New Game Engine Tests

sprague · 07-04-2022, 06:20 AM

I am making this thread not to attack OOTP 23.
I am making it to show that the game engine of this version is vastly different from other versions. I am not saying GOOD or BAD, just it is very different.
The way to show that is with a real sample. That is by running the SAME LEAGUES in this version and older versions and compare the results.
I have done this with historical games (as they are unchanging) and with the same fictional leagues from older versions moved up.

What I have found, as will be laid out below, is that the older versions (I am using 21) and the new 23 version will give you the same yearly totals.
About the same number of home runs, hits, walks will happen in the two games.
What is changed is DISTRIBUTION. Two of the same leagues where 5000 home runs are hit, will have in 23 certain players hitting much more than previous. I have also noticed this with pitcher's strike outs and stolen bases.

So let us get to the fist post of what is being shown here.

sprague · 07-04-2022, 06:22 AM

Ok here is a historical league. 1890
Slight alterations to the actual historical game.
dropped to 14 teams, 1 league.
But will have the same players, same number of teams, same GP, same
settings (such as all ballparks neutralized to 1.000) and the same totals to run the modifier from.
I modernized the totals for the games (as that is how I play) but as you see it is the same in both games.

Here is the stats in 21 and the team totals.
That is mostly what I have come to expect from my years of historical games.

sprague · 07-04-2022, 06:25 AM

Here is the totals in 23. Notice how different the distribution is. 10 more home runs for top hitters. In other tests I have had Bill Joyce hitting 75 home runs.
Strikeouts for top guys like Toad Ramsey are vastly off the charts, as is the players now stealing a ton of bases.

These are the first tests. I will be showing other years as well as the results of the same fictional leagues.
But there is nothing different from what I have just showed you. DISTRIBUTION of totals is vastly different now.
This is not necessarily criticism, it is showing all of us who have played this game for a long while that the results are going to be spread out very differently now.

sprague · 07-04-2022, 06:27 AM

Here is 1891.
first is the leaders from 21:
Pitcher strike out totals were cut off. League Leader was Ramsey with 271, Amos Rusie next at 248.

Next 2 screens from 23 to see the batting and pitching totals

Reed · 07-04-2022, 06:52 AM

Thanks for doing this.
Yes I noticed the stolen bases were way different. I hadn’t noticed that much of a difference in home runs.

Karl Pagel Blues · 07-04-2022, 08:20 AM

Quote:

Originally Posted by sprague

What I have found, as will be laid out below, is that the older versions (I am using 21) and the new 23 version will give you the same yearly totals. What is changed is DISTRIBUTION.

Fascinating. Thank you for doing this. Definitely some significant differences.

By the way, seeing the names of these players (Sam Thompson, Dan Brouthers, Mike Tiernan, Hugh Duffy, etc.) brought back memories! The first computer baseball game I ever played was MICRO LEAGUE BASEBALL in 1984. I had the GM disk and the Baseball Encyclopaedia book. I manually created an entry in Micro League for every player I considered an MLB Great and saved the data on the GM floppy disk. In hindsight, I should not have included Alan Knicely's 1981 MLB season (4-7, 2HR). My brother drafted Knicely and no matter what I offered him, he wouldn't trade him to me.

Anyway, thank you for taking the time to do this. It will be interesting to hear what's behind the different numbers between 21 and 23 - because I've been thinking about starting a historical game.

LansdowneSt · 07-04-2022, 01:55 PM

It seems the spirit is to give the exceptional player more of an opportunity to be exceptional. So, whomever the equivalent is of Ruth or Rickey is in that era, they have a chance of being more like their outlier self, versus being squeezed toward the mean. Or to give someone a chance to be Maris or Vince Coleman. Toad was that kind of outlier in strikeouts back in the day (looking at the irl 1887 leaderboard as I type). That's my initial takeaway. Thanks for running these.

sprague · 07-05-2022, 11:41 AM

Ok here is the results for 1892
i will show some fictional leagues next

NoOne · 07-05-2022, 04:15 PM

Quote:

Originally Posted by LansdowneSt

It seems the spirit is to give the exceptional player more of an opportunity to be exceptional. So, whomever the equivalent is of Ruth or Rickey is in that era, they have a chance of being more like their outlier self, versus being squeezed toward the mean. Or to give someone a chance to be Maris or Vince Coleman. Toad was that kind of outlier in strikeouts back in the day (looking at the irl 1887 leaderboard as I type). That's my initial takeaway. Thanks for running these.

since i've been playing, that's what i thought it needed.

the bulk of the league should be volatile with a few consistent manny ramirez-types etc (extremely simplified, still a spectrum of course)... how many? whatever histoical data says as i have no preconceived notion, but the distribution of resulting stats was always too spread out.

it was impossible for a max steal/speed guy to hit peak numbes like that of ricky henderson without boosting SB league total modifier a bit. I usually used a pre-honed LTM that translated to a similar average SB total, too. so, my baselines didn't fluctuate much release to release.

totals and modifiers have to jive* for ~130sb to be possible no matter what they do with distribution of talent.

this is why i started picking totals/modifiers that produced individual results i wanted vs worrying about league totals too much. I'd done it with SB for a long time (roughly same total and %-success would get the upper end results i wanted release to release, but maybe not this year sounds like).

koohead · 07-08-2022, 11:52 AM

I too actually think this is 'better' and more realistic, though a part of me worries about the single season records and what kind of impact this engine change will have on those records.

brotherblues · 07-08-2022, 12:15 PM

Quote:

Originally Posted by koohead

I too actually think this is 'better' and more realistic, though a part of me worries about the single season records and what kind of impact this engine change will have on those records.

I'll chip in that I'm glad they took that step even with the risks it carries.

Plus, real life records are based off what is essentially a single sim. Real life only had one shot at it. We're running countless sims here. It stands to reason real life records should be broken a few times.

Reed · 07-08-2022, 05:53 PM

Just curious, are you playing 1 year or 3 year recalc??

mschroe · 07-09-2022, 07:38 PM

There seem to be a few of us that are very frustrated with the extreme outliers, especially on HRs, with this 23's game engine. I'm having to go into commissioner mode and suspend my best player, my current season for 40 games, to bring some normalcy. He had 52 HRs in 87 games started. Other years were absurd when he and a couple other players from other teams would hit 50 - 60+ and rest of league leaders topping out in low-40s. Note I play with 2018 season totals. Yes, a higher home run season but the top 5 were 48, 43, 40, 39, 39.

Before the last patch I had PCMs changed enough that the HR leaders weren't so absurd. Had 50+ HR leader every season typically once I hit the 2040s and an occasional 60+ HR person over 20 - 30 years (I think it was 3 - 4). Went back to default PCMs with last patch updates and I think the problem has gotten worse compared to the release version once you advance into 2040s+.

I honestly cannot understand how others that play fictional and sim into future think this engine is okay. This does not make it more real. Doesn't seem like this has received much acknowledgement on the forums that it's an issue. Just a few people like Sprague and a couple posts from others like me grumbling. This seems like a fundamental issue to the game. In fact, I've heard Markus multiple times cite how the game engine doesn't produce crazy HR numbers as an example of what's great about the game.

HRBaker · 07-09-2022, 07:53 PM

Quote:

Originally Posted by mschroe

There seem to be a few of us that are very frustrated with the extreme outliers, especially on HRs, with this 23's game engine. I'm having to go into commissioner mode and suspend my best player, my current season for 40 games, to bring some normalcy. He had 52 HRs in 87 games started. Other years were absurd when he and a couple other players from other teams would hit 50 - 60+ and rest of league leaders topping out in low-40s. Note I play with 2018 season totals. Yes, a higher home run season but the top 5 were 48, 43, 40, 39, 39.

Before the last patch I had PCMs changed enough that the HR leaders weren't so absurd. Had 50+ HR leader every season typically once I hit the 2040s and an occasional 60+ HR person over 20 - 30 years (I think it was 3 - 4). Went back to default PCMs with last patch updates and I think the problem has gotten worse compared to the release version once you advance into 2040s+.

I honestly cannot understand how others that play fictional and sim into future think this engine is okay. This does not make it more real. Doesn't seem like this has received much acknowledgement on the forums that it's an issue. Just a few people like Sprague and a couple posts from others like me grumbling. This seems like a fundamental issue to the game. In fact, I've heard Markus multiple times cite how the game engine doesn't produce crazy HR numbers as an example of what's great about the game.

I, among others, DO think this engine is better than the last. It was heartbreaking to see season after season produce almost identical distributions. Ruth NEVER hit 60 HRs for me in almost 20 years of playing OOTP. This year in my 1927 replay he hit 62 HRs. That's realism. I welcome it.

mschroe · 07-09-2022, 08:06 PM

I specially mentioned fictional sims, advanced into the future. I'm glad the engine recreates the seasons nicely for you. But to be honest...I didn't really know or even conceive that anyone would replay seasons from the past until I spent time reading these forums. To each their own...it's one of the great things about the game.

But Ruth hitting 50-ish HRs in '27 is reasonable. To get to 60 HRs, if it takes an engine that makes future sims often unplayable without extreme intervention in commissioner mode - when future sims and franchise mode seems a bit like a core part of the game - doesn't seem like a reasonable tradeoff. So you can get from 95% reality to closer to 100%, the future sims players find themselves needing extensive modifications to get the simulation quality we had in prior seasons.

jbergey22 · 07-09-2022, 11:07 PM

Quote:

Originally Posted by mschroe

There seem to be a few of us that are very frustrated with the extreme outliers, especially on HRs, with this 23's game engine. I'm having to go into commissioner mode and suspend my best player, my current season for 40 games, to bring some normalcy. He had 52 HRs in 87 games started. Other years were absurd when he and a couple other players from other teams would hit 50 - 60+ and rest of league leaders topping out in low-40s. Note I play with 2018 season totals. Yes, a higher home run season but the top 5 were 48, 43, 40, 39, 39.

Before the last patch I had PCMs changed enough that the HR leaders weren't so absurd. Had 50+ HR leader every season typically once I hit the 2040s and an occasional 60+ HR person over 20 - 30 years (I think it was 3 - 4). Went back to default PCMs with last patch updates and I think the problem has gotten worse compared to the release version once you advance into 2040s+.

I honestly cannot understand how others that play fictional and sim into future think this engine is okay. This does not make it more real. Doesn't seem like this has received much acknowledgement on the forums that it's an issue. Just a few people like Sprague and a couple posts from others like me grumbling. This seems like a fundamental issue to the game. In fact, I've heard Markus multiple times cite how the game engine doesn't produce crazy HR numbers as an example of what's great about the game.

I am with you here. These 60+ home run seasons are not outliers with this "new engine" they are seasonal.

mschroe · 07-15-2022, 03:04 PM

Update: I changed the PCMs after running with the new defaults for 40 years in my first sim since last patch. The changes eliminated the 60+ HR every season, multiple players issue. Have had only one or two 60+ seasons going 30 years forward. My PCM changes were:

Batting Contact: 1.050
Batting Power: 1.100
Avoid Ks: 0.800

I see two issues that keep driving the skewed distribution of HRs across all the sims I've been running:

(1) The spreading out of power skills. An increasing % of players on active rosters have power ratings (100% scouting accuracy) of 45 or less. In 2022, there's 0%, 3%, 14%, and 31% of batters with 30, 35, 40, 45 for power. By 2040, even with power PCM @ 1.100 I'm getting 1%, 6%, 15%, 27%, and 40% for 25, 30, 35, 40, and 45 power. By the time I get to 50 rating, the differences are smaller (53 vs 57%). This distribution is one half the problem but is then exacerbated by the avoid K ratings

(2) Avoid K just skews much higher in PCMs, even with latest patch. Before the last patch, I had to take the avoid K PCM to 0.6 to control the excessive HR outliers every season. The first sim after the last patch I left everything at 1.000, which when combined with power more widely distributed throughout the 20 - 80 scale, the best power hitters have such good avoid K skill that they make up a disproportionate number of the HR total relative to real-life normals.

I'll keep people updated as this sim progresses. But so far so good.

jcard · 07-16-2022, 12:30 AM

Quote:

Originally Posted by mschroe

Update: I changed the PCMs after running with the new defaults for 40 years in my first sim since last patch. The changes eliminated the 60+ HR every season, multiple players issue. Have had only one or two 60+ seasons going 30 years forward. My PCM changes were:

Batting Contact: 1.050
Batting Power: 1.100
Avoid Ks: 0.800

I see two issues that keep driving the skewed distribution of HRs across all the sims I've been running:

(1) The spreading out of power skills. An increasing % of players on active rosters have power ratings (100% scouting accuracy) of 45 or less. In 2022, there's 0%, 3%, 14%, and 31% of batters with 30, 35, 40, 45 for power. By 2040, even with power PCM @ 1.100 I'm getting 1%, 6%, 15%, 27%, and 40% for 25, 30, 35, 40, and 45 power. By the time I get to 50 rating, the differences are smaller (53 vs 57%). This distribution is one half the problem but is then exacerbated by the avoid K ratings

(2) Avoid K just skews much higher in PCMs, even with latest patch. Before the last patch, I had to take the avoid K PCM to 0.6 to control the excessive HR outliers every season. The first sim after the last patch I left everything at 1.000, which when combined with power more widely distributed throughout the 20 - 80 scale, the best power hitters have such good avoid K skill that they make up a disproportionate number of the HR total relative to real-life normals.

I'll keep people updated as this sim progresses. But so far so good.

I do not know if the issue that I had raised in the thread linked below is related, but I do wonder if it means that the game is generating players more suited to the statistical DNA of the MLB fifteen years ago. That scenario would at least be consistent with the trends you describe as the game moves forward and seed players are phased out. In any event, I was surprised by the total disinterest among forum members in the matter. With the emphasis placed by developers on the relative ratings among players (and, axiomatically, the distribution of those ratings) I feel that the lack of a self-evident (at least to my mind) reason for the matter I noted in the linked thread warrants an explicit explanation be provided:

https://forums.ootpdevelopments.com/...d.php?t=339261

Reed · 07-16-2022, 07:33 AM

Quote:

Originally Posted by jcard

I do not know if the issue that I had raised in the thread linked below is related, but I do wonder if it means that the game is generating players more suited to the statistical DNA of the MLB fifteen years ago. That scenario would at least be consistent with the trends you describe as the game moves forward and seed players are phased out. In any event, I was surprised by the total disinterest among forum members in the matter. With the emphasis placed by developers on the relative ratings among players (and, axiomatically, the distribution of those ratings) I feel that the lack of a self-evident (at least to my mind) reason for the matter I noted in the linked thread warrants an explicit explanation be provided:

https://forums.ootpdevelopments.com/...d.php?t=339261

I and I assume most others, don’t have a clue as to what PCMs do? I would love a good explanation.

kidd_05_u2 · 07-17-2022, 01:00 AM

Quote:

Originally Posted by Reed

I and I assume most others, don’t have a clue as to what PCMs do? I would love a good explanation.

Player creation modifiers dictate the ratings that players created for a given league will have.

For example, if you have league A with a PCM of 1.000 for the contact rating and a league B with a PCM of 1.200 for contact, players created in league B will have approximately 20% higher contact ratings/potential.

You can check the manual on PCMs here: http://manuals.ootpdevelopments.com/...vanced_players

07-04-2022, 06:20 AM	#1
sprague All Star Starter Join Date: Aug 2011 Posts: 1,947	New Game Engine Tests I am making this thread not to attack OOTP 23. I am making it to show that the game engine of this version is vastly different from other versions. I am not saying GOOD or BAD, just it is very different. The way to show that is with a real sample. That is by running the SAME LEAGUES in this version and older versions and compare the results. I have done this with historical games (as they are unchanging) and with the same fictional leagues from older versions moved up. What I have found, as will be laid out below, is that the older versions (I am using 21) and the new 23 version will give you the same yearly totals. About the same number of home runs, hits, walks will happen in the two games. What is changed is DISTRIBUTION. Two of the same leagues where 5000 home runs are hit, will have in 23 certain players hitting much more than previous. I have also noticed this with pitcher's strike outs and stolen bases. So let us get to the fist post of what is being shown here.

07-04-2022, 06:22 AM	#2
sprague All Star Starter Join Date: Aug 2011 Posts: 1,947	Ok here is a historical league. 1890 Slight alterations to the actual historical game. dropped to 14 teams, 1 league. But will have the same players, same number of teams, same GP, same settings (such as all ballparks neutralized to 1.000) and the same totals to run the modifier from. I modernized the totals for the games (as that is how I play) but as you see it is the same in both games. Here is the stats in 21 and the team totals. That is mostly what I have come to expect from my years of historical games. Attached Images Last edited by sprague; 07-04-2022 at 06:23 AM.

07-04-2022, 06:25 AM	#3
sprague All Star Starter Join Date: Aug 2011 Posts: 1,947	Here is the totals in 23. Notice how different the distribution is. 10 more home runs for top hitters. In other tests I have had Bill Joyce hitting 75 home runs. Strikeouts for top guys like Toad Ramsey are vastly off the charts, as is the players now stealing a ton of bases. These are the first tests. I will be showing other years as well as the results of the same fictional leagues. But there is nothing different from what I have just showed you. DISTRIBUTION of totals is vastly different now. This is not necessarily criticism, it is showing all of us who have played this game for a long while that the results are going to be spread out very differently now. Attached Images

07-04-2022, 06:27 AM	#4
sprague All Star Starter Join Date: Aug 2011 Posts: 1,947	Here is 1891. first is the leaders from 21: Pitcher strike out totals were cut off. League Leader was Ramsey with 271, Amos Rusie next at 248. Next 2 screens from 23 to see the batting and pitching totals Attached Images Last edited by sprague; 07-04-2022 at 06:29 AM.

07-05-2022, 11:41 AM	#8
sprague All Star Starter Join Date: Aug 2011 Posts: 1,947	Ok here is the results for 1892 i will show some fictional leagues next Attached Images

07-04-2022, 06:52 AM	#5
Reed Hall Of Famer Join Date: Oct 2013 Posts: 2,339	Thanks for doing this. Yes I noticed the stolen bases were way different. I hadn’t noticed that much of a difference in home runs.

07-04-2022, 01:55 PM	#7
LansdowneSt Global Moderator Join Date: Feb 2004 Location: From Duxbury, Mass residing Baltimore Posts: 7,565	It seems the spirit is to give the exceptional player more of an opportunity to be exceptional. So, whomever the equivalent is of Ruth or Rickey is in that era, they have a chance of being more like their outlier self, versus being squeezed toward the mean. Or to give someone a chance to be Maris or Vince Coleman. Toad was that kind of outlier in strikeouts back in the day (looking at the irl 1887 leaderboard as I type). That's my initial takeaway. Thanks for running these.

07-08-2022, 11:52 AM	#10
koohead Hall Of Famer Join Date: Dec 2001 Location: Phoenix, AZ Posts: 3,211	I too actually think this is 'better' and more realistic, though a part of me worries about the single season records and what kind of impact this engine change will have on those records. __________________ GM - New Jersey Bears of the NPBL;

07-08-2022, 05:53 PM	#12
Reed Hall Of Famer Join Date: Oct 2013 Posts: 2,339	Just curious, are you playing 1 year or 3 year recalc??

07-09-2022, 07:38 PM	#13
mschroe Minors (Rookie Ball) Join Date: Nov 2021 Posts: 25	There seem to be a few of us that are very frustrated with the extreme outliers, especially on HRs, with this 23's game engine. I'm having to go into commissioner mode and suspend my best player, my current season for 40 games, to bring some normalcy. He had 52 HRs in 87 games started. Other years were absurd when he and a couple other players from other teams would hit 50 - 60+ and rest of league leaders topping out in low-40s. Note I play with 2018 season totals. Yes, a higher home run season but the top 5 were 48, 43, 40, 39, 39. Before the last patch I had PCMs changed enough that the HR leaders weren't so absurd. Had 50+ HR leader every season typically once I hit the 2040s and an occasional 60+ HR person over 20 - 30 years (I think it was 3 - 4). Went back to default PCMs with last patch updates and I think the problem has gotten worse compared to the release version once you advance into 2040s+. I honestly cannot understand how others that play fictional and sim into future think this engine is okay. This does not make it more real. Doesn't seem like this has received much acknowledgement on the forums that it's an issue. Just a few people like Sprague and a couple posts from others like me grumbling. This seems like a fundamental issue to the game. In fact, I've heard Markus multiple times cite how the game engine doesn't produce crazy HR numbers as an example of what's great about the game.

07-09-2022, 08:06 PM	#15
mschroe Minors (Rookie Ball) Join Date: Nov 2021 Posts: 25	I specially mentioned fictional sims, advanced into the future. I'm glad the engine recreates the seasons nicely for you. But to be honest...I didn't really know or even conceive that anyone would replay seasons from the past until I spent time reading these forums. To each their own...it's one of the great things about the game. But Ruth hitting 50-ish HRs in '27 is reasonable. To get to 60 HRs, if it takes an engine that makes future sims often unplayable without extreme intervention in commissioner mode - when future sims and franchise mode seems a bit like a core part of the game - doesn't seem like a reasonable tradeoff. So you can get from 95% reality to closer to 100%, the future sims players find themselves needing extensive modifications to get the simulation quality we had in prior seasons. Last edited by mschroe; 07-09-2022 at 08:15 PM.

07-15-2022, 03:04 PM	#17
mschroe Minors (Rookie Ball) Join Date: Nov 2021 Posts: 25	Update: I changed the PCMs after running with the new defaults for 40 years in my first sim since last patch. The changes eliminated the 60+ HR every season, multiple players issue. Have had only one or two 60+ seasons going 30 years forward. My PCM changes were: Batting Contact: 1.050 Batting Power: 1.100 Avoid Ks: 0.800 I see two issues that keep driving the skewed distribution of HRs across all the sims I've been running: (1) The spreading out of power skills. An increasing % of players on active rosters have power ratings (100% scouting accuracy) of 45 or less. In 2022, there's 0%, 3%, 14%, and 31% of batters with 30, 35, 40, 45 for power. By 2040, even with power PCM @ 1.100 I'm getting 1%, 6%, 15%, 27%, and 40% for 25, 30, 35, 40, and 45 power. By the time I get to 50 rating, the differences are smaller (53 vs 57%). This distribution is one half the problem but is then exacerbated by the avoid K ratings (2) Avoid K just skews much higher in PCMs, even with latest patch. Before the last patch, I had to take the avoid K PCM to 0.6 to control the excessive HR outliers every season. The first sim after the last patch I left everything at 1.000, which when combined with power more widely distributed throughout the 20 - 80 scale, the best power hitters have such good avoid K skill that they make up a disproportionate number of the HR total relative to real-life normals. I'll keep people updated as this sim progresses. But so far so good.