Video: Drafting w/ MalcPow and Quik

[Squirrel]

Big thanks to MalcPow for doing this.

MalcPow and Quik on drafting in FOF 8 - YouTube

Since Quik and I started making videos about FOF by far the most requested one was for MalcPow to join us and talk about how he does it on drafting...anyone who is been in a multiplayer league with him knows what I mean.

So much fun. Hope you enjoy it.

Some links to things referred to in the video:

Draft Analyzer - Front Office Football Central

Comparing Two RB's - IHOF Forums

Some Combine Benchmarks (and Drafting Thoughts) - Front Office Football Central

[Squirrel]

Another link to add:

The Autumn Wind: A GML Scratchpad - Front Office Football Central

[QuikSand]

Well, I enjoyed it.

[Ternvig]

So did I!

Watched the whole thing on my Zwift-bike as I rode the Virtual Mont Ventoux, so might have mist a detail or two.

It is interesting how some GMs just keep draft gold - even after the first 3 rounds. We have a few GMs in the CFL who releases pure value 5th year studs every year.

[corbes]

Re: Draft File Generator

Is there a way to create a player who is an overall 9 or an overall 625? It seemed during the video that there was an oblique reference to this. I can see how to create a player who has a 625 with respect to each attribute, but I can't see how to create a player who is an overall 625, with each attribute being what they may? Am I missing something or misunderstanding something?

Enjoyed the video, many thanks for this.

[MalcPow]

Roster file generator is what you need to use to only set a talent number. I know, it’s all pretty strange. Draft file generator I think you can only set individual attributes and it’s 375-625.

[corbes]

All this time and I don't believe I've ever clicked on the roster file generator. Interesting. Thanks!

[finkellll]

So I was curious about the new player file stuff, because I've never played around with it but I knew what it was doing (roughly).

For those who don't know what they were talking about you can go to SteamApps/common/fof8 and then double click the playerfile.exe. there are help files in the executable, and here is the relevant part:


9 - Rare Superstar - 0.3%
8 - Superstar - 0.4%
7 - Near Superstar - 0.6%
6 - Elite Starter - 1.1%
5 - Excellent Starter - 1.5%
4 - Very Good-Plus Starter - 2.6%
3 - Very Good Starter - 3.9%
2 - Good Starter - 12.2%
1 - Replacement-Level Starter (decent backup) - 23.9%
0 - Roster Filler - 53.4%

I never considered an "all 9s" player file so I went ahead and made an "all 0s" "all 1s" etc. I exported immediately after game start, so there is no pre2 randomness at work here. I also turned off X-factor to minimize randomness.


While the details didn't reveal anything surprising (spoiler alert, higher # = higher number of talented players), I was surprised at the range of bar values. The only range that had 0 players in it was 70+ for 0s, 80+ for 1s, and 90+ for 2s, 3-9 had at least 1 player in every bar range.



I did see that undrafted rookies cluttered up the 0-39 current ovr. range in all groups, which is why I dropped that range in the second chart. The all 9s player file had 1240 players rated 0-39, I think there were like 5 guys that weren't rookies in this range. I didn't want to drop rookies from the data because obviously there were a lot of talented rookies spread throughout in every case, so it would do more harm than good.



I was expecting a 9 to have a bar range of 75-90 current overall, not 40-100. I'm trying to see if stuff like height, weight, and "personality view" changes, but I'm not seeing it (probably because I'm using the initial player file, just modified current overall values).



I wonder if a 70 ovr. "9" player more consistently produces stats to match his bars compared to a 70 ovr. "3", but I don't know where to even begin testing that without an unreasonable amount of brute force.




























edit: I added the 80-99 graph for completion, but it should be obvious to everyone that as the 0-9 climbs there are more players 80-99 overall.

[Squirrel]

These are fascinating data...thanks for posting this

[MalcPow]

It’s been years since I’ve looked at this specifically, but make sure you’re doing something to call out the 9s (or whatever value) in your roster file. I used last names. There will be a good number of players generated to fill out the universe that aren’t being built off of a 9 overall talent level and you’ll want to easily see that.

[finkellll]

table_flip.gif


I'll take another (brief) look. I was wondering if that was the case because of the swarm of free agent rookies that were 12/29 overall.


edit: after taking another look the only filler players were undrafted rookies. Maybe 10 total on rosters, the rest are free agents. All rookies. My charts for 40-59 & 60-79 are unaffected, the lower rated counts are.


I did a new output based on all 9s and got the following output. It's really interesting how close the number of 40+ players are since I had to create a new game to get the new output. I expect similar results with the other outputs, the number of players with a current rating of 0-39 will drop by about 1200 total and the 40+ group will be unaffected.

range new value old value 0-9 12 115 10-19 15 559 20-29 8 437 30-39 31 129 40-49 137 136 50-59 322 329 60-69 477 455 70-79 545 573 80-89 382 409 90-99 94 83 100 1 0

[QuikSand]

Okay, nobody else brought this up, so I will. Haven't listened to the vid/chat yet, this is just from my recollection...

At one point, MP said something like "I just put in the minimum weight for scouting and interviewing for the staff draft." (referencing the tool used in the leagues run by BELCo)

Now, the easy interpretation for this, the one I was anticipating, was more or less (my paraphrasing here): "I'm strong enough with the combines and other things that I find the scout's view to be less useful than you little people do, so I might as well stock up on staff who are good at other stuff like playcalling and whatnot." And I would have completely understood and appreciated that.

But...what I think I heard from MP instead, was closer to (again my paraphrasing): "With terrible-rated scouting from my staff, I actually gain some scouting insight into the players, so that works our for me."



So my takeaway question is... is there a way to Bizarro-scout players in FOF and make that work for you?

Simplest way this would work seems to be a mirror-image theory. I have seen savvy players like Ben sometimes say "well, my good scouts saw this guy with bigger bars than that, so that's how I sensed this guy was worth that higher pick than you'd have been willing to make." Could the reverse be true? MP saying to himself "my awful scouts see this guy as only a 30-something rated guy... but they're terrible and every other signal tells me this guy is good, so I'm interpreting that as a really good sign" or something of the sort?

Is it possible that rubes like me, who really focus on good scouting, are missing the actual moneyball play, and that the real value lies with staff who are either excellent or terrible in scouting, just not in between?

Please discuss.

[finkellll]

I find it hard to believe that he has truly terrible scouting/interviewing, because a 0 interviewing bar in the oc/dc means that there is a 48 point spread in the pre interview widths, which would make any attempt at analysis pretty useless since you only have combines and raw grade at that point.

I'm guessing that he is OK with a 25 or 50/100 interview & scouting bar, where most GMs would stay far away at that point. That'll give him blue bar widths in the 30s, and he can probably see more of the stair stepping he was talking about with the GD bar in his ihof picks.

With a 48 point spread, the overall layout of bars will be too wide for any accurate analysis to happen.

[tzach]

Originally Posted by QuikSand

So my takeaway question is... is there a way to Bizarro-scout players in FOF and make that work for you?

the short answer is yes -- this is what i do in gml. i deliberately go after the worse scouts i can (focusing on other coaching stuff) and use the reverse approach you mentioned above. i like to call that approach the 'malcpow draft predictor'


for that approach interviews or bar widths are nearly useless to me

[MIJB#19]

Originally Posted by QuikSand

Now, the easy interpretation for this, the one I was anticipating, was more or less (my paraphrasing here): "I'm strong enough with the combines and other things that I find the scout's view to be less useful than you little people do, so I might as well stock up on staff who are good at other stuff like playcalling and whatnot." And I would have completely understood and appreciated that.

But...what I think I heard from MP instead, was closer to (again my paraphrasing): "With terrible-rated scouting from my staff, I actually gain some scouting insight into the players, so that works our for me."

The way I heard it, was closer to the first one, fwiw. And then, not quite in that context, but more like: it takes too much time for too little gained to take a look at the bars and decipher whether my staff/scout is good or bad at scouting, the numbers (combine scores, bar profiles) I have are enough to not really care about how reliable the staff/scout is. As a result not doing interviews either.

There's one person out there (MP himself) who can clear it all up though.

[Ushikawa]

I am pretty dumb but never found the draft file generator, I am in:
C:\Users\Administrator\AppData\Local\Solecismic Software

Could it be hidden? If so, is it findable via search? Or do I need to go into Terminal (been years haha)?

[garion333]

If you have crap scouting and interview then I expect one would be focusing on combines and static bars (and masked pairs?). I suppose that makes sense.

Personally, I prefer high scouting because it gives me a better look at the actual player. I'm not generally looking for boomers, but avoiding busts.

[QuikSand]

Originally Posted by garion333

If you have crap scouting and interview then I expect one would be focusing on combines and static bars (and masked pairs?). I suppose that makes sense.

Personally, I prefer high scouting because it gives me a better look at the actual player. I'm not generally looking for boomers, but avoiding busts.

right, this is conventional wisdom

i'm just plumbing whether there's a contrarian view... not that you ignore your terrible scout, but you use his terribleness as an asset... kinda like a Constanza Rule for scouting, you start with what he thinks and do the opposite

if MP and tzach are on board with this, i think it's worth considering

[finkellll]

Originally Posted by Ushikawa

I am pretty dumb but never found the draft file generator, I am in:
C:\Users\Administrator\AppData\Local\Solecismic Software

Could it be hidden? If so, is it findable via search? Or do I need to go into Terminal (been years haha)?

Steam/SteamApps/common/front_office_football