Predictor Week 6

L

LegendaryGT

Dodd-Like
Joined
Oct 18, 2009
Messages
62,557
Predictor Top 25 said:
#1: ('ALA', [1.0, 125, 0, 5, 0])
#2: ('OSU', [0.996, 124, 1, 4, 0])
#3: ('MICH', [0.992, 123, 2, 5, 0])
#4: ('TA&M', [0.988, 122, 3, 4, 0])
#5: ('CLEM', [0.98, 120, 5, 4, 0])
#6: ('MIAMI', [0.96, 115, 10, 3, 0])
#7: ('TENN', [0.948, 112, 13, 5, 0])
#8: ('HOU', [0.948, 112, 13, 4, 0])
#9: ('WASH', [0.936, 109, 16, 4, 0])
#10: ('NEB', [0.916, 104, 21, 5, 0])
#11: ('MD', [0.912, 103, 22, 3, 0])
#12: ('WMU', [0.904, 101, 24, 4, 0])
#13: ('WVU', [0.884, 96, 29, 3, 0])
#14: ('BSU', [0.876, 94, 31, 4, 0])
#15: ('BAY', [0.868, 92, 33, 4, 0])
#16: ('FLA', [0.856, 114, 11, 4, 1])
#17: ('WIS', [0.852, 113, 12, 4, 1])
#18: ('STAN', [0.819, 111, 14, 3, 1])
#19: ('LOU', [0.811, 109, 16, 3, 1])
#20: ('COLO', [0.799, 106, 19, 3, 1])
#21: ('VT', [0.7973, 116, 9, 2, 1])
#22: ('AFA', [0.792, 73, 52, 3, 0])
#23: ('AUB', [0.76, 115, 10, 3, 2])
#24: ('WSU', [0.7493, 104, 21, 2, 1])
#25: ('LSU', [0.734, 121, 4, 2, 2])
...
#39: ('UNC', [0.659, 71, 54, 3, 1])
...
#49: ('UGA', [0.594, 86, 39, 2, 2])
...
#54: ('GT', [0.578, 82, 43, 2, 2])
...
#66: ('PITT', [0.506, 64, 61, 2, 2])
...
#69: ('GASO', [0.4853, 38, 87, 2, 1])
#70: ('UVA', [0.466, 54, 71, 2, 2])
...
#92: ('DUKE', [0.301, 44, 81, 1, 3])
Click to expand...

Last week we significantly exceeded the predictor's expectations against Miami's defense, and our defense held up better than it expected us to as well. Let the trend continue.

Predictor said:
GT vs PITT
Week 6 of 2016

GT - Offense
Yardage: 313 ypg
Performance: -2 (higher is better)
Scoring: 20
Efficiency: 15 (lower is better)
Performance: 11 (higher is better)
GT - Defense
Yardage: 341 ypg allowed
Performance: 9 (higher is better)
Scoring: 20
Efficiency: 16 (higher is better)
Performance: 19 (higher is better)

PITT - Offense
Yardage: 478 ypg
Performance: 1 (higher is better)
Scoring: 39
Efficiency: 12 (lower is better)
Performance: 10 (higher is better)
PITT - Defense
Yardage: 476 ypg allowed
Performance: -10 (higher is better)
Scoring: 37
Efficiency: 12 (higher is better)
Performance: -5 (higher is better)

Prediction:

Pure statistical: GT 34 (413 yards), PITT 26 (389 yards)
GT best case: GT 49 (512 yards), PITT 20 (314 yards)
PITT best case: GT 21 (336 yards), PITT 31 (455 yards)

GT talent: 5342
PITT talent: 6571
GT talent of schedule: 8149
PITT talent of schedule: 6933

Suggested talent adjustment: 0% towards PITT best case
Final: GT 34 (413 yards), PITT 26 (389 yards)
Click to expand...
 
TIL the Predictor does NOT like Bobby Petrino.
 
coit said:
TIL the Predictor does NOT like Bobby Petrino.
Click to expand...

When UNC beat FSU while being outgained and surrendering 600 yards, the predictor threw FSU in the garbage and erased every ounce of credit it had given Louisville for that win. It also didn't give UNC much credit at all, either.

It occurs to me that I forgot to run the rest of the season. Oops.
 
GT 42, GASO 14 (Win)
GT 27, DUKE 16 (Win)
GT 25, UNC 22 (Win)
GT 15, VT 36 (Loss)
GT 27, UVA 17 (Win)
GT 19, UGA 21 (Loss)

No change in the projected W/L. Still 8-4. Got a lot closer to a W in the UGA prediction though.
 
Good one on Bristol. Only other alternative is Fenway Park but there may be a baseball game that day. Thanks for the write-up on games
 
ThisIsAtlanta said:
GT 42, GASO 14 (Win)
GT 27, DUKE 16 (Win)
GT 25, UNC 22 (Win)
GT 15, VT 36 (Loss)
GT 27, UVA 17 (Win)
GT 19, UGA 21 (Loss)

No change in the projected W/L. Still 8-4. Got a lot closer to a W in the UGA prediction though.
Click to expand...
If this came to be I would be happy. Besides the uga loss. I honestly think we will beat them this year.
 
IMO, it's hard to see the combined score of our game with UNC as only 47. Feels like it'll be in the 60s.
 
Does your model factor in any sort of home field adjustment?
 
G776T said:
Does your model factor in any sort of home field adjustment?
Click to expand...

Nope. I'm not smart enough to quantify it in the structure of the model and with the limited data set. After the season is over, I could probably tell you what it worked out to, but because the predictor needs 5 or 6 games to do what it does, if you divide the data set in half you won't get enough to use until the end of the regular season. I could probably add it as a fudge factor, but everything in the model is self-tuning, and I also don't want to add parameters I have to tweak. I just live with it the way it is for now.
 
You must log in or register to reply here.
Back
Top