Questions on rating in the Historical-file ?

[clamel]

I'm about to launch an updated version of the old Historical file, but I got some questions on how ratings are worked on by the game.

Since I added some 150-200 Free Agent players with lower OVR to just fit in and replace the fictional players that in the first FA-window dominate that part, resulting that teams are signing scores of those fictional names. Not my cup of tea.

My concern is that even putting those FA players on low OVR, like 65 for most they could pop up in FA pool with OVR up to 90 and over, then making them top players in the league and get hunted by most teams. Which is not the intention.
Like the game insist on having star players in the FA pool, fictional or otherwise.

The next issue is on Defensive Backs. The game behave strange IMHO. Most players are setup as DB but they end up when starting on any of the CB,SS or FS positions resulting in a problem with min-max postions.
I went ahead and really fixed them in their proper most common position, but sadly found out the game really didn't mind. It's OK, but the crazy things like having a balanced team turned up 14 CB, 0 SS and 1 FS initially. Why do the game resort, and on what credentials ???

I have scanned over how the 2002 season started with training camps and pre-season to make that year fully fledge to start a historical run in this game. I added players so every team got around 60 players in the rosters at the get go. In time I will add the players joining teams for the last years missing but must know how the game really judge players.
It looks weird and very sporadic depending on the OVR you give the player.
Brooks already told me that if a player in the old historical file had over 100 in OVR that was to make sure he became a top player, with about 90, but important to know if their also is a LOW limit. Lets say 50-60 to make sure they don't become top players. Then what will the game do with creating fictional names ?
The ideal situation of course is if rating stayed very close to the ones given in that json file. Can't understand why not.

It would be nice if I could publish this file shortly since these are the last worries I got.
I must say I can't test this proper on the latest PF21 version since my demo period has gone out, but while I had the time I could see that no big difference was up to the PF20 version.

Some gamers have asked for an updated file and this is on a road to be that. I think most also like to keep fictional players out as much as possible and that "real player" names performed close to what they did in real life.
I picked the 2002 season to start because it was when the alignment and teams was as it is today. So far starting from any other season will just be as the old Historical.json. However if this projects comes out nice I will over time fill it up with missing players mostly in the later years.

[Rebel]

Are you saying that AI teams might build their roster with 14 CBs and 1 FS and that's okay as long as they've met the minimum # of players for DBs? I think that could be fixed by simply changing it to require a minimum on the specific positions, not just the grouping of DBs. Perhaps I'm misunderstanding.
I'm in the middle of modding the 2020 rosters, so I want to know as much as I can about how the game works so I'm doing it right.

[clamel]

Are you saying that AI teams might build their roster with 14 CBs and 1 FS and that's okay as long as they've met the minimum # of players for DBs? I think that could be fixed by simply changing it to require a minimum on the specific positions, not just the grouping of DBs. Perhaps I'm misunderstanding.
I'm in the middle of modding the 2020 rosters, so I want to know as much as I can about how the game works so I'm doing it right.

OK, what I was trying to say was that in the Historical File the players are setup as DB,CB,FS,SS. I tried to change all the DB to what they really mostly played. The result was that the game during the setup made it's own decision on what special position all the player should have.
That resulted in that from the file we could have 8 CB, 3 SS and 3 FS (example) but looking at the roster for the first time it could say 10 CB,4 FS and ZERO SS. Not even being the same players filling in at those FS positions. So I took it that the game decided all on it's own.
The same situation don't pop up with example OL and DL. They are fixed on the position as in the json-file.
Of course this brings the AI already in the first round of Free Agency to go after to fill those gaps. I have "flooded" the FA pool with "real" player names that was not in the original historical json file, just to hope they sign "real" players and not the added on fictional players. Even on that case I get trouble seing some fictional players rated very high and make the AI teams signing them. I put all FA very low around 65, but I could find some at 80-90 rating OVR after the game setup. In my file they are, as they should be, just fringe players and fill-in if injury strike.

I perhaps make the error on comparing with historical runs in OOTP, but in my mind a historical run should have enough players that really played and not any fictional players. The upcoming seasons after 2002 is needed to be filled up with "real" players too and not get bunches of fictional players mixed in. I'm on that issue looking at draft-classes onwards.

But feel a bit discourage from putting down this huge work updating the file if the game itself works against me.
Lets hope Brooks can look over this and make a historical save play out as it should.
After all if you like fictional players you should run the game from the getgo with that and create a fantasy world. Just hoped we could have an option, this or that.

The min-max values are in PF20 not editable, but even what I saw during my demorun in PF21 you don't get to this Adv setting until after the AI made its roster setup. Whatever you should'nt set any position to zero just the same, so the editing of min-max should mostly be to use max.
As I said several times min-max may solve some problems in programming to help the AI, but create perhaps just as many.

[brooks_piggott]

If you set the positions to CB/FS/SS you should be fine. If you use DB or S or something like that the game will randomly choose between related official positions.

The game will generate fictional players if the player pool is not enough to cover full rosters. So if you have 53 man rosters + 10 practice squad, and you pick 1960 as a year, there most certainly will not be enough players in the history file to cover that (when teams were 45 man rosters and no PS at all). Feel free to add your own fake players if you want to whatever season you're targeting by creating guys with like 10 OVR which should drop them to the bottom of the list and only sign them in emergency cases.

[clamel]

I'm on it and it looks good so far. I concentrate on filling up 2002 so no fictional players will pop up inside rosters. All team around 60 or more at the start, and then some 150 FA real ones. It's the first round of FA that is a concern, but since all teams now got their min. in place only a handful signings are done, and only a couple of fictionals. I will stop that with more players on that position.
Still a bit troubled that some fringe players I added that just had a time in training camp and/or pre-season during 2002 could be re-rated by the AI from my 65 OVR to up to 75-80 OVR. It's of course different at different starts. If I keep them lower (below 60OVR) even more fictionals will be on the top FA list.

When I test forward in 2003,2004 it looks OK, but will add some more players not presently in the original Historical file. Have confident to keep "fictionals" out of the teams, but further down the years they might sneak in.

I will publish a file before the SuperBowl this year, but then update with adding the missing last seasons. That will take a lot of time, but since the file works in all versions of PFxx I have hope some will make use of it.

Note that this setup will be for start in the year of 2002, other years the file is untouched.

[Invisible Witness]

Disabling training camp gains should prevent changing players overall , no ?
On a side note it would be nice to have an option where we could lower fake players overall so they don't get picked over real players.

[brooks_piggott]

The historical file takes a rating (Similar to OVR but not exactly the same) and uses it to extrapolate player ratings in the game... there is randomness built into it so 2 100 Rated QB's won't come in with the exact same ratings.

[clamel]

Small (or maybe big) concern on ratings from the OVR numbers.
Could you inform me if one do have a lower OVR settings for players you surely don't like to pop up in high 80 and 90 and over. Still they should be better than the flood of fictional players the AI creates at gamestart. (I got most of mine new at 65 OVR.)
I understand that if you have a player that in real life really made an impact one should put him on 100 OVR or higher.
What scares me is to see some FA low ranked (65) player from the file pops up with high 80 and even over 90 in the FA pool.

I found a small remedy cure for them by actually putting them in the teams they had a small time in during training camp/pre-season.
Expanding all teams to around and over 70 players stops the AI from signing any players during that time. Of course imprtant to have correct min. sets, but that I have now.

Still should not this historical file have more stats, like the "normal" player file. Now a player got his OVR but not anything direct the game to make him anything close to what was his important skills in real-life.

I always compare with other games of course, OOTP, FOF and such were you get the finer details. Overall ratings could be set on most players, but it would have been top if the detailed fields was in the historicalfile. Just say, if not any number in those fields the AI will decide, but if you got a special running QB with speed rating he should pop up in-game with that.
My ever problem Lamar Jackson, as an inmobile QB doesn't sound nice. Sad.

Lots of players have asked for this update and I'm very close to release now, with starts in 2002.
The task has been to fill out the meager rosters that file had for 2002, that created loads of fictional names. Not fun to play a historical save then see a fictional QB take his team to a win in the Super Bowl. Not my cup of tea. I'm on adding first year players for 2018 and will go on with 2019 and 2020. That will take more time but like to get some hint on if this is a lost cause or not.

We all like this game to be so much more I guess, and I believe it can be.

[tyronethecalzone]

Small (or maybe big) concern on ratings from the OVR numbers.
Could you inform me if one do have a lower OVR settings for players you surely don't like to pop up in high 80 and 90 and over. Still they should be better than the flood of fictional players the AI creates at gamestart. (I got most of mine new at 65 OVR.)
I understand that if you have a player that in real life really made an impact one should put him on 100 OVR or higher.
What scares me is to see some FA low ranked (65) player from the file pops up with high 80 and even over 90 in the FA pool.

I found a small remedy cure for them by actually putting them in the teams they had a small time in during training camp/pre-season.
Expanding all teams to around and over 70 players stops the AI from signing any players during that time. Of course imprtant to have correct min. sets, but that I have now.

Still should not this historical file have more stats, like the "normal" player file. Now a player got his OVR but not anything direct the game to make him anything close to what was his important skills in real-life.

I always compare with other games of course, OOTP, FOF and such were you get the finer details. Overall ratings could be set on most players, but it would have been top if the detailed fields was in the historicalfile. Just say, if not any number in those fields the AI will decide, but if you got a special running QB with speed rating he should pop up in-game with that.
My ever problem Lamar Jackson, as an inmobile QB doesn't sound nice. Sad.

Lots of players have asked for this update and I'm very close to release now, with starts in 2002.
The task has been to fill out the meager rosters that file had for 2002, that created loads of fictional names. Not fun to play a historical save then see a fictional QB take his team to a win in the Super Bowl. Not my cup of tea. I'm on adding first year players for 2018 and will go on with 2019 and 2020. That will take more time but like to get some hint on if this is a lost cause or not.

We all like this game to be so much more I guess, and I believe it can be.

@dev please answer this! I am so happy somebody is updating the historical file (Which I am still saddened that you guys stopped doing)

[Rebel]

Where can I get my hands on a historical file? I'd like to take a look at it.