not saying it doesnt work for some or for most people. At the same time though, the only 2 times I've ever used it, it gave me ridiculous results that seemed to be extreme over-compensations from the previous season.

Best thing I can say is if you've never used it before and have been happy with how your league plays out -------- you'd be foolish to mess with it.

[Brad K]

I'm puzzled When people are playing fictional players in the future, why do they expect them to perform like 2015 or whatever?

[1991Twins]

Originally Posted by Sweed

Thanks for posting and adding screens in your other posts to "show your work". It's a lot easier to have a discussion, whether we agree or disagree with each other, when we are all on the same page and not guessing at the data.

Sweed, you are 100% right. I wanted to post data so that we could examine what is going on. I do feel after my test that the number of home runs in a season directly correlate to the top end guys in a league. The higher the season total, the more likely it is that the upper end guys have historical seasons. I wonder if that is because evidence from history is inclined to support this premise. Take the late 1990s and the early 2000s for example. When record numbers of home runs were hit, all of a sudden you have guys like McGwire going for 70, Bonds 73, and Sosa with a bunch of 60 home run seasons. I wonder if the game's distribution curve is modeled after historical data to replicate that.

On a side note, from my test, I have found that using the real stats for the 2022 season helps to keep the high end power guys from setting individual and career records. It seems the league leaders top out at the low 50's and some years end up in the mid-40s. The highest I saw was 57, which seems plausible. Since the stats are based off of the real life season, with approximately 5200 home runs being the current pace, I have finally found the setting that I am happiest with when playing out games.

[MathBandit]

Originally Posted by md40022

I've ran auto-calc twice in my entire life.

The first time was after the first completed season in this year's OOTP game. The very following season the entire league wide BABIP broke the all-time MLB record by a mile. The season concluded with a league wide BABIP of roughly .325.

The 2nd time I ran it was after completing the 2nd season. League wide BABIP proceeded to drop to .271..... a .050 point swing league wide. Complete insanity

In both scenarios I ran it manually on opening day as was suggested by users on here. My opinion is it completely over compensated both times that I ran it. In my first season on here league wide offense seemed a little low. I don't have the energy to look up the stats, but it just seemed a little low. Ran auto-calc for the first time ever after that season and BABIP jumped to .325 which shatters MLB records by a mile. Total over compensation. Ran it again after that season and it drops league wide BABIP to .271 which is like the lowest it's been in 100 years or something.

I'm not saying it doesnt work for some or for most people. At the same time though, the only 2 times I've ever used it, it gave me ridiculous results that seemed to be extreme over-compensations from the previous season.

Best thing I can say is if you've never used it before and have been happy with how your league plays out -------- you'd be foolish to mess with it.

Auto-Calc doesn't know anything about the previous season I don't think, so it's not compensating (or overcompensating) for anything.

When you ran Auto-Calc the first time, what BABIP did you tell it you wanted for the league? What about the second time?

[kidd_05_u2]

Originally Posted by MathBandit

When you ran Auto-Calc the first time, what BABIP did you tell it you wanted for the league? What about the second time?

BABIP can go haywire regardless of what you input below the league totals and especially if you input league totals that are inconsistent with each other.

[Sweed]

Originally Posted by 1991Twins

Sweed, you are 100% right. I wanted to post data so that we could examine what is going on. I do feel after my test that the number of home runs in a season directly correlate to the top end guys in a league. The higher the season total, the more likely it is that the upper end guys have historical seasons. I wonder if that is because evidence from history is inclined to support this premise. Take the late 1990s and the early 2000s for example. When record numbers of home runs were hit, all of a sudden you have guys like McGwire going for 70, Bonds 73, and Sosa with a bunch of 60 home run seasons. I wonder if the game's distribution curve is modeled after historical data to replicate that.

On a side note, from my test, I have found that using the real stats for the 2022 season helps to keep the high end power guys from setting individual and career records. It seems the league leaders top out at the low 50's and some years end up in the mid-40s. The highest I saw was 57, which seems plausible. Since the stats are based off of the real life season, with approximately 5200 home runs being the current pace, I have finally found the setting that I am happiest with when playing out games.

Your theory makes sense to me. On the simplest terms if you increase HR 20% the guy that hit 50 now hits 60. The guy that hit 20 only jumps to 24.
I'm not saying it's quite that simple in the engine (there is some randomness baked it) but it's probably not far off either with the LT and the way LTMs work.

My game was started in v4-2002, with real players, and imported into every version since. I play out every game and am now in my 2040 season. So overtime it's morphed into a fictional game. I've only changed my HR LT once in all these years. OOTP4, in 2002, obviously came with the steroid era HRs and I rolled with those for around 10 seasons, IIRC. Got tired of the high HRs and MLB IRL was getting steroids out of the game. It was time for me to get back to reality in my OOTP dynasty. I didn't think about how much I wanted to decrease HRs, instead I looked for a season where leaders were in the window I wanted. IE anything in the 50's was great and rare. Mid 40's would be a common occurrence. Can't recall what year I pulled my HR LT from but, in the season I used, HR were 11.5% of hits. I reduced my LT for HR to 11.5% of hits, left all other LTs the same, ran AC and voila I had exactly what I wanted. I'm probably 20 to 22 seasons with that "new" setting. In that time I had one guy hit 60 with second best being 54 IIRC, and the "normal" leader coming in low 40's to 48ish. My third-baseman one year won MVP and lead the NL in HR with 38. For clarity I do run AC on opening day every year. Doing that means there isn't going to be much drift in stats. There are times I've contemplated doing it every 2 or 3 years and see what happens but, in my mind, I also think "why mess with something that is working so well?".

The success I've had using AC since it's inception along with how easy it is to use is why I started this thread. I cringe when I see blanket "don't use it" based on "another user told me x,y, and z".
By all means a user should post if they AC and get "bad" results. Just take a minute to screenshot your settings and output (or anything you might think relevant) so it can be looked at. f12 does a screenshot it less than a second, attaching it to your post another 5. If that is done we can have a discussion. If not, what is there to say?

[md40022]

Originally Posted by MathBandit

Auto-Calc doesn't know anything about the previous season I don't think, so it's not compensating (or overcompensating) for anything.

When you ran Auto-Calc the first time, what BABIP did you tell it you wanted for the league? What about the second time?

I didn't adjust the league totals manually. I didn't set it to anything that I "wanted it to be" because I just wanted it to play out however it played out, operating under the assumption that it would be in line with reality. Of course there's going to be some variation. I know that. I'm fine with that. But leaguewide BABIP is ALWAYS in the .292-.305 range. Since Forever. It doesn't jump to .325 for a full season by random variation. And at the same time, it doesn't go from .325 down to .270 the following year on random variation. Something is screwed up there, obviously.

To answer your question though, both times that I ran auto-calc BABIP was in the mid .290s on the league settings/modifier screen at all points. It has adjusted a little bit after running auto calc. before each of the opening days, but it never visibly showed anything that was clearly insane in the league settings screen. It was always in the .290s..... which makes my situation all the more bizarre.

Here are a few pics. You'll see how my 2023 season reallllllly went haywire with big BABIP compared to MLB history. I'm not yet finished with my 2024 season, so best I can do is show a 2nd picture here and you'll see how BABIP did a complete flip flop.

I don't know what happened here. Something obviously is screwy. When league wide BABIP is normally in the .295 range, you don't just randomly jump to .325 for a full season on a complete whim only to then drop to .273 the very following season...... As a very long time OOTP customer, this year's version is the only time I have ever run auto-calc. And needless to say, it will be the last time. Obviously it works for a lot of you - that's awesome. Did I screw something up? I don't know. But I can say for a fact that I was operating off of instructions that other users gave me saying to run auto calc only on opening day when the opening day rosters are set and before the first game. That's exactly what I did. I didn't manually tweak anything. I've never been a fan of manual tweaks, I feel like it's cheating the system. So I just let the engine do its thing, but man...... something got jacked up on my end, obviously.

[itsmb8]

Quick question for the experts here, didnt want to start a new thread... How viable is copying stats directly from Baseball Ref and hitting auto calc?

I'd like for all leagues to follow the IRL numbers for 2022 for the most part and go from there, but I'm worried if I try this method that something will go very very bad.

[kidd_05_u2]

Originally Posted by itsmb8

Quick question for the experts here, didnt want to start a new thread... How viable is copying stats directly from Baseball Ref and hitting auto calc?

I'd like for all leagues to follow the IRL numbers for 2022 for the most part and go from there, but I'm worried if I try this method that something will go very very bad.

It is perfectly viable, but keep an eye for possible changes to all the modifiers that are below the league totals (starting with starting pitcher stamina). My recommendation is to take a screenshot of those modifiers before hitting auto calc and then compare that screenshot to what you get after auto calcing. If there are big differences (say 0.3 or more) you probably want to put back the original modifiers or check back here.

Two recommendations:
1) Hitting auto calc on opening day is better than doing it in advance.
2) If possible, run a full season sim after auto calcing to make sure everything is in order stats wise. Sometimes the main stats can look right but league-wide Babip goes to weird places, or you notice that pitcher usage just doesn't look right.

[Reed]

Originally Posted by md40022

I didn't adjust the league totals manually. I didn't set it to anything that I "wanted it to be" because I just wanted it to play out however it played out, operating under the assumption that it would be in line with reality. Of course there's going to be some variation. I know that. I'm fine with that. But leaguewide BABIP is ALWAYS in the .292-.305 range. Since Forever. It doesn't jump to .325 for a full season by random variation. And at the same time, it doesn't go from .325 down to .270 the following year on random variation. Something is screwed up there, obviously.

To answer your question though, both times that I ran auto-calc BABIP was in the mid .290s on the league settings/modifier screen at all points. It has adjusted a little bit after running auto calc. before each of the opening days, but it never visibly showed anything that was clearly insane in the league settings screen. It was always in the .290s..... which makes my situation all the more bizarre.

Here are a few pics. You'll see how my 2023 season reallllllly went haywire with big BABIP compared to MLB history. I'm not yet finished with my 2024 season, so best I can do is show a 2nd picture here and you'll see how BABIP did a complete flip flop.

I don't know what happened here. Something obviously is screwy. When league wide BABIP is normally in the .295 range, you don't just randomly jump to .325 for a full season on a complete whim only to then drop to .273 the very following season...... As a very long time OOTP customer, this year's version is the only time I have ever run auto-calc. And needless to say, it will be the last time. Obviously it works for a lot of you - that's awesome. Did I screw something up? I don't know. But I can say for a fact that I was operating off of instructions that other users gave me saying to run auto calc only on opening day when the opening day rosters are set and before the first game. That's exactly what I did. I didn't manually tweak anything. I've never been a fan of manual tweaks, I feel like it's cheating the system. So I just let the engine do its thing, but man...... something got jacked up on my end, obviously.

How in the world did you BA come out exactly as IRL every year? Mine is always off by .2 or .3% or more every year up to 3%.
Fielding REALLY baffles me. Does that mean you have zero errors for all those years?

[Syd Thrift]

Originally Posted by kidd_05_u2

It is perfectly viable, but keep an eye for possible changes to all the modifiers that are below the league totals (starting with starting pitcher stamina). My recommendation is to take a screenshot of those modifiers before hitting auto calc and then compare that screenshot to what you get after auto calcing. If there are big differences (say 0.3 or more) you probably want to put back the original modifiers or check back here.

Two recommendations:
1) Hitting auto calc on opening day is better than doing it in advance.
2) If possible, run a full season sim after auto calcing to make sure everything is in order stats wise. Sometimes the main stats can look right but league-wide Babip goes to weird places, or you notice that pitcher usage just doesn't look right.

Yeah, so just to open the hood a bit here... autocalc follows this workflow:

1. Sim the season using the current rosters/rotations/bullpen/sub matrices/etc. and assuming no injuries.

2. Take the league totals and compare them with the expected totals. Are hits 10% too high? If so, set the LTMs for hits to 0.9. And so on for all the stats that it covers.

3. Repeat steps 1 and 2 two more times.

Usually this gets you pretty close but yes, there are things that can skew it. If you run autocalc when the league doesn't have rosters completely set then that can be a big issue, so it's ideal to run it right before Opening Day (the game will automatically do this the "evening" before Opening Day if you set it to). Also, if you have a smaller league and some key player or five is on the DL or goes on the DL for the season right after you run the auto-calc, that can monkey with the results. The larger your league is, the less likely this is to be a factor though. And for numbers that are naturally very small - balks, for instance, but also sometimes stats like triples or wild pitches - even in a larger league one guy who does a lot of those missing time or getting called up can skew the league totals.

And of course if the initial run is way off, then the next two sims might still not get it super close.

[darkcloud4579]

Originally Posted by Sweed

Edit:
I have edited this post and added other things to the list of "too many (whatever stat). I guess I should have at the beginning instead of relying on "etc. etc. etc." that I thought would convey the message that it wasn't just HRs. Nothing wrong with discussing the HR issue some have but, it wasn't the intent of the thread to highlight that issue. The intent was to show, using data, how AC works for me. Too many times users post an issue without data and don't even give AC a try as a way to fix their issue because someone posted "it doesn't work". It's an effort to curb the "too many (pick your stat)" that are done without context. With data one can, as we are seeing with HR here, have a discussion. Without data and context there isn't much do discuss IMHO. End of edit.


Every year autocalc is questioned on these boards. Users tell other users "AC doesn't work", "I get unrealistic stats", "HR leaders with way too many", "too many balks", "too many wild pitches", "SB success too high" etc. etc. etc. Rarely do posters actually post data. If someone (I've done this many times) posts in the thread with data in context with the league and invites others to do so? Well, the thread will usually come to a crashing halt and dies. Why don't the anti-AC reply with data?

I just completed my first season in v23 and am going to share the results. The league was autocalced, manually, on opening day so the rosters were settled. Note that I did the AC "manually". Some have said that the "automatically run AC" doesn't seem to work. Why one would work and the other not? Who knows. All I know is I've always done AC manually and it has never given me any issues. I'll first post some data with LT and output along with some thoughts.

Also screens of the "stats&ai" screen. This will not only show the LT and LTM's but also the generic strategies of the league for context. Screens of league batting, pitching, fielding, and leaders. I didn't use the sortable stat screen but rather the league wide one so you could get a quick look at all categories and the top "7 to 10 ish" or so in each.

It would be nice if users would post, at a minimum, their LT and League Batting, Pitching, or whatever is needed to see their problem in context. IE too many HR by leaders? Please post LT, League Batting, and HR leaders. This will help show whether you have an actual AC issue or maybe a stat distribution issue that would be caused by ratings. Simply saying "too many of x,y, or z" doesn't really tell anyone anything. IOW if your LT is set to 4500 HR, your league hits 4500, but the leader hits 120 you have a distribution issue. The AC worked as intended as it did influence the league to output 4500 HRs.

I would hope users, both new and old, instead of reading posts that say "don't use AC" will actually try it for themselves before dismissing it all together.

I will also add in that I play with default PCM, talent changes, aging, etc. etc. I have no idea how these affect the game. I would submit that changing these will still not break AC. What they may do is cause "stat distribution issues".

Onto the reports..

LT.... OOTP output

AB   167.353   166.161
H     44.522    42.489
Dbl    8.919     8.500
Tr       898       891
HR     5.120     4.742
BB    16.222    16.410
HBP    1.850     1.848
K     31.828    31.511
BABIP   .297      .288

Slash  .266/.338/.422    .256/.327/.403

The following are real life numbers though long enough ago I don't know from what year they came.

Real..... OOTP

DP     3817    4160
WP     1458    1750
Balk    146     133
SB Att 3839    2760
SB%    .734    .641

SB attempts low, probably because of General Strategic Tendencies. IE set to "normal" in the
context of OOTP23 and today's baseball. My 3839 RL attempts would be from early 2000's.

Success rate at only .641 is surprising as the game was started with the earliest build of v23
and the "too easy to steal" bug(?). My success rate was only 51% again, with the it's too easy
to steal for the human bug(?). Probably the patch fix working with my prefix modifiers is the
cause?

HR and doubles a bit low in the total but... as a percentage of total hits? They are spot on.

Note the HBP, Balk, and WP numbers are right where they should be. You know, the numbers that
are always questioned every year on the forums. But when one suggests autocalc will make them
"right" they are met with "I was told to not use autocalc because it screws thing up".

This season is pretty much the norm for me using AC. The batting average is actually a bit farther off than usual being .010 low. It's usually with in 0.005 but still, the 0.010 is only 3-4% off. Acceptable to me and, could end up being my outlier in v23. Have to wait 'til next season's end to find out.

Screenshots...

Great thread thanks for boosting this convo. I never used to use autocalc, now I swear by it to keep my leagues relatively static statistically, I'm not interested in having evolving eras, I like baseline stats so I can compare eras easily and having totals where I want and autocalc to regulate as the league expands, etc., is super handy.

[1991Twins]

Originally Posted by itsmb8

Quick question for the experts here, didnt want to start a new thread... How viable is copying stats directly from Baseball Ref and hitting auto calc?

I'd like for all leagues to follow the IRL numbers for 2022 for the most part and go from there, but I'm worried if I try this method that something will go very very bad.

My earlier tests that I posted in this thread, I used the IRL stats. I did bump BA to .248 and BABIP to .295. Loved the results that I was seeing. Most years too HR guy was mid 40s to low 50s.

[itsmb8]

Originally Posted by kidd_05_u2

It is perfectly viable, but keep an eye for possible changes to all the modifiers that are below the league totals (starting with starting pitcher stamina). My recommendation is to take a screenshot of those modifiers before hitting auto calc and then compare that screenshot to what you get after auto calcing. If there are big differences (say 0.3 or more) you probably want to put back the original modifiers or check back here.

Two recommendations:
1) Hitting auto calc on opening day is better than doing it in advance.
2) If possible, run a full season sim after auto calcing to make sure everything is in order stats wise. Sometimes the main stats can look right but league-wide Babip goes to weird places, or you notice that pitcher usage just doesn't look right.

Thanks for the great help! I made a quick MLB quickstart and did what you suggested, and sure enough stats looked great! Bichette lead all with a .340 avg, next was around .315. Ohtani led with 48 HRs and a mid 2 ERA. League BA was .245 and league BABIP was .295. No outliers or errors from what I could tell!