Front Office Football Central

Front Office Football Central (http://forums.operationsports.com/fofc//index.php)
-   FOF9, FOF8, and TCY Discussion (http://forums.operationsports.com/fofc//forumdisplay.php?f=33)
-   -   QB/WR Nerf Draft File and Full League Player File Generators (http://forums.operationsports.com/fofc//showthread.php?t=82542)

Ben E Lou 10-14-2011 10:13 AM

QB/WR Nerf Draft File and Full League Player File Generators
 
NEW LINKS:

Draft Class Generator: http://www.younglifenorthdekalb.com/....php?year=2011

Full League File Generator: http://www.younglifenorthdekalb.com/...egenerator.php




First off, all credit to QuikSand for floating this idea in the FOxL Lounge the other day.

I didn't realize it until the last 48 hours or so, but we can "fix" the over-importance of the passing game in FOF MP (and SP, if anyone cares) in a much more eloquent and precise manner than what's been done in the past. The Draft File Generator from FOF2K4 allows the input of a csv. I've now created a php app that will generate a random FOF draft class csv with a bit of a nerf for QB/WR talent. Input that csv into Draft File Generator, then Import the resulting .faf file into FOF2K7, and voila.





Here's a list of what I'm doing with the file that I'd consider improvements:

PLAYER FILE GENERATION NOTES

THINGS THAT WILL AFFECT GAMEPLAY/STRATEGY
  • QB/WR overall rating no higher than 60. (Keep in mind that X-Factor and/or VSOL can push these ratings up, though.)
  • BPR for WRs no higher than 60. (See above. X-Factor and/or VSOL can increase it.)
  • Special Teams Rating not tied to overall player talent, so more ST specialists will exist. Could make for more interesting roster decisions.
  • Endurance is tied to height/weight ratio, so no longer will you get a 5'11", 180-pound NFL-caliber CB with 10 Endurance.
MINOR/COSMETIC STUFF
  • Much larger pool of potential names.
  • Huge majority of draftees are ages 21-23.
  • Greatly reduced chance of bar oddities like guys with 85 Long Passing and 5 Very Long Passing.
  • Just about eliminated the chance of 1.1 pick coming from a tiny school.
  • Players at some positions are taller than in standard FOF draft classes, better reflecting current NFL.
The primary difficulty with this undertaking is that Jim's documentation regarding talent distribution is outdated/wrong/something. What you see ain't what you get, so I've had to make some educated guesses and assumptions regarding the distribution of talent. And that's what I need tested.
HOW TO TEST THIS
1. Browse to http://www.younglifenorthdekalb.com/....php?year=2011 to generate a draft file csv. (TO GET A DIFFERENT DRAFT CLASS YEAR, CHANGE THE YEAR IN THE TITLE TO THE YEAR OF THE CLASS YOU WANT. I didn't have time to include an input field for this. I'll do that at some point this weekend.)
2. Download the csv to your machine.
3. Open the Draft File Generator from FOF2K4. (If you don't have a copy of FOF2K4, just download it from Solecismic.com here: http://www.solecismic.com/orders/Fro...otball2004.exe
4. Hit "Generate Draft File" in the Draft File Generator. It will prompt you to select the csv that you just downloaded from my app, then it will generate a draft file (.faf) for FOF.
5. Immediately prior to the draft class being generated, (i.e. during the Franchise/Summer/Ticket stage), the "Import Draft Class" button appears. Import the .faf file you generated in step 4.

That's it. When you hit "Begin Free Agency," the draft file will be in your game. You can test however you want, but I recommend that you let the AI handle everything so you can blow through seasons and import new draft classes. I'm most interested in the talent makeup of a league after 15-20 years of using classes generated with this app. Too much talent, too little, just right?

Thanks in advance, and I appreciate any feedback I get on this.

--Ben



Ben E Lou 10-14-2011 10:18 AM

Dola:

I've coded this in a way that tweaking the player pool talent is very easy, so the more feedback, the better. I really want to get this "right."

MIJB#19 10-14-2011 11:12 AM

Quote:

•Greatly reduced chance of bar oddities like guys with 85 Long Passing and 5 Very Long Passing.
I like that thinking for a some positions, like your example. Care to elaborate more on this for other positions?

Ben E Lou 10-14-2011 11:28 AM

Quote:

Originally Posted by MIJB#19 (Post 2549041)
I like that thinking for a some positions, like your example. Care to elaborate more on this for other positions?

Don't have time to dig back into the code right now, but the two I remember off the top of my head are RB and Speed to Outside/Breakaway Speed, and DB coverages. Guys with 70 Breakaway and 10 Speed to Outside will be far more rare, as well as guys with 90 M2M and 15 Bump.

If thing takes off a bit, I can definitely do more to make players' skill sets feel more realistic. An idea that I have is creating certain "types" at positions for non-star players. In other words, say it creates a tight end with overall rating less than 60, why not have it then roll the dice to choose between Speed TE, Blocking TE, Possession Receiver TE, and Balanced TE? So maybe I roll a TE who is rated 55/55 overall, but he's very good in Avoid Drops, Courage, Route Running, and Third Down Catching, and mediocre or worse in every other category.

chinaski 10-14-2011 12:17 PM

I have some time today, ill sim out 20 or so seasons. Would you like the league data tracked?

Ben E Lou 10-14-2011 12:31 PM

Quote:

Originally Posted by chinaski (Post 2549100)
I have some time today, ill sim out 20 or so seasons. Would you like the league data tracked?

Thanks!

First off, wait a few minutes. I'm going to upload some changes in a few moments. Overall talent is too low, and I'm increasing it a bit.

Second, I'm not terribly interested in tracking stats, but more in what the talent distribution of a league looks like 15-20 years down the road.

I'll let you know when the updated code goes up.

chinaski 10-14-2011 12:36 PM

Ok, will do. First thing i've noticed is the top 3 QB's in the first draft had high long pass, but very low deep pass. The highest rated QB is 29/66.

Ben E Lou 10-14-2011 12:38 PM

Ok. It's there. Go for it.

Ben E Lou 10-14-2011 12:40 PM

Quote:

Originally Posted by chinaski (Post 2549113)
Ok, will do. First thing i've noticed is the top 3 QB's in the first draft had high long pass, but very low deep pass. The highest rated QB is 29/66.

Oh, one thing to keep in mind is that FOF is going to apply masking to some players. I should have suggested one other thing: do the testing on Main Street level, so you can see closer to real ratings. Or just don't pay much attention to the bars until a player is at or near full development. ;)

Ben E Lou 10-14-2011 12:43 PM

Dola:

ONe other thing, chinaski. I have a way to generate multiple csvs for multiple seasons in one shot. What draft class years do you want? I'll create 25 or so csvs for you and zip 'em up so at least you don't have to do that part.

TRO 10-14-2011 12:55 PM

This is potentially revolutionary BEL, thanks!

chinaski 10-14-2011 01:05 PM

Quote:

Originally Posted by Ben E Lou (Post 2549119)
Dola:

ONe other thing, chinaski. I have a way to generate multiple csvs for multiple seasons in one shot. What draft class years do you want? I'll create 25 or so csvs for you and zip 'em up so at least you don't have to do that part.



Any years, thanks!

Ben E Lou 10-14-2011 01:08 PM

Quote:

Originally Posted by chinaski (Post 2549142)
Any years, thanks!

Ok. 2009-2039 coming your way. I'll send a link via PM in a few minutes. I'm not posting it publicly because if someone else wants to do this, I'll want them testing with new files.

gstelmack 10-14-2011 01:19 PM

Remember that you can use Extractor to pull all the player card info for all the player cards. Ben could then use that to analyze all the league talent.

Alf 10-14-2011 01:29 PM

Ben, what kind of feedback do you want us to collect ? And in which way ?

Ben E Lou 10-14-2011 01:50 PM

Quote:

Originally Posted by Alf (Post 2549158)
Ben, what kind of feedback do you want us to collect ? And in which way ?

Biggest thing is to look at overall talent.

1. Sim 15-20 seasons.
2. Sim to end of a regular season, so that all good players are on teams.
3. Look at each position's talent on the Recommend Player screen with Current Rosters (not Free Agents) and look for issues.

The kinds of issues you might see...
  • Too much "Bunching" of talent (i.e. 20 players rated 50/50-->55/55 at a position that only needs one like QB/RB/FB/TE)
  • Too much talent overall (like 15 guys rated 80 or better at a position)
  • Too little talent overall (no guys rated 80 or better at several positions)
Then start sorting on specific bars. Look for things like having no players with 70 or higher in zone, or having tons and tons of guys who excel in a skill. In short, I just want the file to be well-balanced, and because the documentation is way off on what numbers produce what kinds of players, I had to create the algorithms based on my own perceptions and guesses at how FOF distributes talent. I think I'm reasonably close (i.e. I'm not going to be embarrassed to have people take a look at this,) but I doubt it's as close as I want it to be, or as close as I can get it.

Ben E Lou 10-14-2011 01:55 PM

Dola:

An example of the kind of stuff to look for: in the first test league I tried, after seven draft classes there were zero players at any position rated above 77 or so. That's the kind of thing I'm talking about. Compare the talent in a league using these draft classes with the talent in a SP or MP league that has run for 15+ seasons and that uses FOF default draft classes.

Ben E Lou 10-14-2011 02:02 PM

I just remembered something: I think .faf imports work differently if you don't have TCY installed. I'm installing it now, and I'd encourage others who test to do so as well just in case. (I don't think you hvae to have it licensed, just installed.)

http://www.solecismic.com/tcyorders/...llegeYears.exe

QuikSand 10-14-2011 02:50 PM

Really cool idea, and I'm thrilled that it's much closer to do-able than I would have guessed. I may have some time to work with this a bit over the weekend.

Julio Riddols 10-14-2011 04:32 PM

This sounds pretty interesting. May be equal to a major patch update in what it adds to the game.

ozias 10-14-2011 06:47 PM

Ben,

So far I've done 12 seasons and here are some averages for them.

QB Current is 28 and Future is 60 The highest current for a rookie so far is 42, and the highest future for a rookie is 68.

RB 52 and 61 The highest current for a rookie was 61 the highest future was 68.

FB 38 and 57 with 44 and 73.

TE 39 and 57 with 45 and 65.

WR 36 and 58 with 43 and 67.

C 36 and 62 with 46 and 82.

G 38 and 62 with 58 and 82.

T 36 and 62 with 42 and 77.

P 43 and 60with 57 and 70.

K 41 and 60 with 58 and 71.

DE 38 and 59 with 46 and 72.

DT 39 and 59 with 48 and 70.

IL 37 and 60 with 42 and 68.

OL 42 and 65 with 47 and 79.

CB 51 and 65 with 69 and 74.

S 43 and 64 with 49 and 78.

None of these numbers include players from inside the game, they are only from your draft class generation.

After the 1st draft there was 4 players with a current above 50 and 83 players with a future above 50.

Here is the breakdown from each draft class(including the first) with how many players were rated 50 and above.

2011
Cur 4 Fut 83
2012
Cur 3 Fut 92
2013
Cur 2 Fut 76
2014
Cur 1 Fut 58
2015
Cur 0 Fut 56
2016
Cur 4 Fut 61
2017
Cur 1 Fut 59
2018
Cur 2 Fut 68
2019
Cur 3 Fut 92
2020
Cur 1 Fut 59
2021
Cur 4 Fut 59
2022
Cur 2 Fut 68

Year end totals went like this.
Cur Fut
32 155
46 163
65 159
83 158
113 181
117 199
140 230
151 255
167 251
180 243
181 259

I didn't copy the first year down, but you can see the current players above 50 increase every season, and the future move a little as players got older and/or retired. I think the current will get over 200 as I continue to sim more seasons, but I have yet to see an absolute stud at any position.

The 2019 season was the only season where I had 3 players rated 80 or better, and currently the highest is a 79 SS, who was an 82 in 2019.

The best QB is 67 followed by a 56.

Here is a breakdown of positions to go with the current total of 181 and future of 259 listed above.

QB 8 and 23
RB 19 and 25
FB 1 and 4
TE 9 and 14
WR 7 and 10
C 11 and 12
G 17 and 22
T 11 and 14
P 4 and 6
K 3 and 5
DE 16 and 20
DT 7 and 12
IL 10 and 18
OL 19 and 25
CB 19 and 24
S 20 and 25

As you can see some of the position groups are a little light on the 50+ players. Not sure if there is a way to raise certain position groups though, it may be worth leaving alone, so it would tend to be like the real NFL, where some years certain positions are lean in the draft.

I haven't seen any great WR's even though there is a 67 and 66 rated guy at the top, followed by 52 and lower. The BPR is 74 and 68 respectively as well and no one has a maxed out bar there.

For the QB's 67 is the leader followed by a couple of 56's and lower. No one with a maxed out bar here either, except a QB with 100 on 2 minute offense. It's actually a great variety of QB bars on all of the QB's, and the most balanced QB is a 12 year vet at 47 with just above and just below 50 for all of his bars, except RD at 96.

I would say, so far, it looks like you've found a way to nerf the QB/WR, but I think some of the other position groups need better players. I'll keep going with the testing, as I created enough draft files to get to the year 2040, so 18 more seasons should round the test out nicely.

Ben E Lou 10-14-2011 07:21 PM

Quote:

Originally Posted by ozias (Post 2549331)
Ben,

So far I've done 12 seasons and here are some averages for them.

QB Current is 28 and Future is 60 The highest current for a rookie so far is 42, and the highest future for a rookie is 68.

RB 52 and 61. 61 was the highest current for a rookie and 68 was the highest future.

FB 38 and 57 with 44 and 73.

TE 39 and 57 with 45 and 65.

WR 36 and 58 with 43 and 67.

C 36 and 62 with 46 and 82.

G 38 and 62 with 58 and 82.

T 36 and 62 with 42 and 77.

P 43 and 60with 57 and 70.

K 41 and 60 with 58 and 71.

DE 38 and 59 with 46 and 72.

DT 39 and 59 with 48 and 70.

IL 37 and 60 with 42 and 68.

OL 42 and 65 with 47 and 79.

CB 51 and 65 with 69 and 74.

S 43 and 64 with 49 and 78.

None of these numbers include players from inside the game, they are only from your draft class generation.

After the 1st draft there was 4 players with a current above 50 and 83 players with a future above 50.

Here is the breakdown from each draft class(including the first) with how many players were rated 50 and above.

2011
Cur 4 Fut 83
2012
Cur 3 Fut 92
2013
Cur 2 Fut 76
2014
Cur 1 Fut 58
2015
Cur 0 Fut 56
2016
Cur 4 Fut 61
2017
Cur 1 Fut 59
2018
Cur 2 Fut 68
2019
Cur 3 Fut 92
2020
Cur 1 Fut 59
2021
Cur 4 Fut 59
2022
Cur 2 Fut 68

Year end totals went like this.
Cur Fut
32 155
46 163
65 159
83 158
113 181
117 199
140 230
151 255
167 251
180 243
181 259

I didn't copy the first year down, but you can see the current players above 50 increase every season, and the future move a little as players got older and/or retired. I think the current will get over 200 as I continue to sim more seasons, but I have yet to see an absolute stud at any position.

The 2019 season was the only season where I had 3 players rated 80 or better, and currently the highest is a 79 SS, who was an 82 in 2019.

The best QB is 67 followed by a 56.

Here is a breakdown of positions to go with the current total of 181 and future of 259 listed above.

QB 8 and 23
RB 19 and 25
FB 1 and 4
TE 9 and 14
WR 7 and 10
C 11 and 12
G 17 and 22
T 11 and 14
P 4 and 6
K 3 and 5
DE 16 and 20
DT 7 and 12
IL 10 and 18
OL 19 and 25
CB 19 and 24
S 20 and 25

As you can see some of the position groups are a little light on the 50+ players. Not sure if there is a way to raise certain position groups though, it may be worth leaving alone, so it would tend to be like the real NFL, where some years certain positions are lean in the draft.

I haven't seen any great WR's even though there is a 67 and 66 rated guy at the top, followed by 52 and lower. The BPR is 74 and 68 respectively as well and no one has a maxed out bar there.

For the QB's 67 is the leader followed by a couple of 56's and lower. No one with a maxed out bar here either, except a QB with 100 on 2 minute offense. It's actually a great variety of QB bars on all of the QB's, and the most balanced QB is a 12 year vet at 47 with just above and just below 50 for all of his bars, except RD at 96.

I would say, so far, it looks like you've found a way to nerf the QB/WR, but I think some of the other position groups need better players. I'll keep going with the testing, as I created enough draft files to get to the year 2040, so 18 more seasons should round the test out nicely.


Good stuff, ozias! A few things...

1. The way the code is working, any variations from position to position apart from QB/WR are due to either the way FOF interprets the file, or just dice rolls. When the overall talent rating is generated for a player, the code doesn't know what position he plays. (He already has a position at that point, but it's not a part of the initial "overall talent" calculation.) So I suspect that the "bad" and "good" positions you see are just dice roll related. That said, if we get consistent reports that , say, LBs have less talent than any other position, and no reports of LBs being solid, then we can be reasonably sure that FOF does something when the .faf is imported that causes LB talent to be lower, and from that it might be decided that I need to raise LB talent.

2. I'm not as much interested in the talent at the time of the draft, because FOF is definitely masking guys at that juncture. But once a class is in, say, year 5 or so, then those numbers are much more important.


3. Quick question: that far in, what do overall league stats look like, particularly QB Rating, ppg, and ypa?

4. Yeah, I'm 100% confident in the QB/WR nerf, 'cause it's very simple and straightforward, and only two lines of code.

ozias 10-14-2011 07:58 PM

The past 5 years of stats go as follows with the QB I just went with the most Yds and TDs per season followed by the range of all QB's completion percentage.

Code:

Yards  TD  Completion %
3831    25  54-65%
4037    26  53-64%
3761    22  52-68%
3807    27  51-63%
4165    24  54-64%


RB's

Code:

Yards  TD
1444  11
1395  13
1267  14
1813  11
1513  13


WR's
Code:

Yards  TD  Catches
1559  13  126
1330  11  111
1207  10  104
1159  11  96
1275    9  103



Defensive stats, just listing the leaders for the past 5 years.
Code:

Tackles  Sacks    Int
104      10        6
136        9.5      6
105      10.5      5
104      15        7
128      12.5      9



All purpose yards for the last 5 years as well.

2228
2143<--the was done by a WR
1956
2445
2340

If there is any particular stat you really want let me know and I'll post it as well.

I'll get the QB Rating, ppg and ypa up in a few.

Ben E Lou 10-14-2011 08:00 PM

Thanks. I'm not looking for stats from individual players, fyi. I'm talking about league-wide stats that you can see at the bottom of the Team Statistics page.

ozias 10-14-2011 08:13 PM

The QB rating for the past 5 seasons, best to worst.

Code:

92.7 - 62.1
92.0 - 57.4
90.0 - 58.7
94.2 - 57.0
93.4 - 57.4


League averages for Yards/Comp and Yards/Att past 5 seasons

Code:

Y/C  Y/Att
9.80  5.82
10.0  5.81
9.81  5.94
10.11  6.04
10.02  6.00


League rushing averages for past 5 seasons.

4.24
4.31
4.24
4.16
4.13

Average PPG for past 5 seasons

17.6
18.5
17.8
19.1
18.4

ozias 10-14-2011 08:17 PM

Here are the average completion percentages as well.

59.3
58.1
60.5
59.8
59.8

ozias 10-14-2011 08:21 PM

Here are the average rush and pass yards per game.

Code:

Rush    Pass
104.9    219.7
107.5    222.0
110.2    215.3
111.7    211.8
109.0    211.0


ozias 10-14-2011 10:20 PM

Here are the rest of the stats thru the year 2040.

Not much deviation which is a good thing.


Code:

    Rush      Pass
Year Avg  Avg/G  Yd/C  Y/Att  Avg/G  Comp%  PPG
2026 4.25  109.0  9.85  5.82  213.2  59.1  17.8
2027 4.26  107.8  9.87  5.96  219.1  60.3  18.6
2028 4.16  106.0  9.61  5.63  208.4  58.6  17.4
2029 4.14  104.5  9.80  5.78  212.6  58.9  17.8
2030 4.14  105.3  9.92  5.90  218.8  59.5  18.0
2031 4.17  107.4  9.88  5.99  218.1  60.6  18.2
2032 4.22  109.3  10.01 6.04  219.7  60.3  18.9
2033 4.11  105.5  9.79  5.71  212.4  58.3  17.9
2034 4.15  106.5  9.92  5.84  215.9  58.8  18.5
2035 4.17  107.4  9.91  5.76  214.3  58.1  17.4
2036 4.13  107.3  9.82  5.76  211.4  58.6  17.9
2037 4.27  109.0  9.90  5.85  216.7  59.0  18.0
2038 4.15  106.5  9.73  5.68  211.6  58.3  17.5
2039 4.11  106.5  9.86  5.89  213.6  59.7  17.7
2040 4.20  108.3  9.76  5.90  214.7  60.5  18.7


I currently don't have anyone rated higher than 79, but one of them is a rookie and the other is a 6yr DE.

Here is the current list of the top 25 at each position.

QB 56 - 45 <--2 players with 59 future here
RB 70 - 47
FB 69 - 41
TE 75 - 43
WR 61 - 42
C 74 - 41
G 72 - 44
T 71 - 46
P 77 - 37
K 71 - 37
DE 79 - 44
DT 73 - 46
IL 79 - 46
OL 70 - 45
CB 74 - 49
S 62 - 46 <--there is a 75 future here

I think after 29 seasons that the above is a reasonable breakdown of players. Out of the Top 25, all positions are there except, QB, WR, and S, but the safety whose future is a 75 is a rookie and will make the list in a year or two.

The 2 positions that you wanted to nerf didn't make the top 25, and the best BPR is a 68 followed by a 61.

I like the way the stats look and the position distribution looks pretty good as well.

I never did come across the 90+ player, and only had a handful of 80+ players in the 29 seasons, so I'd say the 80+ are the rare once in a lifetime superstars, and the 70+ are your elite players, 60+ would be the great, 50+ are your good, 40+ is your backup/starter types, and 30+ should be your roster filler.

At least that's how the draft files turned out, and I think it could be a good SP challenge, depending on whether or not you game plan the cpu and/or house rules that may need to be established.

For MP, I think it would be much harder, as you are not going to come across great WR's or QB's with the draft files, and when those are out there in FA a bidding war will ensue. More than likely, none of those players will ever hit FA, unless a team is in cap trouble, but I'd think a trade would come along for that team, and that may just end up becoming the bidding war instead of money.

Now too see what others will come up with during there test runs, and whether or not were all in the same area, or if others come across the 90+ player(s).

All in all, I like what you've done with this test Ben, and it looks like it will make the game "new" again for a lot of people.

Ben E Lou 10-15-2011 02:17 AM

Nice work, Ozias!

I've been looking at this from a different angle, and I can say definitively that the talent apart from QB/WR is too low. And I have a theory as to why. I suspect that most of the documentation is correct and updated, but that one little section in it is wrong. In the Draft File Generator help, it states that the percentages should only be applied to the top 200 or so draftees. I suspect that somewhere along the way, Jim changed it so that those percentages are applied to the entire draft class. I've been working with percentages higher than Jim's, but only applying them to 200 players out of the class. I just ran a test with my percentages against the entire class, and Draft Analyzer shows this draft class being a bit more talented than a "normal" draft class, which would jibe with this theory given that my odds are a little higher.

So right now, I'm going to update the percentage table to reflect Jim's defaults, update the ratings ranges to Jim's, and update the code so that it applies those percentages to the entire draft class. We'll see how this turns out.

Ben E Lou 10-15-2011 02:24 AM

First draft class generated this way looks, on the surface, very much like a standard FOF draft class, just without any stud QBs or WRs. I'm going to run Analyzer on this one. If it comes out like I think it will, I'll update the main code so that everyone can generate this way.

Ben E Lou 10-15-2011 02:43 AM

Yeah, this looks much better. And it makes more sense that just one or two bits of the documentation would be incorrect or outdated rather than all of it, which is how it appeared initially. The code is updated. If you're interested in testing, please re-test. Ozias, chinaski, and Quik, I'm going to auto-generated 30 different draft classes for each of you now and send you links to them.

Ben E Lou 10-15-2011 03:13 AM

1 Attachment(s)
Yeah, this is the sort of top-of-the-class look that I had in mind.

Ben E Lou 10-15-2011 03:18 AM

Oh, one thing that needs a close look and likely a tweak is going to be the Endurance algorithm. It may be that I need to tweak it my positions. I suspect that offensive linemen in particular may not be getting quite enough endurance.

Anyway, on to some testing of my own now.

Ben E Lou 10-15-2011 04:50 AM

Y'know, I've been ignoring college stats in this, but it just hit me that we could introduce evaluating players by stats to this. I'd make it very transparent by sharing the correlation percentages, and of course the good ol' dice would be involved too. But I'm thinking something like...


RB
Rushing Attempts-->Average of All Rushing Bars: 65%, Endurance, 35%
Yards Per Carry-->Breakaway 50%, Hole Rec 25%, Elusiveness 25%
TDs---->Breakaway 33%, Hole Rec 33%, Elusiveness 34%


..and so on. I could generate the stats based on the real ratings, in other words. QBs with high avoid interceptions would be more likely to throw more picks in college, that sort of thing.

aston217 10-15-2011 07:07 AM

I think it'd be great - if this isn't part of it already - if you could adjust the degree of nerf-ness. I can imagine different leagues would have different needs. To me, the 60 BPR thing and QB<60 seems a bit harsh (though I wouldn't mind SP). I think for some leagues at least, the WR and QB adjustments should be a lot more minor.

I like the ST adjustment. I didn't realize it was tied to overall talent that much already.

Other changes: bar oddities, player heights, endurance - these make sense on a cosmetic level but I'm not sure about their effect on the game. The endurance thing is most concerning because it obviously affects talent distribution / balance in some way but I'm not sure how. So I'm not sure if it makes sense on a level other than a purely cosmetic one.

Quote:

Guys with 70 Breakaway and 10 Speed to Outside will be far more rare, as well as guys with 90 M2M and 15 Bump.


That doesn't make too much sense to me. There should be players who excel at one or the other. I can see some corners being very good in man coverage, but who struggle with the physicality of BnR. Or some RBs who have great straight-line speed, but don't do well running laterally.
.
I don't see a need for restricting player skills, and also not really a need for the stats part. The mystery of the draft is what makes it fun! College stats also do not mean too much.

This is all very cool though, so keep us posted.

Yoda 10-15-2011 07:23 AM

Is this because of the over inflated stats your offense puts up by running up the score?

Ben E Lou 10-15-2011 07:53 AM

Quote:

Originally Posted by aston217 (Post 2549737)
I think it'd be great - if this isn't part of it already - if you could adjust the degree of nerf-ness. I can imagine different leagues would have different needs. To me, the 60 BPR thing and QB<60 seems a bit harsh (though I wouldn't mind SP). I think for some leagues at least, the WR and QB adjustments should be a lot more minor.

I can't imagine that many existing leagues would want to use this, unless they're willing to just position-switch every existing WR and QB of consequence into oblivion. It'd create a huge imbalance until the league was purged of the existing players. Assuming I get this right, it'll be used for a new league, and I'd imagine some people will want it for SP. I know I will. I'm tired of the WR house rules I have to use, given that the AI has no clue about their value. Guys who'd go top 20 in the top FOF MP leagues are still sitting there in the late 3rd on a routine basis.
Quote:

Other changes: bar oddities, player heights, endurance - these make sense on a cosmetic level but I'm not sure about their effect on the game.
I am sure, especially when it comes to endurance.[/quote]The endurance thing is most concerning because it obviously affects talent distribution / balance in some way but I'm not sure how. So I'm not sure if it makes sense on a level other than a purely cosmetic one.[/quote]When's the last time you were watching an NFL game and heard the announcer say anything like "the QB picked on the backup corner on that play while the starter was out taking a breather?" It happens rarely in real life for some positions. RB and DT? Sure. Center? No. DB? No. But FOF routinely creates guys with awfully low Endurance and they're sitting out plays all over the place, with no real rhyme or reason.
Quote:

That doesn't make too much sense to me. There should be players who excel at one or the other. I can see some corners being very good in man coverage, but who struggle with the physicality of BnR. Or some RBs who have great straight-line speed, but don't do well running laterally.
Sometimes, sure. But not as frequently as this happens in FOF.
Quote:

I don't see a need for restricting player skills, and also not really a need for the stats part. The mystery of the draft is what makes it fun! College stats also do not mean too much.
After 5+ years with FOF2K7, there is little/no mystery to the draft for some of us. Adding more obvious information in the stats could be a significant equalizer for those who haven't caught up yet. Someone who doesn't dig into combine correlations might not understand or have time to get that high Solecismic plus low intelligence = high avoid interceptions. But if they see a QB with only 5 INTs in a full season, they'll be clued in that he's worth drafting.

Ben E Lou 10-15-2011 07:53 AM

Quote:

Originally Posted by Yoda (Post 2549739)
Is this because of the over inflated stats your offense puts up by running up the score?

No. The extra TD pass or two 4 or 5 games a year for just one team is merely a symptom of a much deeper issue. FOF in its current state remains fundamentally flawed with regard to the deep passing game being overpowered. Four of the first five picks in the IHOF Draft went WR. Six of 11 in FOFL the same. WRs who aren't even all that good are going top 20 all over the place. It's getting monotonous to see the same magic bullet in every league. Apart from a new game, Quik's idea here is very likely just the shot in the arm that FOF MP needs to get more interesting again.

aston217 10-15-2011 08:14 AM

Quote:

When's the last time you were watching an NFL game and heard the announcer say anything like "the QB picked on the backup corner on that play while the starter was out taking a breather?" It happens rarely in real life for some positions. RB and DT? Sure. Center? No. DB? No. But FOF routinely creates guys with awfully low Endurance and they're sitting out plays all over the place, with no real rhyme or reason.


Yeah, I completely get this not making any sense compared to RL.

Re-reading it though, I kind of like the idea of EN being tied into height/weight ratio, so the 6'4-335 lb DL chugs around and needs frequent breathers, while the 6'4-305 guy can get in on more plays. Is this something that introduces a lot more cohesion into the game, to the point that depth becomes less important, or young players don't develop because they are sitting for a lot more plays behind star vets? Or will it just redistribute the endurance to make more sense given a player's ht/wt profile?

Re: draft...you know what would be interesting, is if you could equalize in the other direction. By putting the mystery back in for everyone. Not sure if that's possible. But just randomly, with no tells whatsoever, nerf some guys who will come out looking great otherwise. Maybe take the obvious masks and turn them into 33% shots, where 66% of the time they will end up being mediocre or worse.

Or/and on top of that, take the top 100 players in the draft talent pool and jack up the volatility for say, 80 of them. Make players think twice at the top of the draft because there very likely won't be any "safe" options, the superstud with the 10 volatility to go to in case you're nervous about taking a 97 guy at 1.4. Just some random thoughts.

Ben E Lou 10-15-2011 08:37 AM

Quote:

Originally Posted by aston217 (Post 2549751)
Yeah, I completely get this not making any sense compared to RL.

Re-reading it though, I kind of like the idea of EN being tied into height/weight ratio, so the 6'4-335 lb DL chugs around and needs frequent breathers, while the 6'4-305 guy can get in on more plays. Is this something that introduces a lot more cohesion into the game, to the point that depth becomes less important, or young players don't develop because they are sitting for a lot more plays behind star vets? Or will it just redistribute the endurance to make more sense given a player's ht/wt profile?

Interesting questions there, and it of course is going to be partially dependent on the injury setting. If using 150 or higher, I'm guessing that depth is still as important even at positions like S where irl they don't come out of the game much at all. But yeah, if you have two starting 90-endurance safeties (which are unquestionably more frequent in these draft files) and you're using an injury setting of, say, 100 or less, there's not much motivation to get a decent backup at safety. As far as young guys, that's just up to the GM. My perception for me is that using these files would stop me from doing what I do currently: set every single position at 100 playing time, especially if I have a young guy needing development time. In SP, it appears that AI teams use the same Rex that we'd get for depth charts, so the playing time isn't based on Endurance, but on the delta in talent between the starter and backup. So in that case, I don't see it making a ton of difference in development: the worthy guys will get their share of snaps, and the scrubs won't.

Quote:

Re: draft...you know what would be interesting, is if you could equalize in the other direction. By putting the mystery back in for everyone. Not sure if that's possible. But just randomly, with no tells whatsoever, nerf some guys who will come out looking great otherwise. Maybe take the obvious masks and turn them into 33% shots, where 66% of the time they will end up being mediocre or worse.
Not doable. FOF applies the mask and X-Factor after the import. If I send FOF a guy with flat 550 ratings, the X-Factor may change those to all the in the range of, say, 500 to 600, but he is who he is, and FOF is going to decide what masking, if any, is aplpied. In fact, the process that MalcPow used to create his opus on drafting and masking was basically this same one. He was just manually creating a draft files with all of the players rated the same, running them through the Draft File Generator, and then observing the results of how FOF created combines and masks for guys whose ratings he knew.

Quote:

Or/and on top of that, take the top 100 players in the draft talent pool and jack up the volatility for say, 80 of them. Make players think twice at the top of the draft because there very likely won't be any "safe" options, the superstud with the 10 volatility to go to in case you're nervous about taking a 97 guy at 1.4. Just some random thoughts.
I don't think people pay enough attention to volatility to make something like this meaningful. I know I don't. The odds of VSOD/VSOL are too remote for it to be a factor in drafting, even if the guy is a 100. I think it'd just create an annoyance. People tend to get angry when their draftees bust.

ozias 10-15-2011 08:49 AM

Ok, done with the files you've sent me Ben.

A few musings...

I usually ended up with 4-6 players with a 7.0+ adjusted rating in the draft. There was a season I had only 1 player, and I had another season that had 9 of them. The last few seasons there have been 2 in each of them. I like how that looks.

Generally I have around 10-12 players rated 80+, and that has stayed consistent all the way to 2040. I did have a peak of 21 players rated 80+ at one time. These are just from the draft files you've sent me.

QB's only two have been above 60/60, one was from early on, and the other has creeped his way up. He is a 69/69 now in his 10th year, but he was 59/59 in is 5th, and iirc he was 52/52 the year before that, and was hit with the VSOL in TC. I currently have 5 others in the 50-56 range, and the worst of the top 25 is a 44/44.

WR's only a few have ever been rated higher than 55, and the best I've seen was 57/57 and he had a BPR of 52. The BPR has been held in check I think as the highest I've seen was 68. There are currently 4 WR's rated above 50/50, and the worst top 25 is a 38/38.

Stats for the past 10 seasons in rushing ranged from 4.15-4.22 per carry with 110.0-113.2 yards per game.

Stats for passing 58.5-60.5 completion %, 9.67-9.89 yds/comp, 5.66-5.98 yds/att, 199.5-210.9 yards per game.

Scoring has been consistent as well, 16.6-18.2 ppg. The 16.6 was the 2040 season, and in the past 10 seasons 16.9 was the only other time scoring was below 17 ppg. The 18.2 was also the only time scoring was above 18 ppg in the past 10 seasons as well. If you drop those, the scoring was 17.2-17.8 ppg.

I have seen studs at all positions, except QB and WR. The QB's that have been rated over 7.0 in the draft haven't come out with a future grade higher than 80, and that number drops after a few seasons. I've only seen 1 WR in the draft rated near 6.0 and he ended up at 52/52.

Whatever changes you've made has definitely added more stud players than the previous test.

If you need any numbers just let me know.

Ben E Lou 10-15-2011 09:19 AM

Quote:

Originally Posted by ozias (Post 2549758)
Ok, done with the files you've sent me Ben.

A few musings...

I usually ended up with 4-6 players with a 7.0+ adjusted rating in the draft. There was a season I had only 1 player, and I had another season that had 9 of them. The last few seasons there have been 2 in each of them. I like how that looks.

Generally I have around 10-12 players rated 80+, and that has stayed consistent all the way to 2040. I did have a peak of 21 players rated 80+ at one time. These are just from the draft files you've sent me.

QB's only two have been above 60/60, one was from early on, and the other has creeped his way up. He is a 69/69 now in his 10th year, but he was 59/59 in is 5th, and iirc he was 52/52 the year before that, and was hit with the VSOL in TC. I currently have 5 others in the 50-56 range, and the worst of the top 25 is a 44/44.

WR's only a few have ever been rated higher than 55, and the best I've seen was 57/57 and he had a BPR of 52. The BPR has been held in check I think as the highest I've seen was 68. There are currently 4 WR's rated above 50/50, and the worst top 25 is a 38/38.

Stats for the past 10 seasons in rushing ranged from 4.15-4.22 per carry with 110.0-113.2 yards per game.

Stats for passing 58.5-60.5 completion %, 9.67-9.89 yds/comp, 5.66-5.98 yds/att, 199.5-210.9 yards per game.

Scoring has been consistent as well, 16.6-18.2 ppg. The 16.6 was the 2040 season, and in the past 10 seasons 16.9 was the only other time scoring was below 17 ppg. The 18.2 was also the only time scoring was above 18 ppg in the past 10 seasons as well. If you drop those, the scoring was 17.2-17.8 ppg.

I have seen studs at all positions, except QB and WR. The QB's that have been rated over 7.0 in the draft haven't come out with a future grade higher than 80, and that number drops after a few seasons. I've only seen 1 WR in the draft rated near 6.0 and he ended up at 52/52.

Whatever changes you've made has definitely added more stud players than the previous test.

If you need and numbers just let me know.

I also created a bunch of files for me. I'm not as far along as you, but my results are quite similar. I'm pretty confident that the code is now producing players with very similar talent distributions to FOF, with QB and WR nerfed. A couple of observations from my league that I'd like you to check in your league to see if they might be just dice rolls.
  • The RB position has seemed talent-heavy. I wonder if I'm creating too many of them, causing the odds to be too high of starter-quality ones being created in any given season. How many RBs do you have rated 50 and up, 55 and up, and 60 and up?
  • I think the endurance/BMI tie is too strong for the OL, and possibly for the DL as well. Take a look at the endurance for some of those positions and tell me what you see.
I don't see anything else that's of any real concern right now. Unless other issues are noticed, I'd like to do some minor cleanup stuff, implement a some player types as mentioned earlier, then maybe teach myself another programming language so I can release this as an Executable.

ozias 10-15-2011 09:49 AM

Quote:

Originally Posted by Ben E Lou (Post 2549765)
I also created a bunch of files for me. I'm not as far along as you, but my results are quite similar. I'm pretty confident that the code is now producing players with very similar talent distributions to FOF, with QB and WR nerfed. A couple of observations from my league that I'd like you to check in your league to see if they might be just dice rolls.
  • The RB position has seemed talent-heavy. I wonder if I'm creating too many of them, causing the odds to be too high of starter-quality ones being created in any given season. How many RBs do you have rated 50 and up, 55 and up, and 60 and up?
  • I think the endurance/BMI tie is too strong for the OL, and possibly for the DL as well. Take a look at the endurance for some of those positions and tell me what you see.
I don't see anything else that's of any real concern right now. Unless other issues are noticed, I'd like to do some minor cleanup stuff, implement a some player types as mentioned earlier, then maybe teach myself another programming language so I can release this as an Executable.


For the RB's currently in the league I've got

4 50+
7 55+
5 60+
2 65+
4 70+ <--none above 74
0 75+
1 80+ <--a rookie rated 87/87, which is the highest rated RB

I guess you could say there is a little to much talent there. Looking at individual stats, I have about 21 players a year gaining over 1,000 yards, in each of the last 10 seasons so that's on the high side.

A quick check of the top 25

C has 13 with no end.
G has 23
T has 23

DE has 15
DT has 20

I'll go through each team as well just to see how bad it is, and post when I complete the results.

ozias 10-15-2011 10:28 AM

Well for have the 16 teams I've done, the Endurance is bad.

OL 89 are 0-25, 50 are 26-50, 13 are 51-75, and 0 are 76-100
DL 67 are 0-25, 42 are 26-50, 21 are 51-75, and 2 are 76-100

I don't think I'll need to go thru the other 16 teams as you can see endurance is on the low side for the OL and DL.

Masked 10-15-2011 10:30 AM

In MP wouldn't people just draft RBs and switch their positions to WR. If so, can you make all RBs heavy so they can't switch effectively.

edit: Alternatively, it would be a simple house rule to say no position switches to WR

Dutch 10-15-2011 10:37 AM

Quote:

Originally Posted by Masked (Post 2549804)
In MP wouldn't people just [insert human ingenuity here] and dominate anyway?


Yes.

For instance, no ammount of QB and WR nerfing is going to help if Player X can just take any player in the league he wants because the rest of us are too stupid or too gullible to stop him from doing so.

Ben E Lou 10-15-2011 11:03 AM

Quote:

Originally Posted by Masked (Post 2549804)
edit: Alternatively, it would be a simple house rule to say no position switches to WR

Yup. Already thought of that, and that's the best solution. Yes, I could also easily make all RBs too big to switch, but then you might get into weight training silliness and all that, so just making it illegal is the cleanest solution to that issue.

FWIW, right now the players are rated in all categories, so I assume CB-->WR and S--->WR would work as well with the files that are currently being generated. That's on my list for cleanup--that players won't be rated on the opposite side of the ball. It's actually fairly tedious to do that, but definitely necessary.

Ben E Lou 10-15-2011 11:08 AM

Quote:

Originally Posted by ozias (Post 2549803)
Well for have the 16 teams I've done, the Endurance is bad.

OL 89 are 0-25, 50 are 26-50, 13 are 51-75, and 0 are 76-100
DL 67 are 0-25, 42 are 26-50, 21 are 51-75, and 2 are 76-100

I don't think I'll need to go thru the other 16 teams as you can see endurance is on the low side for the OL and DL.

Yeah, that's what I figured. That's a simple correction that I'll do in a bit. I'll be opening up the code in the next half hour or so. The list is...
  • Increase OL/DL Endurance
  • Investigate numbers of RBs generated in standard FOF draft classes and tweak if necessary
  • Make top-tier players' bars less flat.
I'm not going to tweak top-tier talent generation just yet. It may be that having more linemen with decent endurance will solve that "problem" on its own.

Ben E Lou 10-15-2011 12:02 PM

Heh. I thought I had already posted this, but it was still sitting in a window. It explains my earlier commetn about tweaking top-tier talent upward...

Quote:

Originally Posted by ozias (Post 2549758)
Generally I have around 10-12 players rated 80+, and that has stayed consistent all the way to 2040. I did have a peak of 21 players rated 80+ at one time. These are just from the draft files you've sent me.

Looking deeper, this is still a bit low, but not unreasonably so. I need to do a few more tweaks anyway. When I do, I'll raise the talent level very slightly. My league has 11 80+ guys, like yours. Here's what this looks like in MP and SP leagues with default draft classes:

CURRENT NON-QB/WR RATED 80+ CURRENT

WOOF: 20
CCFL: 28
IHOF: 24
PFL: 22
MATURE SP LEAGUE 1: 18
MATURE SP LEAGUE 2: 22
MATURE SP LEAGUE 3: 21
MATURE SP LEAGUE 4: 20
MATURE TEST MP LEAGUE: 27

So it seems fair to say that a league "should" average somewhere in the low 20s (average of those is 22) for 80+ current-rated players.

Ben E Lou 10-15-2011 12:47 PM

Quote:

Originally Posted by Ben E Lou (Post 2549824)
Yeah, that's what I figured. That's a simple correction that I'll do in a bit. I'll be opening up the code in the next half hour or so. The list is...
  • Increase OL/DL Endurance--Done.
  • Investigate numbers of RBs generated in standard FOF draft classes and tweak if necessary--Checked five FOF Draft Classes. Range for RBs was 59-63. In my code, the possible range of RBs generated is 57-63. And comparing ozias's league (23 50+ RBs) to a couple of existing MP leagues (CCFL and IHOF), I've realized that's probably pretty normal. CCFL has 20 50+ RBs, and IHOF has 22. RB Endurance is definitely significantly higher than in FOF, so I lowered that a bit. (Maybe so many quality RBs are generated because many teams should need two of 'em.)
  • Make top-tier players' bars less flat. Done.
I'm not going to tweak top-tier talent generation just yet. It may be that having more linemen with decent endurance will solve that "problem" on its own.


Fixes above implemented. I'm going to run one more round of tests, but overall I'd say these are ready to use for SP, and should be MP-ready fairly soon.


All times are GMT -5. The time now is 08:18 AM.

Powered by vBulletin Version 3.6.0
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.