Jump to content
Sign in to follow this  
  • entries
    78
  • comments
    48
  • views
    43,416

DUSHEE beat Vegas! (barely)

Duquesne Frog

513 views

A few have asked how DUSHEE did picking the bowl games this year and the answer is a mixed bag. Historically, DUSHEE has done very well picking bowl winners while the regular season was more sketchy. (Since most of the documentation of prior DUSHEE success was cataloged on kf.c, you'll have to take my word for it or see if you can find it over there.) This year, DUSHEE wasn't really much better in the bowls than it was in the regular season, which is to say, not that great.

Using the same ESPN 'confidence points' technique used by our esteemed panel of college football experts, DUSHEE would have garnered 422 points, which would have placed him sixth out of nine. DUSHEE's top 10 confidence picks, based on the largest differentials in the teams' DUSHEE scores were:

Bowl | Winner Score1 DUSH1 | Loser Score2 DUSH2 | ActualPM DUSHPredPM DUSHConfidence

Boca Raton | Marshall 52 21.66 | Northern Illinois 23 -1.75 | 29 23.41 38

Heart of Dallas | Louisiana Tech 35 12.97 | Illinois 18 -6.47 | 17 19.44 37

Foster Farms | Stanford 45 10.77 | Maryland 21 -4.17 | 24 14.94 36

Texas | Arkansas 31 15 | Texas 7 2.64 | 24 12.36 35

Rose | (3) Oregon 59 23.33 | (2) Florida State 20 12.06 | 39 11.27 34

Independence | South Carolina 24 2.42 | Miami (FL) 21 12.27 | 3 -9.85 33

Belk | (13) Georgia 37 19.1 | (20) Louisville 14 10.67 | 23 8.43 32

Fiesta | (21) Boise State 38 15.34 | (12) Arizona 30 7.43 | 8 7.91 31

Cactus | Oklahoma State 30 -5.84 | Washington 22 2.05 | 8 -7.89 30

Las Vegas | (23) Utah 45 0.26 | Colorado State 10 8.02 | 35 -7.76 29

The games in red, DUSHEE missed: the Independence where an expected 10 point Miami win (DUSHPredPM stands for DUSHEE Predicted Point Margin) was a 3 pt win for South Carolina, the Cactus where an expected 8 point win for Washington was an 8 point win for Oklahoma State, and the Las Vegas bowl where an expected 8 point win for Colorado State was a 35 point win for Utah.

Not surprisingly, the Las Vegas Bowl was the game DUSHEE missed by the most, reality deviating from expectation by almost 43 points. The 5 games that deviated from expectation the most were:

Bowl | Winner Score1 DUSH1 | Loser Score2 DUSH2 | DUSHDiff

Las Vegas | (23) Utah 45 0.26 | Colorado State 10 8.02 | 42.76

Russell Athletic | (18) Clemson 40 10.5 | Oklahoma 6 16.26 | 39.76

Peach | (6) Texas Christian 42 24.67 | (9) Mississippi 3 21.94 | 36.27

Rose | (3) Oregon 59 23.33 | (2) Florida State 20 12.06 | 27.73

GoDaddy | Toledo 63 0.26 | Arkansas State 44 3.11 | 21.85

It is interesting to note that DUSHEE predicted the correct winner in 2 of the 5 games (Peach and Rose) but predicted much closer games than the blowouts that occurred.

In total, DUSHEE picked 21 out of 39 bowls winners correctly, barely getting half right. Mediocre, by definition. Although DUSHEE did predict the correct winner in 6 of the 7 "New Years Day 6" plus MNC games, missing only Georgia Tech's upset of Mississippi State in the Orange.

However, DUSHEE was less mediocre than Vegas, meaning if I'd had the courage to actually put some money on DUSHEE, I would have done decently.

Taking the spreads from TeamRankings.com (not being a gambler, I don't know if that is a reputable source for spreads or not) and setting those spreads as Vegas' expected point margin, Sin City's confidence picks look slightly worse than DUSHEE's, earning 418.5 points (compared to DUSHEE's 422), the half confidence point coming from the fact that Vegas only gives lines with a precision of 0.5 points so that there are often multiple games with the same point spreads. For instance, there were 6 bowls tied for 8th most confident with a point spread of 7. Thus, instead of randomly assigning those games confidence points ranging from 32 to 27 (corresponding with the 8th through 13th most confident picks) all the games were assigned an average confidence of 28.5.

Vegas missed on 4 of their most confident 13 picks, colored in red:

Bowl | Winner Score1 Loser Score2 | VegasPredPM

Foster Farms | Stanford 45 Maryland 21 | 14

New Mexico| Utah State 21 Texas-El Paso 6 | 9.5

Boca Raton | Marshall 52 Northern Illinois 23 | 8.5

Music City | Notre Dame 31 (22) | Louisiana State 28 | -8

Rose | (3) Oregon 59 (2) Florida State 20 | 7.5

Sugar | (5) Ohio State 42 (1) Alabama 35 | -7.5

Sun | (15) Arizona State 36 Duke 31 | 7.5

Cactus | Oklahoma State 30 Washington 22 | -7

Belk | (13) Georgia 37 (20) Louisville 14 | 7

Texas | Arkansas 31 Texas 7 | 7

Birmingham | Florida 28 East Carolina 20 | 7

Holiday | (24) Southern California 45 (25) Nebraska 42 | 7

Outback | (17) Wisconsin 34 (19) Auburn 31 | -7

DUSHEE and Vegas agreed on 6 of the 10 Bowls selected for higest confidence (Foster Farms, Boca Raton, Rose, Cactus, Belk, and Texas) and both missed on one of those 6, Oklahoma State's upset of Washington in the Cactus. Both DUSHEE and Vegas got the other 5 correct and the 5 favorites in those games won by an average of 4 TDs, easily covering the spread in each case.

Vegas' most spectacular miss was the Russel Athletic Bowl (DUSHEE's 2nd worst miss) where the 6 point favorite Sooners were beaten by Clemson by 34, a 40 point miss.

All told, Vegas expected the correct winner in only 18 of the 39 games (compared to DUSHEE's 21 of 39). So keep that in mind the next time you hear someone argue that we should just let Vegas determine the college football playoff rankings.

Vegas and DUSHEE both correctly predicted 14 games and both incorrectly predicted 14 games. DUSHEE correctly predicted 7 games that Vegas didn't:

Bowl | Winner Score1 Loser Score2 | ActualPM DUSHPredPM VegasPredPM

Fiesta | (21) Boise State 38 (12) Arizona 30 | 8 7.91 -3

Poinsettia | Navy 17 San Diego State 16 | 1 4.94 -3.5

MNC | (4) Ohio State 42 (2) Oregon 20 | 22 2.60 -5.5

Cotton | (7) Michigan State 42 (4) Baylor 41 | 1 2.33 -2.5

Sugar | (5) Ohio State 42 (1) Alabama 35 | 7 1.45 -7.5

Quick Lane | Rutgers 40 North Carolina 21 | 19 0.47 -3.5

Military | Virginia Tech 33 Cincinnati 17 | 16 0.34 -2.5

For whatever it's worth, and I don't think it's worth much as I think it's all pretty random, DUSHEE did a better job picking winners in the big games, getting 4 of the "New Years 6" plus the MNC right where Vegas missed (Fiesta, Cotton, Sugar). In particular, DUSHEE had Boise as a substantial favorite against Arizona and nailed the outcome almost exactly. Blind squirrel finds acorn.

On the flip side, Vegas got 4 games that DUSHEE missed:

Bowl | Winner Score1 Loser Score2 | ActualPM DUSHPredPM VegasPredPM

Citrus | (16) Missouri 33 Minnesota 17 | 16 -0.82 4.5

Holiday | (24) Southern California 45 (25) Nebraska 42 | 3 -2.68 7

GoDaddy | Toledo 63 Arkansas State 44 | 19 -2.85 3.5

Las Vegas | (23) Utah 45 Colorado State 10 | 35 -7.76 2.5

Finally, let's look at how DUSHEE would have done if we'd used it to pick games against the spread. If the DUSHEE Predicted Point Difference was larger than the Vegas point spread and the team DUSHEE favored covered, DUSHEE covers. If the DUSHEE Predicted Point Difference was less than the point spread and the Team DUSHEE didn't favor did not cover, DUSHEE covers.

Using this scheme, DUSHEE went 23-16 against the spread during the bowl season. Better than it did straight-up. Go figure. Just goes to show, the "experts" in Vegas can't control random noise any better than anybody else.



0 Comments


Recommended Comments

There are no comments to display.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...