2016 Preseason Probability Poll Results

coit

coit

Persecuted for his beliefs
Joined
Nov 29, 2007
Messages
88,474
I'm going to start throwing some results in here and a spreadsheet showing how everyone voted.

BC 0.667 W
Mercer: 0.874 W
Vandy: 0.706 W
Clemson: 0.390 L
Miami: 0.462 L
Pitt: 0.585 L
Ga Sou: 0.738 W
Duke: 0.666 W
UNC: 0.537 L
VT: 0.543 W
UVA: 0.689 W
UGA: 0.472 W

Total: 7.33 wins

Adjusted total thru Game 1: 7.66 wins (+0.333)
Adjusted total thru Game 2: 7.79 wins (+0.126)
Adjusted total thru Game 3: 8.08 wins (+0.294)
Adjusted total thru Game 4: 7.69 wins (-0.390)
Adjusted total thru Game 5: 7.23 wins (-0.462)
Adjusted total thru Game 6: 6.65 wins (-0.585)
Adjusted total thru Game 7: 6.91 wins (+0.262)
Adjusted total thru Game 8: 7.24 wins (+0.333)
Adjusted total thru Game 9: 6.70 wins (-0.537)
Adjusted total thru Game 10: 7.16 wins (+0.457)
Adjusted total thru Game 11: 7.47 wins (+0.311)
Adjusted total thru Game 12: 8.00 wins (+0.528)

Changes in probability from 2015 repeat opponents:

Clemson: -0.157
Miami: -0.177
Pitt: -0.159
Duke: -0.155
UNC: -0.211
VT: -0.116
UVA: -0.116
UGA: -0.154

Ranking of opponents by probability:

1. Clemson (L)
2. Miami (L)
3. UGA (W)
4. UNC (L)
5. VT (W)
6. Pitt (L)
7. Duke (W)
8. BC (W)
9. UVA (W)
10. Vandy (W)
11. Ga Sou (W)
12. Mercer (W)

Here's the complete voting spreadsheet.

 
Last edited:
eneq said:
You have way too much time on your hands
Click to expand...

If you know what you are doing it doesn't take that long. I've probably spent 3 hours tops on this, 10 or 15 minutes at a time.
 
ThisIsAtlanta said:
The 0.874 probability of beating Mercer is actually the dumbest thing on this entire website.
Click to expand...

Well, keep in mind that the maximum possible value is 0.9. If you take the top value as 0.95 the result is 0.916. If you take the top vote as 1.0, the result is 0.96.

If you delete @eneq, you get an overall increase in intelligence and poll result.

If anything, the poll is helpful in revealing trolls. bennywy is one to watch for.
 
coit said:
Well, keep in mind that the maximum possible value is 0.9. If you take the top value as 0.95 the result is 0.916. If you take the top vote as 1.0, the result is 0.96.

If you delete @eneq, you get an overall increase in intelligence and poll result.

If anything, the poll is helpful in revealing trolls. bennywy is one to watch for.
Click to expand...

There's got to be a better way to do this. I don't want to live in a world where 87% could be 96%.
 
Oh öööö, I voted 0.2? I don't remember that.
 
As suggested by @beej67, the total EV will be adjusted up and down each week based on the probability poll results.

So, for Week 1 a victory increases the 0.667 poll result by 0.333. The total expected win value is now 7.66.
 
Our Week 2 victory has increased the 0.874 poll result by 0.126. The total expected win value is now 7.79.
 
Our Week 3 victory has increased the 0.706 poll result by 0.294. The total expected win value is now 8.08.
 
This is pretty neat. I should have answered more polls...

(no sticky though, we got enough already)
 
Anybody who has praised the wisdom of the crowds hasn't seen too many actual crowds.
 
40% seems generous. ESPN has us at 20%. Thats still pretty good against a presumed playoff team.
 
You must log in or register to reply here.
Back
Top