Twitch coaches football

Diseqc

Diseqc

Dodd-Like
Joined
Apr 15, 2008
Messages
49,593
Given that we’re taking crowd input into play calling nowadays. I’m working on a twitch plays football channel. It’ll collect chat votes on what play to run and then push it to an intern on the sideline to relay into the team.

If twitch could beat Pokémon this way,surely we can beat an FCS team.
 
GTCrew4b said:
Guaranteeing a positive outcome at that point in the game was the correct move. The time not be “weak” was on the 3 previous plays.

not to mention it cost us an additional timeout.
Click to expand...
2 yds vs NIU when you need it has to be as much of a “guaranteed positive outcome” as a FG at the program level. We have to execute to that philosophy for it to be true or even perceived as true. Speaking it into existence, if you will.

I’m sure there are metrics that assigned a probability to us converting there. It would be interesting to know what they said but I’d bet that it skews towards a TD, extrapolating from us being an 18pt favorite etc. Play selection and execution aside and thinking strictly of the go/no-go decision, I am ok with the risk that was taken when understanding that one can’t see the future. I realize not everyone will see it that way and that’s ok too.
 
Diseqc said:
Given that we’re taking crowd input into play calling nowadays. I’m working on a twitch plays football channel. It’ll collect chat votes on what play to run and then push it to an intern on the sideline to relay into the team.

If twitch could beat Pokémon this way,surely we can beat an FCS team.
Click to expand...
I know you’re joking, but I’m surprised there’s no AI software to recommend plays
 
gtphd said:
I know you’re joking, but I’m surprised there’s no AI software to recommend plays
Click to expand...

With today's technology, it really shouldn't be so hard. I bet the DeepMind team would love the challenge of building AlphaCoach at Tech. We should get them on the line.
 
MeatWrench said:
Settling for 3 there was weak and I’m glad he went for it.
Click to expand...

Both 3rd and 4th down play calls were weak. Kicking was the right call at the time. There was too much time left to leave points on the table.
 
ElCidBUZZingFAN said:
It lost us the game, you tard.
Click to expand...
We had already missed 2 field goals at that point. If we couldn’t stick it in there, then we didn’t deserve to win the game anyway. Going for it was the right call. The playcall and execution was bad. The indecisiveness was atrocious though and cost us a much needed timeout.
 
I can’t wait until our third down play is just Collins yelling the n word on the sideline for 40 seconds.

which would be an improvement.
 
Diseqc said:
Given that we’re taking crowd input into play calling nowadays. I’m working on a twitch plays football channel. It’ll collect chat votes on what play to run and then push it to an intern on the sideline to relay into the team.

If twitch could beat Pokémon this way,surely we can beat an FCS team.
Click to expand...

I've been saying since Collins got here there's no reason we can't calibrate an ML algorithm to recommend plays most likely to succeed based on down, distance, player height/wt, position of the sun, crowd noise level, etc etc. It need not call the plays, but it could rank them and present them on a handheld to the O.C. who makes a selection.

It's basically letting Vegas and stone-cold reality slap a coach for having messed up priorities.
 
Flywheel said:
I've been saying since Collins got here there's no reason we can't calibrate an ML algorithm to recommend plays most likely to succeed based on down, distance, player height/wt, position of the sun, crowd noise level, etc etc. It need not call the plays, but it could rank them and present them on a handheld to the O.C. who makes a selection.
Click to expand...
Above the Line Algorithm imo
 
You must log in or register to reply here.
Back
Top