A losing record in the third week and a review of Sagarin's predicted lines and retrodictive accuracy caused me to revisit my predicted margin formula. So below are the season results so far followed by this week's predictions using the new formula. What I noticed was that I was more accurate on games between closely matched teams than Sagarin (lower predicted margins) but worse on games between unevenly matched teams (predicted blowouts). I tweaked some things and hopefully it will improve. As always, if you straight use this advice to gamble without adding common sense, you deserve to lose the money.
| Confidence | Record |
|---|---|
| Overall | 83-69-2 |
| 14+ | 7-4-0 |
| 10+ | 7-8-1 |
| 7+ | 15-17-0 |
| 54-40-1 |
This week's lines:
| Line | Pick | Conf |
|---|---|---|
| Penn St. -2 @ Ohio St. | Penn St. | 8.20 |
| Oklahoma -19 @ Kansas St. | Oklahoma | 7.63 |
| Ohio +4 @ Temple | Temple | 7.04 |
| Georgia +2.5 @ LSU | Georgia | 6.84 |
| Rice +2.5 @ Tulane | Rice | 6.71 |
| Eastern Michigan +24 @ Ball St. | Ball St. | 6.42 |
| Boise St. -7.5 @ San Jose St. | Boise St. | 6.39 |
| Texas Tech +2.5 @ Kansas | Texas Tech | 6.18 |
| Colorado +22 @ Missouri | Missouri | 6.10 |
| UCLA +17.5 @ California | UCLA | 5.71 |
| Rutgers +9.5 @ Pittsburgh | Pittsburgh | 5.70 |
| Nevada -3.5 @ Hawaii | Hawaii | 5.51 |
| Cincinnati -1 @ Connecticut | Connecticut | 5.30 |
| Minnesota pk @ Purdue | Minnesota | 5.23 |
| Oklahoma St. +13 @ Texas | Texas | 5.17 |
| SMU +12.5 @ Navy | Navy | 4.96 |
| UNLV +23 @ BYU | UNLV | 4.83 |
| Michigan St. -5.5 @ Michigan | Michigan St. | 4.72 |
| Wyoming +32 @ TCU | TCU | 4.69 |
| Middle Tennessee St. +11.5 @ Mississippi St. | Middle Tennessee St. | 4.46 |
| Duke +9.5 @ Vanderbilt | Duke | 4.03 |
| Mississippi -4.5 @ Arkansas | Mississippi | 4.03 |
| Baylor +10.5 @ Nebraska | Nebraska | 3.97 |
| Auburn +3.5 @ West Virginia | Auburn | 3.83 |
| Notre Dame -11.5 @ Washington | Washington | 3.75 |
| Troy -25 @ North Texas | Troy | 3.61 |
| Bowling Green +8 @ Northern Illinois | Northern Illinois | 3.47 |
| Virginia Tech +4.5 @ Florida St. | Virginia Tech | 3.29 |
| Kentucky +24 @ Florida | Kentucky | 3.07 |
| Wake Forest +3 @ Miami (FL) | Miami (FL) | 3.05 |
| Fresno St. -17 @ Utah St. | Utah St. | 2.99 |
| Colorado St. -7.5 @ San Diego St. | San Diego St. | 2.90 |
| Texas A&M +3.5 @ Iowa St. | Iowa St. | 2.89 |
| Virginia +11.5 @ Georgia Tech | Virginia | 2.59 |
| Florida Atlantic -2 @ Louisiana-Monroe | Florida Atlantic | 2.53 |
| Southern Cal -15 @ Arizona | Arizona | 1.99 |
| Alabama -7 @ Tennessee | Alabama | 1.97 |
| New Mexico St. -15 @ Idaho | New Mexico St. | 1.75 |
| Central Florida +22.5 @ Tulsa | Tulsa | 1.53 |
| Northwestern -9 @ Indiana | Northwestern | 1.32 |
| Central Michigan -3.5 @ Toledo | Central Michigan | 0.99 |
| New Mexico +5.5 @ Air Force | New Mexico | 0.95 |
| South Florida -4.5 @ Louisville | South Florida | 0.90 |
| North Carolina St. +10.5 @ Maryland | North Carolina St. | 0.58 |
| Boston College +3 @ North Carolina | North Carolina | 0.57 |
| Southern Miss pk @ Memphis | Memphis | 0.43 |
| Kent St. +7.5 @ Miami (OH) | Miami (OH) | 0.39 |
| Louisiana Tech +1.5 @ Army | Louisiana Tech | 0.35 |
| Oregon -4 @ Arizona St. | Oregon | 0.34 |
| Illinois -2 @ Wisconsin | Wisconsin | 0.24 |
A couple of things to note. The first is that I've mentioned elsewhere that I don't like the Penn St. prediction. Also, the combination of the new formula and the later date in the season has clearly limited the number of high confidence picks from the system. Of course, given the performance so far this year that might be a good thing. The formula seemed to have the intended effect, though, as the number of high confidence numbers spit out in blowout spreads was way down.