Absurd run schemes

[QuikSand]

Just out of curiosity - I wonder if any of the FOF players out there have tinkered with absurdly imbalanced running schemes.

I know people have tinkered extensively with various air attacks, and that's lots of fun to see the eye-popping numbers that can be reached by throwing all the time. I've also seen heavy run-oriented attacks, using multiple backs to build a powerful offense.

What about an offense built around, for instance, one offensive lineman? Centers are pretty easy to come by in this game -- what about a superstar centar, with topped-out skills? has anyone tried to build a running game where 100% of the runs are directed to go right in behind the center? (Or the right tackle, or whomever is the superstar player on the line)

Just curious - I've never tinkered with it, but it's the kind of thing that might be intriguing in a restricted career (like a "misfit toys" team where you luck out with one breakout lineman).

[The Shadow]

Great idea, QS.

Can you try it on the FOF2004 beta and let us know how it turned out? Thanks.

BTW, I really enjoyed your beta diary.

[QuikSand]

Originally posted by The Shadow
Can you try it on the FOF2004 beta and let us know how it turned out? Thanks.

That might be the most clever ploy to get more out of me on this topic that I've yet seen.

[cthomer5000]

Intruiging idea. My best guess is you'd end up with 1 injured offensive lineman, but run stats that probably wouldn't be all that different from normal.

Of course, the lineman would be a pro-bowler every season and a lock for the hall of fame with those KRB stats.

[Buzzbee]

It would probably be successful, but I'm guessing that the defense eventually would "look very familiar with that play" and the success would see diminishing returns.

[QuikSand]

Originally posted by Buzzbee
It would probably be successful, but I'm guessing that the defense eventually would "look very familiar with that play" and the success would see diminishing returns.

One would hope so... but I'm not so sure.

[Bonegavel]

The game should smack that down and actually reduce the stats. If you know they run up the middle every run, that easy to counter. Even for a crappy team.

[sabotai]

I'll do a little experiment for it. I'll check back with results.

[albionmoonlight]

But I remember Jim saying something about the AI not being able to handle "absurd" user values. For instance, if you throw the ball 100% of the time on second and short--you will get better numbers than you would IRL because the game engine has limits to how much it can react to extreme values.

Am I remembering that right?

If so, I wonder how that will affect multiplayer, where you have the ability to go in and set your D up to counter extreme values.

[Buzzbee]

Originally posted by QuikSand
One would hope so... but I'm not so sure.

I'm not saying that you would end up with negative yardage everytime, but that long term the YPC would decline. It would eventually reach some limit, and that limit would be less than if you had equal players on the OL and spread the ball around.

In other words, what might be 3 YPC in a "normal" scheme might be reduced to 2.5 YPC in the "one slot" scheme.

Just my theory though.

[Samdari]

I clicked on this thread thinking it would be about things like the "help me move my money from Namibia" type schemes.

What it this discussion about a football sim doing in this forum?

[TroyF]

Originally posted by QuikSand
One would hope so... but I'm not so sure.

I'm not either. I'm betting you could get some massive rushing performences out of it.

TroyF

[sabotai]

How I did it.

Complete draft, I drafted Corey Dillon and a stud C. The rest of the line was filled with average linement.

I set everything to the AI. I then changed all runs to be 50% Inside Left Guard and 50% inside right guard.

I ran 10 seasons with the AI set up to change depth charts and to sign players (and set to NO on AI controls game plan)

Results:

Season 1: Record 7-9, PF 269, PA 340 (2nd in Div)
Runs: Left 0-0, Middle 409-1165, Right 0-0. (Total: 2.85 Ave)

Season 2: Record 10-6, PF 301, PA 264 (Won Division. Lost WC)
Runs: Left 0-0, Middle 439-1154, Right 0-0 (Total: 2.63 Ave)

Season 3: Record 8-8, PF 321, PA 328 (Won Divison, Lost WC)
Runs: Left 0-0, Middle 400-1084, Right 0-0 (Total: 2.71 Ave)

Season 4: Record: 12-4, PF 377, PA 258 (Won Division, Lost CC)
Runs: Left: 0-0, Middle 466-1098, Right 0-0 (Total: 2.35 Ave)

Season 5: Record: 9-6-1, PF 312, PA 315 (Won Division, Lost WC)
Runs: Left: 0-0, Middle 424-1022, Right 0-0 (Total: 2.41 Ave)

Season 6: Record: 8-8, PF 281, PA 269 (3rd in Div)
Runs: Left 0-0, Middle 428-1105, Right 0-0 (Total: 2.58 Ave)

Season 7: Record: 6-10, PF 253, PA 273 (4th in Div)
Runs: Left 0-0, Middle 393-891, Right 0-0 (Total: 2.26 Ave)

Season 8: Record: 9-7, PF 294, PA 246 (Won Division, Lost WC)
Runs: Left 0-0, Middle 418-1133, Right 0-0 (Total: 2.71 Ave)

Season 9: Record: 11-5, PF 325, PA 299 (Won Division, Lost DivGame)
Runs: Left 0-0, Middle 396-1079, Right 0-0 (Total: 2.72 Ave)

Season 10: Record: 10-6, PF 287, PA 255 (Won Division, Lost WC)
Runs: Left 0-0, Middle 421-1167, Right 0-0 (Total: 2.77 Ave)


Totals: 90-69-1, PF 3020, PA 2847
7 Playoffs, 5 Lost in Wildcard, 1 Lost in Divisional Round and 1 Lost in the Conf. Championship

Middle: 4194-10,898, 2.6 Ave Per Run

(I hope everyone can figure out the per season averages for themselves. )

Next, I'll do the "Control"

[Ben E Lou]

Looks like the defense looked very familiar with that play...

[sabotai]

Yeah, no season with over a 3.0 yards per gain overall. (2.85 being the highest)

Still...made the playoffs 7 out of 10 times. Weird.

[albionmoonlight]

When you are doing controls, perhaps you should try it with a tackle to make sure that the 2.~ avg. was not more a function of where you ran than how much you ran there.

[sabotai]

Yeah, I'll try it with a different location another time. Right now I'm just hitting the recommend button so it spreads it around. (Note: I am using the same team for all 20 seasons)

[BishopMVP]

Are you running it 100% of the time in this? Because otherwise you must have great defense and special teams.

[sabotai]

No, I'm running everything as the AI set it except the direction of the runs, so I could get a "real world" result. Not one that depended on a player running 100%.

[sabotai]

The same rules as above, except I'm setting everything in the offensive game plan to recommend. (But still NO to AI control game plan, to keep the run %'s the same)

Season 1: Record 9-7, PF 347, PA 266 (2nd in Div)
Runs: Left 141-552, Middle 252-823, Right 53-240
Total: 446-1615 (3.62 Ave)

Season 2: Record 8-8, PF 274, PA 263 (3rd in Div)
Runs: Left 133-599, Middle 239-681, Right 4-157
Total: 412-1437 (3.49 Ave)

Season 3: Record 9-7, PF 333, PA 288 (2nd WC, Lost in WC)
Runs: Left 155-664, Middle 253-773, Right 50-213
Total: 458-1650 (3.60 Ave)

Season 4: Record 12-4, PF 363, PA 270 (1st, Lost in SB)
Runs: Left 156-656, Middle 240-759, Right 33-71
Total: 429-1486 (3.46 Ave)

Season 5: Record: 11-5, PF 381, PA 317 (1st, Lost in CC)
Runs: Left 143-587, Middle 265-956, Right 44-134
Total: 452-1677 (3.71 Ave)

Season 6: Record: 9-7, PF 345, PA 342 (2nd WC, Lost in WC)
Runs: Left 145-634, Middle 259-877, Right 57-217
Total: 461-1728 (3.75 Ave)

Season 7: Record: 11-5, PF 355, PA 295 (1st, Lost in WC)
Runs: Left 138-601, Middle 245-887, Right 48-196
Total: 431-1684 (3.91 Ave)

Season 8: Record 9-7, PF 356, PA 304 (1st, Lost in WC)
Runs: Left 129-425, Middle 255-845, Right 43-172
Total: 427-1442 (3.38 Ave)

Season 9: Record: 9-7, PF 368, PA 269 (2nd in Div)
Runs: Left 123-482, Middle 250-843, Right 42-141
Total: 415-1466 (3.53 Ave)

Season 10: Record: 10-6, PF 346, PA 271 (1st in Div, Lost in DivG)
Runs: Left 132-451, Middle 240-790, Right 61-245
Total: 433-1486 (3.43 Ave)


Totals: 97-63, PF 3468, PA 2885
7 Playoffs, 4 WC loses, 1 DivG lose, 1 CC lose and 1 SB lose

More wins, little better playoff showing.

[sabotai]

Control runs:

Left: 1395-5651 (4.05 Ave)
Middle: 2498-8234 (3.30 Ave)
Right: 471-1786 (3.79 Ave)

Total: 4364-15,671 (3.59 Ave)

[BishopMVP]

Originally posted by sabotai
No, I'm running everything as the AI set it except the direction of the runs, so I could get a "real world" result. Not one that depended on a player running 100%.

This makes sense for the tests you're running, but if you're trying to compare it to the All-Air Attack and see if the computer can respond, you should set the percentages to 99% running. If a team can still make the playoffs at least 1/2 the time under those circumstances I'd be surprised. If not, then it would seem to prove that passing is the way to go in FOF4.

[sabotai]

170 More runs in the control group. 17 more runs per season. According to the percentages, they should be the same, but this also says that my total time of possesion was probably the same. If I held the ball shorter or longer in the control, I would probably see a bigger difference.

4,773 More yards in the control group, 477.3 yards more per season. 0.99 more yards per run in the control group. So, it would seem, as SkyDog implied above, that the defense does in fact recognize when you are running it straight up the middle all of the time.

[sabotai]

Originally posted by BishopMVP
This makes sense for the tests you're running, but if you're trying to compare it to the All-Air Attack and see if the computer can respond, you should set the percentages to 99% running. If a team can still make the playoffs at least 1/2 the time under those circumstances I'd be surprised. If not, then it would seem to prove that passing is the way to go in FOF4.

Yeah, I would if I wanted to compare it to other extreme game plans. But I figured I would check to see if just using extreme directional running itself could go under "AI exploit" game plan.

But, since running up the middle yeilded almost an entire yard less per run, running it in the same direction all the time seems detrimental.

But I will try a new location to see if there is a difference there.

[sabotai]

After running a few seasons to just the left tackle, I'm noticing season averages of less than 3 yards per run again. So I'm going to stop here. I don't see th epoint of continueing.

[The Shadow]

Sabotai,

Great job. Are you running the same year over and over again for 10 years or is it 10 consecutive years?

BTW, I'm guessing that running it behind center all of the time has less chance for a long run, which lowered the average. If you have a stud OT and ran it behind him (inside tackle and outside tackle) all of the time, your YPC should improve.

Thanks for your hard work.

[sabotai]

Same season over and over.

I'll probably try it again later by using a stud T and doing it that way. You also have less of a chance of being stopped behind the line for negative yardage when running behind the center, so in that way it also increases the average and kind of cancels the less chance of a long run out.

[mckerney]

Here's results from running a normal offense and only running in two directions (50 - 50 off each tackle):

Year One : 13 - 3 Superbowl Victory
Team - 1906 Yards 4.2 YPC 15 TDs
Starter - 1388 Yards 4.7 YPC 12 TDs

Year Two: 6 - 10
Team - 1328 3.2 YPC 14 TDs
Starter - 824 Yards 3.4 YPC 9 TDs

Year Three - 14-2 Loss in Divisional Playoffs
Team - 2195 4.5 YPC 17 TDs
Starter - 1608 Yards, 5.2 YPC 11 TDs

Year Four: 8 - 8
Team - 1803 Yards 4.0 YPC 12 TD
Starter - 1131 Yards 4.6 YPC 9 TD


Not quite sure what is to be learned from this.

[EagleFan]

Originally posted by Samdari
I clicked on this thread thinking it would be about things like the "help me move my money from Namibia" type schemes.

What it this discussion about a football sim doing in this forum?

With the number of those emails that I have received over the past month, I think I'm moving to that country. It seems like everyone in that country has come across an insane amount of money in one way or another.

[The Shadow]

Originally posted by TroyF
I'm not either. I'm betting you could get some massive rushing performences out of it.

So do we conclude that FOF AI is smart enough to "recognize a familiar play" if we run behind the same OL on every running play?

That's good news, isn't it?

But mckerney's results seem to indicate that as long as you run behind at least 2 different OL, the AI will not recognize that.

[HornedFrog Purple]

I think it indicates more that the AI is guessing and coming up wrong. That is a good thing. Most teams have a side they prefer to run to, plus in your mind you could assume the RB cutback against the grain sometimes.

One thing I would like to see in future editions especially with TCY and the option is running to the short or long side of the field with percentages put to each.

[sabotai]

Running pretty much exclusively to both sides of the line (50% one way, 50% the other) isn't unheard of. Besides, going 50% to each side gives you a good deal of play selection. Running it up the gut or just to one side does not.

I'm guessing that running it behind the C all the time or the LT all the time doesn't give the AI enough choices in play selection and therefore it has to call the same plays over and over and over, and thus the play recognition.

Another good test would be to do it with one QB with few formation choices and one with a lot to see if upping the # of formations will generate better running due to the increase play selection.

So it would see, at least for now, that the AI is not scouting and knowing that the team will run it in a certain direction. It just recognizes a play ran several times in a game already. Not the ideal solution, but it accomplishes the same thing...kind of.

[mckerney]

Originally posted by sabotai
Running pretty much exclusively to both sides of the line (50% one way, 50% the other) isn't unheard of. Besides, going 50% to each side gives you a good deal of play selection. Running it up the gut or just to one side does not.

I'm guessing that running it behind the C all the time or the LT all the time doesn't give the AI enough choices in play selection and therefore it has to call the same plays over and over and over, and thus the play recognition.

Another good test would be to do it with one QB with few formation choices and one with a lot to see if upping the # of formations will generate better running due to the increase play selection.

So it would see, at least for now, that the AI is not scouting and knowing that the team will run it in a certain direction. It just recognizes a play ran several times in a game already. Not the ideal solution, but it accomplishes the same thing...kind of.

I was trying the only running behind two guys to see if it would show anything, but it doesn't really seem to. Next I'm going to try only running to one side.