The 1933 Project

[gambler]

Anyone feel like seeing how far we can take MaxFB? Sure, the game has a few issues, but playing around a bit, it's gotten me excited again about a project I had in mind. Starting a league back in the 30's and slowly moving towards the present. So I began making a checklist of what's needed:

1. Old Style Uniforms: Probably the easiest conceptually to do. Just needs someone with artistic talent.

2. 1930's Offensive and Defensive formations: Already checked out the PDS and made a few plays out of the Single Wing, A Formation and T Formations. Double and Triple Wings should be doable. Some of the formations won't be perfect since there wasn't a requirement to 'cover' both Tackles in the NFL back then but they can be close. 8, 7 and 6 man fronts on defense were the norm.

3. 1930's Offensive and Defensive plays and playbooks: Heavy emphasis on run. Trick plays tended to involve laterals and not forward passes.

4. Ratings to fit: The 1930's player is interesting. Most played 2 way still. Quarterbacks were evolving from basically a blocking fullback to a thrower. Halfbacks were often the primary thrower. I'm eyeing a Perl script to create the player database from which to have the initial draft.

5. Ability to have players play both ways: This is doable. MaxFB lets you put any player into any position.

6. Expansion & League Realignment: Not sure how doable this is. You can't do it WITHIN MaxFB. The database tables make it look like it can be done however. This needs to be tested.

So... Anyone interested in participating? I'm going to be working on the creation of Perl scripts to make the initial players and have already begun creating plays. I especially would like to have some talented uniform creators and people willing to make plays/playbooks, brainstorm on rating ranges, help in testing the game to see how close to the era we seem to be, etc... I'll give a teaser on what I've been doing so far...

[gambler]

Code: Select all

           SPD   STR   AGI   INT   END   DIS   BLK   BLA   PSS   PSA   PPR   CAT   HND   COV   KST   KAC   
 QB (BLK)   35    40    35    30    85    30    40    40    35    25    10    10    25    20    10    10
            60    65    60    80   100    80    65    65    60    50    35    35    50    45    35    35
 
 QB (THR)   40    35    40    30    85    30    25    25    45    35    20    10    25    25    10    10
            65    60    65    80   100    80    50    50    70    60    45    35    50    50    35    35
 

The two numbers in each column represent the high and low values.
Basically the Perl script would create approximately 2-3 QBs per team. Say 3/4 would be the 'BLK' (Blocking) type of quarterback while 25% would be the 'THR' (Throwing) type. Notice that even the throwers aren't going to be spectacularly efficient (in fact 70 might be too high for the era). I'm going to try to set the script to create only around 400 total players. The largest roster in 1933 was 33 players and there were 10 teams, so 400 should be about the right amount.

The numbers in general will be much lower than the current era, except in things like Endurance (which needs to be high to represent the players playing both ways).

[gambler]

And I've now created my first 15 quarterbacks using a pretty simple algorithm. Complexity was added slightly by breaking up the attributes into 3 'skill groups' (strength, dexterity and intelligence) and randomly determining if a player is poor, below average, average, above average, great at the skill group with a subsequent addition to the determined value as follows:

STR (Str, End, Blk, Bla, Pss, Kst)
DEX (Spd, Agl, Psa, Cat, Hnd, Kac)
INT (Int, Dis, Ppr, Cov)

Poor: -6 to -10 in attribute
Below Average: -1 to -5 in attribute
Average: No change to originally generated attribute
Above Average: +1 to +5 in attribute
Great: +6 to +10 in attribute

So to generate an attribute value I:

1. Determined what type of quarterback the player is (Blocking Back(75%) or one of them newfangled Throwers (25%)). Ex: Blocking Back
2. Determined their 'skill group' level for STR, DEX and INT groups. Ex: STR (Above Average), DEX (Average), INT (Poor)
3. Created the original attribute value for each attribute using the ranges listed in the previous post. Ex: BLK (or. attr val = 52; possible range is (40-65)), PSA (or. attr val = 40; possible range is (25-50)).
4. Modified the original attribute value based on their 'skill group' level for that attribute. Ex: BLK (Part of the STR group, player is above average so adds +1 to +5.. randomly determines it's +3 so final value = 52 + 3 or 55), PSA (Part of the INT group, player is poor so subtracts -6 to -10, randomly determines it's -9 so final value = 31).

Should lead to a fair amount of variety among players. I've been having more fun doing this than anything else with MaxFB so far. Which isn't to say it's not a fun toolkit/game, because it is, even with the glitches.

[Shaggyra]

I would be extremely interested in this project. Is there any place you need help?

[gambler]

Primarily in the areas of play/playbook creation, testing them with the players generated and/or graphics. The nice thing about the time period is the plays themselves are generally straightforward. Lots of inside runs, a sweep or three, simple pass plays. Defenses geared to stop the run. If you like low scoring games, this is your era.

[gambler]

And one more teaser... You may have to drag the browser sideways to see all of the table. This is the QB.txt file created via Perl to give me an idea how things look (and it perfectly updates the Players table for the league.mdb database). The ID is straightforward as is 'Type' (this is the first time I got more than 2-3 Throwing QBs. The first three (STR, DEX and INT) fields are the 'Attribute Groups' I referred to above with (-- = poor, - = below average, - = average, + = above average and ++ = great). I'd tag QB11 as a potential 'revelation' if he gets drafted by, oh, the Chicago Bears who started their newfangled 'T Formation' with the QB as a passer rather than a blocker. He'll probably get drafted high by any other team in the era as 'BLK' will be a high priority in QBs for most teams, but his special abilities probably won't be utilized... Of course the irony is that he's tagged as a 'Blocking' QB but what probably happened was he 'rolled in the high range' for PSS and PSA and combined with his ++ in both STR and DEX, he turned into an 'all around QB' for the era. Except for that rather pathetic Intelligence.

Code: Select all

 ID    Type      STR  DEX  INT  SPD  STR  AGL  INT  END  DIS  BLK  BLA  PSS  PSA  PPR  CAT  HND  COV  KST  KAC
 QB0   Throwing  +    +    0    56   59   64   32   91   54   46   36   62   41   20   37   48   45   32   23
 QB1   Throwing  0    0    +    52   38   47   41   92   53   29   30   49   44   25   20   47   40   33   27
 QB2   Throwing  ++   +    0    62   45   61   75   100  66   40   58   64   39   21   28   31   42   41   32
 QB3   Blocking  0    ++   0    56   53   50   61   88   66   61   57   53   48   30   30   37   20   21   35
 QB4   Blocking  0    0    0    47   64   50   62   85   37   62   63   44   39   20   30   29   23   28   34
 QB5   Blocking  +    +    0    52   58   56   55   88   78   60   58   54   49   14   19   30   26   17   16
 QB6   Throwing  0    0    0    56   49   59   53   96   34   40   48   45   52   37   13   44   38   14   35
 QB7   Throwing  0    -    +    59   45   48   54   93   66   28   46   57   57   30   29   40   34   15   25
 QB8   Blocking  0    0    -    42   62   57   44   85   76   53   63   36   45   29   29   31   26   15   28
 QB9   Throwing  +    0    0    56   55   65   31   98   32   41   36   59   46   35   12   39   32   24   22
 QB10  Blocking  --   0    +    54   31   38   78   83   60   46   40   53   35   39   14   39   29   4    32
 QB11  Blocking  ++   ++   -    56   46   66   36   98   42   71   62   61   57   19   42   37   39   43   36
 QB12  Blocking  0    +    0    47   61   47   68   85   65   56   48   37   53   22   25   44   30   24   17
 QB13  Blocking  ++   ++   -    46   51   45   63   98   49   57   62   64   40   17   40   54   22   20   20
 QB14  Blocking  ++   -    -    46   60   34   31   93   33   50   67   68   44   24   25   36   41   30   17
 

[Deft]

You should check with Old Coach. I know he has been creating 40's ERA college plays.

[Phatguy]

Once this game is patched up a bit to get rid of the most obvious bugs I plan on getting it and will then help you out any way I can. Did you consider going back to to the NFLs' 1st season in 1920(i think)? Since it sounds like you can do a lot with the database it should be doable. Cover the whole history instead of starting in 1933.

[nmleague]

Actully from my experience so far with the game I think your project would work great right now. The limitations I have found with the playbooks etc.. would be minimized by the type of offenses and defenses your talking about. The data base needs to be edited especially for speed ratings as I have mentioned in another thread.

[gambler]

I thought about it, but I'm personally most interested in the era when football's 'Old' era started picking up elements that you'd find in the modern era. Nice thing about the scripts I'm writing, it's easy to go in and tweak the ranges, number of each type of player, etc, so if I decide to go back to the 20's I can.

One other big thing that needs to be addressed is how the sim engine handles such a radical change. We already know it comes out with decent overall numbers at the default setting, which match more or less a 'watched' season (or at least it seems like that's the case). Hopefully the sim engine is flexible enough to handle a majorly different set of plays and ratings...

[thegipper]

I've always wanted to play a classic collegue league with the Four Horsemen in the backfield. I'd be glad to help.

[gambler]

An update of where I'm at:

1. Perl Script to generate players is basically done. Really the only things to deal with at the moment are a better spread of players by position and tweaking the ranges.

2. Draft Profile Perl Script was made to assign one of 6 different profiles (created in the program) to the teams instead of having to change each manually. I'm going to completely redo this though, basically replacing the in game draft engine with my own which will allow teams to be much more flexible in their choices and give the user a lot options to customize.

3. Some A-Formation plays, a basic 8-3 defensive formation, one Single Wing play, one T-formation play.

My goal right now is to get the game playing 1930's football, so I'll put the draft profile script on hold for a bit (I think I'll settle into a 'code like heck' mode on the weekends and things like play generation and testing ratings with the plays during the week). One annoying issue was Training Camp. The PlayersSA table (the one that shows the potential ratings of players) seems to be generated after the draft, not when the players are first generated, and the potentials are much higher than the rating range I want.. So when I run training camp, I end up with players much higher in their important attributes than they should be. So I have to do the following to start a new 1933 league:

1. Start New Season
2. Run the Player Generation PerlScript (creates players for each position desired in the number desired with custom ratings, and deletes the rest of the players).
3.Run the Draft Profile Assignment Perlscript
4.Conduct the draft
5.Run a Perlscript to replace all potential values with the actual values of the players (eventually this will create more appropriate potentials, but for purposes of testing ratings this works fine).
6.Set ALL teams to human controlled.
7.Run Training Camp (the training camp section becomes 'useless' due to step 5).
8.Cut players and set depth charts. If teams aren't set to human controlled, this step causes the game to hang (likely due to not having players at a position that the game 'needs' for the depth chart).

I usually make multiple copies of the database after most steps just in case something goes wrong.

What I'd like to do is upload the database I create along with any plays to a site (I'll ask around, see who can host). I'll also upload the perlscripts but right now I have no documentation on how to set up Perl and ODBC on your system (if you don't know what I'm talking about, you aren't going to be able to use the scripts). I will work on the documentation soon but you'll still need to be reasonably savvy in a Windows/Command Line environment. Anyone who wants to create plays/test will be welcome to do so. Please note I'm not looking for plays that 'break' the game's AI. I just want to recreate the 1930's so while it's probably possible to make a play that scores a TD 3/4 of the time against the defenses of the era using the formations of the era, that's not what I want.

An interesting issue is substitution. The rule back then was if a player came out of the game on any play, they had to stay out for the rest of the quarter. This means a player had to play offense, defense and special teams. I very much doubt the computer can handle this so my likely fix will be just to set all depth charts manually and see if the CPU leaves them that way in game and whether I can live with the results or have to control each team's substitution in game.

Oh yeah, and nmleague (or anyone else)... If you want to make some 1930's uniforms, I have two you could work on:

Packers and Spartans (Lions) from a rather low resolution color drawing, but it gives the idea.
http://www.sportsecyclopedia.com/nfl/po ... rtans.html

[gambler]

thegipper: Thanks for the offer! I'm focusing more on the pros, but during the 30's the two weren't anywhere near as distinguishable (college usually experimented more, talent level certainly much more comparable I think).

New status update. Well, I said I would code mainly on the weekend. I tested a few things out and ended up coding most of last night, but relatively futilely, attempting to get Perl to do the new draft mode I have in mind. My problem is I'm a Perl newbie and the draft as I plan it is at least an order of magnitude more complex than the player generation script I already wrote. The nifty thing though is I also just started using Perl at work, and something I learned today should make the draft script a lot easier.

Once that's done, it'll make testing easier as I can recreate the game's database without even starting a new season in game after I make substantial changes to ratings for instance.

I did work on some plays, primarily a base 7-2-2 defense and a couple of short pass plays. The nice thing is the runs tend to be short (1-5 yard range with the occasional stuff behind the line and the long gain) the passes in the 4-12 yard range... one of the problems in testing is that I'm using the depth charts... which often mean the same player is lined up against himself... and in the era the 'competing' skills can often be similar). So fairly soon I hope to actually play out a game seeing how the plays work when unequal players go up against each other.

So far for plays I have:
Base 8-3 defense
Base 7-2-2 defense

A-Formation (Run off center)
A-Formation (Run off RG)
A-Formation (Sweep Right)
A-Formation (Short flat pass)
A-Formation (Medium slant pass)

I need one or two runs to the left, a run off RT and a trick play and I'll call the A-Formation done for now. Probably will make an inside run stuff (was my first attempt with a 7-2-2 but it got beat constantly for big yardage) designed for short yardage, a couple of pass defenses and 1-2 variations on the base defense. And then do similar for a 6-3-2 defense and make plays for the Single Wing, Double Wing and Triple Wing formations in about the same number and type as the A-Formation. I'll probably save the T-Formation until I test out a game and get results I'm happy with, hopefully by this weekend!

[gambler]

So I got the draft utility working more or less which was very cool. A couple of minor glitches to work out but basically it works just like mentioned above. Players change weights and criteria for different positions in a text file and then just run the script and boom, it autodrafts for you (a 33 round draft in about 5 seconds - admittedly with a lot fewer players to search through, my league first starts with a database created already, then re-does the first 400 or so players in the player creation script, and deletes the rest). Eventually I'll want to make the draft script do the following:

1. Do a random draft order each round for league startup (to simulate the chaotic nature of scouting and who's available to pick from in the era).
2. Right now it picks the first player to have the 'high value' (it checks each player against the criteria and weights and whether the player is a starter or backup, and if the value is higher than the currently saved highest player it replaces them, until it's gone through all players). Instead I'll have the script save all of the currently highest rated players, and then pick one randomly.

I changed the ranges a bit in player creation and when testing the plays the spread of outcomes seemed much more hit and miss... the runs were either stopped behind the line or were broken open for big gains. Not too pleased about that... Passing actually looked better, the passer (HB) is pretty likely to take off and run.

[gambler]

Ok... Update time.

I slipped on getting my utilities out this weekend because I spent too much (frustrating) time on trying to get the game to play 30's era football. The big problem has been the tendency of the CPU to:

A) Substitute... This seems to only happen when Injuries are turned on and may actually only occur when an injury occurs. More testing is needed here.

B) Flip formations... I discussed this in more detail in the Support section. Basically because of the unorthodox formations, with the QB as another blocking/running back primarily and the HB as the primary ball handler (runner & thrower), whenever the CPU 'flips' the play, it doesn't come out symmetrical and the play is FUBAR, even more so than with a usual flip. If anyone knows how to turn off the ability of the CPU to flip a play, please let me know.

I had one game with stats pretty darned close to what I would expect in reality, at least for the running game. But that was deceptive, because all the 'sacks' (and 90% of the pass plays became sacks.. still need lots of tweaking on the passing front) were calculated into the running totals but not each player's individual run total. Most of a player's yardage came from a few long runs. So I tweaked ratings down in speed and agility, decreased defensive players' ability in BLA, and tried again and got a disaster.. Nobody broke the line of scrimmage except on a few 5-8 yard plays. So somewhere in between is obviously required. Plus it was made more difficult to evaluate because of all the FREAKING play flips, most of which blew up a play before it started.

The passing game actually works about right except completions are too common. This wouldn't be a problem if the game properly kept track of incompletions that didn't pass the line of scrimmage (it doesn't... I think they're recorded as tackles.. I heard this was fixed in a recent update though).

So... back to the drawing board.

Item #1 is to redo the playbooks from scratch. Until I figure out (or the patch team figures out) how to stop the CPU from flipping plays, I'll have to make the formations more traditional (with the QB as the primary ball handler, the HB as another blocker/runner type, etc) and just have the ratings be appropriate for the 'position' of the era (ie, the QB can throw but is really a running back, the HB is more of a blocker, etc). Then I'll have to make the blocking of the line symmetrical on either side in case the play is flipped (because ONLY the skill players flip... so if the line blocking is different on either side of the Center, your play probably blows up).

Item #2 is to tweak the speed/agility ratings of the ball carriers to achieve proper running.

Item #3 is to download the beta patch and see how the passing stats look when incompletes aren't counted as sacks.

[Old Coach]

TheGipper,
I like the old style uniforms you've created. Would you email me or post at FBMAX.com your templates for the jerseys and pants? I have a template for cross pattern, leather helmets that might be of use to you. I'll post those templates on FBMAX tonight. I can email them to you also if you prefer.

I have a single wing playbook posted at FBMAX that you might find useful as a guide for play creation.

Old Coach

[gambler]

Old Coach: Do you ever notice the Single Wing formation you made 'flipping'? Is the formation your own creation or plays built off the one included in the game? I'll download in either case to take a look, thanks!

Thing is, all my A-Formation, SW and DW plays seem to work in Practice and then just blow up in a 'watched' game, mainly when the formation 'flip' occurs.

[gambler]

Ok... After redoing the A and Single Wing formations, swapping the QB and HB, things are better. It still flips the formations 50% of the time but by making the line blocking more or less symmetrical, it's not as big of an issue. I also upped my ball carriers' Agility and Speed a tad. And got the final results in my first 'test league game'.

Chicago Cardinals: Led by an unusual halfback, Walrond, who has a better arm than most players in the league but isn't much of a runner, the Cardinals also have a couple of players with above average catching ability (ie, in the 50's range). Probably the stud of the team is Mangiamele, playing Quarterback and Weakside Linebacker, with superior speed and agility. The offensive line is anchored by a mammoth of a man, Pillion, standing at 6'3" and topping 240 lbs. He also is a terror on the defensive line. The Cardinals also arguably have one of the best fullbacks in the league, Folger, who is incredibly strong. The rest of their line is somewhat weak though, making them more of a finesse team. They might throw more than 15 times a game!

Philadelphia Eagles This team is a mirror image of the Cardinals in some ways, sporting an impressive player at QB/WLB, Twilligear. They also have a good fullback, Peveler and a strong halfback, Kudla. While their slotback, Birchwood, has good hands, Kudla isn't much of a passer. Their offensive and defensive lines are a bit stronger than the Cardinals, but their defensive backs probably don't match up very well against the explosive passing game Chicago brings into Comiskey Park.

The Cardinals came out throwing, Walrond completing 5 of 6 passes and quickly putting up 70 yards in the first quarter. Their running game bogged down, but the Eagles were completely stymied and the first half was played mostly in their end of the field. The game was 14-0 at halftime, Walrond the undisputed star with 1 touchdown pass of 8 yards to Folger, and a run up the middle of 12 yards for another.

The game stayed the same into late in the 3rd quarter when Twilligear broke a 65 yard run up the middle, tackled just before the goal line (by a player about 20 points slower in speed I might add... who RACED up while Twilligear suddenly seemed caught in molasses... Grrrr...). While Philadelphia was stuffed and then promptly missed the field goal from the 1 yard line, momentum had shifted. They held the Cardinals and took over well inside Chicago territory, and helped by a couple of Cardinals' penalties, they moved into position for Kudla to plunge in on a 5 yard run. The extra point was missed. The Cardinals were able to move again on a couple of broken runs by Walrond against a thoroughly confused Philadelphia defense as they focused too much on the receivers and not on the ball carrier. This led to two field goals which sandwiched a Philadelphia field goal, leaving the final score at 20-9, Chicago.

Code: Select all

               Chicago       Philadelphia
 First Downs     13               6
 Rushing Yards  150              126
 Passing Yards  126               0
 Net Yards      231              106
 
 Plays           53               41
 
 Rushing Att     36               41
 Average         4.2              3.1
 
 Passing Att     17               0
 Passing Comp    11               0
 

Code: Select all

 [b]Notable Statistics[/b]
 CHC: Walrond (11/17 - 126 yards - 1 TD - 0 INT; 12 rushes - 147 yards - 2 TD; 8 Tackles)
      Folger (5 att - 31 yards; 2 Rec - 11 Yards - 1 TD)
      Hibley (6 Rec - 90 Yards)
      Mangiamele (11 Tackles)
 
 PHI: Kudla (27 att - 67 yards - 1 TD)
      Twilligear (7 att - 73 yards; 11 Tackles)
      Peveler (11 Tackles)
 
 

Notes: Need to tweak the defenses to follow the ball carrier (just discovered the trick as posted here of setting defensive players to 'Move to Next Point' to have them do so), otherwise the roll out pass plays lead to too large of a gain.. Almost all of Walrond's yardage came on these plays. I also need to find the range in which the passing will lead to lower completions (the best of the era were under 50%) but also actually pass the ball (Kudla never attempted a pass even though the team ran a pass play maybe 10 times - same plays that the Cardinals had).

Otherwise I'm actually very pleased finally.

Oh yeah, quarters were 11 minutes to match the approximately 50 plays a game for each team that 1933 saw.

And finally, before anyone asks, no these are not historical rosters.

[Old Coach]

Gambler,
The 11 man plays included with the game ARE all my plays. I created those for MaxFB over the beta test years. Some of the plays have been in there since beta test started. The fun and gun playbook that Erik posted on the Members link is the updated playbook. The single wing playbook posted over on FBMAX is a playbook that I had been working on for several months and then revised to work better with the newer code.

I've not seen the flipping problem occur because most of my playtest is me playing against someone else or I call the plays for both teams.

Old Coach

[gambler]

Alright, I'll grab them... I'll be curious to see the blocking schemes on the runs particularly.