Well a funny thing happened since I posted my predictions using Points vs. Yardage. Two funny things, actually. The first one is, I programmed an error into my automatic score-generating algorithm that affected both sets of predictions, so they don't reflect how the systems would work. The second funny thing is, the erroneous Points system was 67% against the spread! Across 50 games that's pretty impressive. Additionally, 8 of the 9 upsets it picked came true. It was 50% against the over/under. The erroneous yardage system picked about 50/50.
I should have known something was amiss when the Points system picked 39 underdogs! I corrected the results here, and give the results for the proper methods. Correct picks are in green, errors in red.
Favorite Line Underdog Over Points rating Yardage Rating
at South Carolina 3.5 Kentucky 57.5 Kentucky - South Carolina +
at Louisville 14.5 Utah 61.5 Utah + Utah +
at Air Force 5.5 UNLV 44 Air Force - Air Force -
at Alabama 10.5 Houston 56.5 Alabama + HOUSTON +
Arizona State 9 at Wash St 63 Arizona St + Arizona St -
Arkansas State 2.5 at LA-Monroe 57.5 Arkansas St - Arkansas St -
at Army 6.5 Tulane 44 Army - TULANE +
at Auburn 7.5 Vanderbilt 42.5 Auburn - Auburn -
at Ball State 12.5 Central Mich 67.5 Ball St - Ball St +
at Boston College 20 Bowling Green 57 Bowling Green + Bowling Green +
at Clemson 5.5 Virginia Tech 40.5 Clemson + Clemson +
Colorado 8.5 at Baylor 48 Colorado - Colorado +
at Colorado State 13 San Diego St 58 Colorado St + Colorado St +
at Florida State 18 NC State 45.5 Florida St - NC State -
Georgia Tech 3 at Maryland 42 MARYLAND - MARYLAND -
at Hawaii 39 Utah State 69 Utah State + Hawaii +
at Illinois 2.5 Wisconsin 51.5 Illinois + WISCONSIN +
at Indiana 14 Minnesota 67.5 Indiana + Indiana +
at Kansas State 3 Kansas 55 KANSAS + Kansas -
at Kent State 9 Miami OH 48.5 Kent State - Kent State +
at LSU 8 Florida 49 la st + La st -
at LA-Lafayette 8 North Texas 67 LA-Lafayette -= ===== +=
Miami (FL) 6.5 at N Carolina 41.5 N CAROLINA - Miami -
at Michigan 31 Eastern Mich 45.5 E Michigan - E Michigan -
at Michigan State 14.5 Northwestern 53 Michigan State + Michigan State +
at Mississippi 13 Louisiana Tech 53.5 La Tech - LA TECH +
at Mississippi St 19 UAB 50.5 alabama-birm + Miss St +
at Missouri 7 Nebraska 68 Missouri + NEBRASKA +
at Nevada 3.5 Fresno State 55.5 FRESNO STATE + FRESNO STATE -
Northern Illinois 3.5 at Temple 53 TEMPLE - TEMPLE +
Ohio 5 at Buffalo 53.5 Ohio U + BUFFALO -
Ohio State 7 at Purdue 52 Ohio St - Ohio St -
Oklahoma 11 at Texas 57 Oklahoma + Oklahoma -
at Oregon State 4 Arizona 54 ARIZONA - Oregon State -
at Penn State 9.5 Iowa 39 Penn St - Iowa -
at Rutgers 3.5 Cincinnati 52 CINCINNATI + Rutgers +
at San Jose State 7 Idaho 51 Idaho - IDAHO -
at Tennessee 1 Georgia 56.5 GEORGIA + GEORGIA -
at Texas A&M 6.5 Oklahoma State 62 Oklahoma St + Texas A&M +
at Texas Tech 24.5 Iowa State 66.5 Iowa State - Iowa State -
Tulsa 3 at UTEP 72.5 Texas El Paso - Tulsa +
UCF 3.5 at E Carolina 53.5 E Carolina + Central Fla +
at UCLA 20.5 Notre Dame 48 ucla + ucla -
at USC 40 Stanford 58 Stanford - Stanford -
Virginia 10.5 at M Tenn 45.5 M Tennessee + M Tennessee -
Wake Forest 7 at Duke 47.5 Wake Forest - Wake Forest -
West Virginia 26.5 at Syracuse 56 West Va + West VA +
at Western Mich 8.5 Akron 57 Akron - Akron +
at Wyoming 3 TCU 40.5 =Wyoming - =Wyoming -
at Boise State 22.5 New Mexico St 61.5 New Mexico St - NEW MEXICO ST +
Points
Spread: 26 of 49 (53%)
Upsets: 6 of 8 picked occurred
Over/Under: 26 of 49 (53%)
Yardage
Spread: 21 of 48 (44%)
Upsets: 4 of 12 picked occurred
Overs vs. Unders: 27 of 49 (55%)
Common picks: 16 of 32 (50%)
Common over/unders: 17 of 30 (56.6%)
Common upsets:
Maryland over Georgia Tech
Fresno State over Nevada
Temple over Northern Illinois
Georgia over Tennessee
...and of course, the corrected system does much worse than the mistaken one. The Yardage system didn't do anything out of the ordinary, either, which was expected. Strangely it did better in the over/under, but not by a significant amount.
Joining the two together as confirmation of each other's picks didn't help either, except a little bit on the over/under. There weren't enough upsets picked to get a real look at that, so perhaps that's one area that the two might work in tandem, but 3 of 4 is no better than the 6 of 8 the Points system got on its own. Yardage alone was a bad predictor of upsets.
So I probably won't go through the work of calculating the yardage spreads again. Instead, I'll look at the fluke Points system and see how it does next week, using the correct Points system as a control.
Comments