Player Progression Test and Notes

[billharris44]

I ran a Dynasty for twenty years to see how player progression had changed from last year.

The good news: offense doesn't fall off a cliff like last year.

The bad news: league totals for home runs increased by almost 35% over the 20 year period. Oops.

So it looks like I may be back into the player progression mod business this year.

The other good news: the player progression system is far more complex and interesting this year.

The other bad news: because of the wholesale differences, we're essentially starting from scratch.

Last year, even though there were several thousand parameters involved, what the player progression system really boiled down to was a three-tier system of min/maxes at three ages: 18,31,45. They were defined for each star level (1-5) and for each position at each age tier. Ratings progression was pretty simple: the player's rating at 18 would go up in a linear manner toward his defined peak at 31, then fall in a linear manner to the end of his career.

With the help of many volunteer testers, we adjusted the progression system to provide reliable Dynasty stats up to 80 years in the future. However, EA's progression system itself was pretty primitive.

Not anymore, at least on first analysis. The min/max files appear to be gone entirely. In their place are a series of tables that will produce far more dynamic and realistic ratings changes.

Here's a look at the files inside the progress.big file:

--the first two, progbadj and progpadj, are probability tables defining the chances of a rating going up or down based on the players age and star rating. For instance, a 25-year old 2* player has a 50% probability of a ratings increase, a 25% probability of a ratings decrease, and a 25% probability of no change.

Here's how it looks:

Age Star Up Down Same

25 2 0.5 0.25 0.25

26 2 0.45 0.3 0.25

27 2 0.4 0.3 0.3

28 2 0.35 0.3 0.35

These probabilities are listed for every age from 18-45 and 1-5 stars, so there are quite a few values involved. But the important point is that it's no longer a linear system--it's probabilistic.

If a rating is supposed to increase, then the files progbinc and progpinc are used ('b' for batters, 'p' for pitchers). These files list, by year the probability that a rating will go up 1-5 points, 6-10 points, and 11-15 points. The probability changes based on the existing value of the rating (0-39, 40-59, 60-79, 80-99).

If a rating is supposed to decrease, then the files progbdec and progpdec are used. they have a similar format to the progbinc and progpinc files.

So what we have to work with now is far better and far more dynamic than last year. But it's going to be much, much more difficult to adjust. And testing will be more difficult than last year because there will be more ratings variances between Dynasties.

I haven't completely decided if I want to do this again, but I'm going to keep poking around in these files for the next day or so. If I do decide to start another project, I'll put out a request for interested testing volunteers next week.

[Cris987]

Sounds confusing Bill, hope you can work it out.

Btw, is the Ballpark factors in the progress file too? If yes, can you explain how it works?

Thx.

[billharris44]

Sounds confusing Bill, hope you can work it out.

Btw, is the Ballpark factors in the progress file too? If yes, can you explain how it works?

Thx.

Yes, it is. It's a file with a line for each stadium (current) with the following modifiers:

Hr, Hit, 3B ,2B.

I don't know if these apply to played games or just simmed ones. As an example, here are the HR modifier values:

Name Hr Modifier %/100

Ana -0.147743202

Ari 0.109589041

Atl -0.057894737

Bal 0.001034554

Bos -0.026315789

ChC -0.055555556

CWS 0.068627451

Cin 0.143617021

Cle -0.198863636

Col 0.173469388

Det -0.299401198

Fla -0.210526316

Hou 0.100591716

KC 0.277739726

LA 0.138211382

Mil 0.142857143

Min 0.017857143

NYM -0.071875

NYY 0.027317073

Oak 0.012738854

Phi -0.12195122

Pit -0.100558659

SD -0.107142857

SF -0.100416765

Sea 0.032679739

StL -0.118811881

TB -0.025641026

Tex 0.008230453

Tor 0.175438597

Was 0.149315069

I don't have the Stats 2005 book, but there are ballpark ratings in them and it would be pretty easy to check them.

I'm not sure if it was ever established that modifying these values last year worked.

[Cris987]

Yes, it is. It's a file with a line for each stadium (current) with the following modifiers:

Hr, Hit, 3B ,2B.

I don't know if these apply to played games or just simmed ones. As an example, here are the HR modifier values:

Name Hr Modifier %/100

Ana -0.147743202

Ari 0.109589041

Atl -0.057894737

Bal 0.001034554

Bos -0.026315789

ChC -0.055555556

CWS 0.068627451

Cin 0.143617021

Cle -0.198863636

Col 0.173469388

Det -0.299401198

Fla -0.210526316

Hou 0.100591716

KC 0.277739726

LA 0.138211382

Mil 0.142857143

Min 0.017857143

NYM -0.071875

NYY 0.027317073

Oak 0.012738854

Phi -0.12195122

Pit -0.100558659

SD -0.107142857

SF -0.100416765

Sea 0.032679739

StL -0.118811881

TB -0.025641026

Tex 0.008230453

Tor 0.175438597

Was 0.149315069

I don't have the Stats 2005 book, but there are ballpark ratings in them and it would be pretty easy to check them.

I'm not sure if it was ever established that modifying these values last year worked.

Funny, I edited the 2004 ( i ahven't got 2005 yet) progress file and made CWS' HR modifier 0.500000000 with wordpad. After that, I simmed a game and everyone started jacking 200+ HRs in a season. So, i switched the number back, but everyone still hit 200+ HRs in 162 games. Strange.

[billharris44]

There were some individual files that were pretty strange last year. The example you mentioned, plus some files that if you modded them with even small changes would crass the game. It was pretty frustrating at times.

[JohnE66]

Howdy Bill, glad to see you on the job with the progress file. I am sure you will have it nailed in no time. Hope things have been doing well for you.

Cheers,

JE

[Cris987]

The progress file in 2004 doens't seem to affect simmed games or managed games. I'll see what happens in 2005.

[billharris44]

The progress file in 2004 only doens't seem to affect simmed games or managed games. I'll see what happens in 2005.

Chris, you lost me there. Do you mean it affected simmed games and managed games, but not played games? And do you mean all the progress file values, or just the stadium values?

[billharris44]

Howdy Bill, glad to see you on the job with the progress file. I am sure you will have it nailed in no time. Hope things have been doing well for you.

Cheers,

JE

Hi John! I was figuring you'd show up any day now. Things are fine and I hope they're going well for you also.

[Cris987]

Chris, you lost me there. Do you mean it affected simmed games and managed games, but not played games? And do you mean all the progress file values, or just the stadium values?

It doesn't affect simmed games and managed games. (I edited the progress file with notepad, and everything seemed fine this time). I don't know if it actually affects the game though, cause i haven't bothered to spend the time hacking at pitches in the actual game. This is all in 2004 only, maybe it's different in 2005.

[billharris44]

[quote="Cris987

It doesn't affect simmed games and managed games. (I edited the progress file with notepad, and everything seemed fine this time). I don't know if it actually affects the game though, cause i haven't bothered to spend the time hacking at pitches in the actual game. This is all in 2004 only, maybe it's different in 2005.

[billharris44]

Okay, here's some more information. With the default rosters, I went through the rosters of ten consecutive teams (starting with L.A.) and counted how many 80+ ratings there were for hitters (against right or left-handed pitching). Here's a look.

Contact:

95-100: 1.

90-94: 9.

85-89: 18.

80-84: 25.

Power:

95-100: 2.

90-94: 1.

85-89: 16.

80-84: 23.

Again, that's for ten teams and both R/L ratings. Now, in 2024, take a look at the same ten teams:

Contact:

95-100: 43.

90-94: 18.

85-89: 32.

80-84: 26.

Power:

95-100: 19.

90-94: 8.

85-89: 18.

80-84: 31.

There's your explanation for why home runs are up 35% in 20 years. And pitching ratings are inflated as well, because having a 99/99 contact/power rating might not even get you 40 home runs in 2024.

I know that's not a true apples to apples test, since it was only 10 teams, but it's enough to veryify that ratings are grossly inflated. And I believe I understand the files enough at this point to start making modifications.

It will take a couple of days before I'm ready, but I'll ask for test volunteers fairly soon. I wasn't going to do this again after last year, but I have a huge amount of respect for the other guys who do mods. It would be a shame to have them put thousands of hours into stadiums and bats and jerseys and have it all be using the default player progression model. So out of respect for their work, I'll work on the progression mod again and do everything I can to make it work properly.

[jackcasper49]

bill, as you work through progression any tips on ratings distribution you can offer based on the multitude of test results you see as to how you're gearing your progression will be key ... I know last year you were able to take MVP and turn it back from a slightly screwed up simulation to a pretty solid one with your progression, so that makes me feel like somewhere in the ratings is the key to fixing my biggest probelm as a roster mod ... getting enough K's and enough variance on BBs between the good and the bad

your work is always top notch and hopefully I can snag a little insight out of the process so I can implement it into my roster set from the get-go and have a better functioning set to start off with so that your progression just makes it perfect

[xedef17]

Hi Again this year Bill

Ahh a new season of MVP, which means a new season of bugs to work out to make the game playable. I picked up this years version and if you need me to help test the progression mod again this year, I can be of some assistance . . . although grad school is taking up a plethora of my time.

[Kamasutra]

I wasn't going to do this again after last year, but I have a huge amount of respect for the other guys who do mods. It would be a shame to have them put thousands of hours into stadiums and bats and jerseys and have it all be using the default player progression model.

yep, a shame indeed, your prog mod is nothing less than a necessity, especially this year (last mvp title for a while), it's much more likely that people will see the effects of progression over a stretched out dynasty running decades. Add to that owner mode which is only 25 years but quite enough to see the effects of the tweaks.

Ahh a new season of MVP, which means a new season of bugs to work out to

make the game playable.

this year's version is actually quite playable out of the box...

As far as bugs in gameplay at least if anything they are nowhere to be seen (yet?)

[x94blair3]

Bill,

Glad to see you back at it! I can see the motivation not to do it, but the game is getting better and better out of the box (in my humble opinion) so it seems more more worth the effort to keep it working through a dynasty.

If you need any help with testing let me know.

Thanks,

Nick

[billharris44]

jackcasper49, I usually tune the player progression system to the default ratings--not individual players but the general bell curve.

xedef17, nice to hear from you. What are you studying in grad school? I have a couple of guys testing and hope to have a full complement in another day or so. I don't want you flunking out on my account.

The reason I'm willing to do this again is because player progression seems to be the only big problem. I'm really impressed at how complete the game is otherwise. And really, I can't imagine any tuning player progression under the pressure of having to get the game shipped. It's very complex and it might have been changing from build to build.

[x94blair3]

Like I posted in the operationsports forums, I think the game this year is excellent. Between the added sliders and tightening of the gamplay they seem to have given the hardcore baseball fans something to work with while still appealing to the average gamer. I agree, I think in most facets it seems great.

At least on the field

I haven't tinkered with dynasty other than the first 5-6 games...I was surprised to see so many of my players unhappy and it looks like contracts are completely botched, but then again I'm running the Tigers organization.

[Cris987]

Bill, I have just found one reason why projected stats after a decade of seasons are screwed up. I was using MVPEdit to find the players with All-Star Star Power and I found the results very ... stupid? There were 91 players under the age of 26/27 in the game that has allstar star power, with 10 of them being current all stars (prior, pujols, beltre), 8 of them prospects (some current, e.g BJ Upton), and all the others fake players. That's EIGHTY SOMETHING Albert Pujols and Pedro Martinez and Barry Bonds you got there! That makes no sense at all, and I suspect that the rookie file creates rookie with star powers close to the same ratio.

Hope this helps (althought it doesn't, lol).

Cris

[billharris44]

Bill, I have just found one reason why projected stats after a decade of seasons are screwed up. I was using MVPEdit to find the players with All-Star Star Power and I found the results very ... stupid? There were 91 players under the age of 26/27 in the game that has allstar star power, with 10 of them being current all stars (prior, pujols, beltre), 8 of them prospects (some current, e.g BJ Upton), and all the others fake players. That's EIGHTY SOMETHING Albert Pujols and Pedro Martinez and Barry Bonds you got there! That makes no sense at all, and I suspect that the rookie file creates rookie with star powers close to the same ratio.

Hope this helps (althought it doesn't, lol).

Cris

Hi Cris,

You're absolutely right. There ae way, way too many guys at the top end over time. It's not just position players, either--there's the same problem with pitchers.

We're working on it. :)

[UncleMo]

does this mean DO NOT start a dynasty until this mod is complete? I don't wanna have to start all over in my dynasty. I'm near ready to start my dynasty so let us know.

[jpup]

does this mean DO NOT start a dynasty until this mod is complete? I don't wanna have to start all over in my dynasty. I'm near ready to start my dynasty so let us know.

it won't take effect until after season 1. i would feel free to play your first season, just dont advance past the playoffs until he is done. i believe that's how it worked last year. you can install it during your dynasty.

[boulet430]

Im curious Bill if you are able to mod the datafile to where strikeouts and such are more realistic across the league as well. I think someone mentioned here that if you sim a season the highest strikeout total for a pitcher is like 150 or something. I don't know if its in the same area as the homeruns and what not, but if it is, I'd love for you take a look at that.

[billharris44]

Im curious Bill if you are able to mod the datafile to where strikeouts and such are more realistic across the league as well. I think someone mentioned here that if you sim a season the highest strikeout total for a pitcher is like 150 or something. I don't know if its in the same area as the homeruns and what not, but if it is, I'd love for you take a look at that.

The seasons we're simming now have the strikeout leader at around 210-220. That's still too low, but it's quite a bit up from 150.

Player progression affects strikeouts because contact ratings, pitch movements, etc., are all affected by player progression. It's not a direct effect like the parameter file in Pure Sim where you have a setting for strikeouts and you can directly increase it.

If I could get 7-10 guys above 200 strikeouts for the season, I'd be very happy with that (right now we're getting 2-3). The league leader is usually 250+, but I don't think the sim engine is ever going to produce a season like that.

[Kamasutra]

as you pointed out, if the contact ratings are progressing exponentially, then it makes sense that the strikout leader totals decrease at the same rate, perhaps even quicker. Hard to tell if there's a way to increase one without sacrifying the other, maybe if they had added an "experience" modifier (sorta like age, but measured in league years experience instead), then that would probably solve it..(ie: pitchers with a certain level experience are subject to the modifier which increases strikouts by a certain amount for example).

I actually read somwhere that the game had modifiers like this integrated in the progression, just don't know which, maybe you'll stumble across them