Now that the 2012-13 college football bowl season has drawn to close, it’s time to take a look back at our preseason projections, and evaluate how we’ve done.
This is the second year we’ve simulated the entire season (minus bowl games) before it began. Last year, things went quite well, and our preseason value picks provided exactly what they were supposed to: value.
As with any new endeavor, it was hard to tell whether we’d been good, good and lucky, or just plain lucky. But a closer inspection shows this season’s projections, while not perfect, were again an overall success:
- Our team win total picks and our BCS conference and division winner value picks were both profitable
- Our biggest disagreements with the public consensus were right twice as often as they were wrong
- 7 of the 10 BCS bowl teams were in the top 20 of our 2012 preseason college football rankings, compared to only 4 in the top 20 of the preseason AP poll
Let’s check out the details.
How Did Our Highlighted Team Projections Do?
Notre Dame. (#13 in our preseason ratings, UNRANKED in the preseason AP) – I think we all know how this one turned out. The Irish were outmatched in the BCS title game, and it took a few lucky bounces (missed field goal by Pittsburgh, overtime against BYU, etc) to get them there, but they were clearly one of the better teams in the country. Here’s what we said about them in the preseason: ”The Irish have been discounted by many this year, not appearing in the AP’s top 25. We’re more optimistic about their talent, ranking them 13th among FBS teams. A rating that high means a few lucky bounces could put them in a BCS bowl.” This one is clearly a HIT.
Florida (#10 in our preseason ratings, #23 in the preseason AP) – The Gators finished #9 in the final AP poll, almost exactly where we slotted them in the preseason. Why did we like them? We said about the Gators: “They have a strong program history, poor 2011 turnover margin (read: bad luck which shouldn’t repeat itself), and loads of returning starters. Those three qualities have historically foreshadowed a team’s improvement, and we think it’ll continue with Florida.” That’s another big HIT.
The Big 12. The Big 12 placed six teams in our preseason top 25, and four in our top 8, plus seven of the ten toughest schedules belonged to Big 12 teams. But take a look at the final AP poll, and the highest ranked Big 12 team is Kansas State, at #12. Unfortunately, this one was a MISS.
Oregon. “Oregon Favored In the Pac-12 Over USC?!” The general consensus before the season started was that USC was the best team in the Pac-12, and possibly the nation. However, we projected the Ducks as the most likely Pac-12 champs, and rated them #3 in the country, ahead of USC. While neither team ended up winning the official Pac-12 title, Oregon finished ranked #2 in the nation, and was likely the conference’s best team. That’s a HIT.
Oklahoma. The preseason polls had Oklahoma #4, while we had them #1 by a hair over Alabama. The Sooners had a disappointing season, and finished ranked #15 in the final AP poll. Clearly this one was a MISS.
Wisconsin/Ohio State. Our claim here was that even though we only rated Wisconsin as the third best Big Ten team, Ohio State’s (and Penn State’s) postseason ban gave them a huge leg up in the race for the Big Ten championship, because their biggest threats in the Leaders division were out of the running. Well, we were right about Ohio State being the biggest threat to the Badgers, as the Buckeyes finished the year as the only undefeated BCS team. Penn State also snuck past the Badgers into second place in the division. Wisconsin finished only 4-4 in the regular season, yet parlayed that 0.500 record into a spot in the conference championship game, where they proceeded to absolutely destroy Nebraska, earning themselves a Rose Bowl berth. So this is a big HIT.
That’s 4 hits and 2 misses. Given that all of these predictions were our “wow, TR is really going out on a limb there” anti-consensus highlights, we’ll take that performance any day.
Regular Season Win Total Value Pick Performance
Our regular season win total picks went 12-13, but because many of our picks were on the less likely side of the line, with positive payout odds, at the then-prevailing lines the net profit from betting one unit on each pick ended up at +4.2 units, for an ROI of +17%.
Below is a table of all our win total picks, ranked by projected ROI.
Team Actual Wins TR Pick Proj. ROI Result Net Units
Ohio State 12 Under 8.5 (+288) 120% loss (by 3.5) -1
Mississippi 6 Over 5.5 (+379) 90% WIN (by 0.5) +3.79
Florida St 10 Under 9.5 (+242) 84% loss (by 0.5) -1
Georgia 11 Under 9.5 (+152) 78% loss (by 1.5) -1
Vanderbilt 8 Over 6.5 (+150) 63% WIN (by 1.5) +1.50
South Carolina 10 Under 8.5 (+124) 62% loss (by 1.5) -1
Wisconsin 7 Under 8.5 (+266) 60% WIN (by 1.5) +2.66
Texas A&M 10 Over 7.5 (+158) 57% WIN (by 2.5) +1.58
USC 7 Under 10.5 (+116) 54% WIN (by 3.5) +1.16
Michigan St 6 Under 8.5 (+103) 47% WIN (by 2.5) +1.03
Kentucky 2 Over 5.5 (+350) 42% loss (by 3.5) -1
Texas 8 Under 8.5 (+190) 40% WIN (by 0.5) +1.90
Virginia Tech 6 Under 9.5 (-125) 37% WIN (by 3.5) +0.80
West Virginia 7 Under 8.5 (-111) 37% WIN (by 1.5) +0.90
TCU 7 Over 8.5 (+153) 33% loss (by 1.5) -1
Clemson 10 Under 8.5 (-101) 31% loss (by 1.5) -1
Oregon 11 Under 10.5 (+119) 28% loss (by 0.5) -1
Missouri 5 Over 7.5 (+146) 22% loss (by 2.5) -1
Louisville 10 Under 9.5 (-309) 18% loss (by 0.5) -1
Marshall 5 Over 6.5 (+249) 14% loss (by 1.5) -1
Alabama 11 Under 10.5 (-118) 12% loss (by 0.5) -1
Duke 6 Over 3.5 (-243) 8% WIN (by 2.5) +0.41
Arkansas 4 Under 8.5 (-107) 7% WIN (by 4.5) +0.93
Nebraska 10 Under 8.5 (+130) 6% loss (by 1.5) -1
Auburn 3 Under 7.5 (-177) 5% WIN (by 4.5) +0.56
Total 42% proj. ROI 12-13 (avg 0.3 wins in our favor) 4.23 (17% ROI)
Conference And Division Champ Value Pick Performance
Our conference or division winners had a similarly positive outcome — or even better, depending on how you grade a few cases where we highlighted two bets for a team.
For some conference and division picks, it was clear one team had some value, and we highlighted both a conference line and a division line. In the table below, we’ve graded each of these as half a unit bet on each option:
|Georgia Tech||ACC Coastal||+300||0.5||WIN||+1.5|
|Boston College||ACC Atlantic||+2000||1||loss||-1|
|Illinois||Big Ten Leaders||+500||1||loss||-1|
|Iowa||Big Ten Legends||+800||1||loss||-1|
|Oklahoma State||Big 12||+2500||1||loss||-1|
|Total*||10||2.5-10.5||10.5 (+105% ROI)|
Using that approach, our preseason conference and division picks had a net profit of +10.5 units on 10 units initially bet, for an ROI of 105%, meaning they would have doubled your money. That’s not as good as last year — when our college football conference and division value picks would have tripled your money — but we’ll take it.
If you just stuck to the more conservative option (a division bet) for each team, you’d have seen a +7.5 unit profit (+75% ROI). And if you went with the riskier option (a conference bet) for each, the results would have been +13.5 units (+135% ROI).
We’re happy with the results from this season, and are looking forward to improving our projections in the offseason. We hope you’ll be around this fall to follow along.