What Should the Louisville vs Florida State Line Be?

The line for FSU vs UofL has had a rocky fall since the opening line was released. The original line released on BOL was FSU -8.5. Since then, there have been 14 line changes over the past 36 hours, causing the line to bounce all the way down to -3, before rebounding a bit and currently sitting at -4. The fact that there have been 14 line changes shows that there is a huge disagreement in the betting marketplace about this matchup. So what is everyone looking at? We’ll use three of Sagarin’s rating systems to take a look.

Predictor – What It Does: The predictor system essentially does a nonlinear regression in excel to solve for power ratings. It essentially retrofits power ratings to minimize the sum of squared errors between each game’s expected outcome and actual outcome. It also uses a few tactics like taking the natural log of the sum of squared errors, rather than simply the absolute value. As a result, this gives diminishing marginal returns to teams who run up the score.

Predictor – What It Doesn’t Do: Predictor cares more about this season, so it doesn’t give much weight to how your team did last year. Sure there is some base starting value from last season, but the further into this season we get the less last season matters. This is honestly the way that it probably should be. 

FSU 85.02
UofL 82.2
HFA 3.52
Line 0.7

The predictor actually has UofL as a slight favorite at home against Florida State, so we can see why tons of bettors were hopping onto the Louisville +8.5 bet.

Elo – What it Does: I grew up with the Elo system from playing chess. It follows a very logical method of looking at who you won and lost against and adjusting accordingly. Beat a good player? You go up. Lose to a bad player? You go down. Lose to a good player, or a beat a bad player? You probably won’t change much in the rankings. While predictor looks more at margin of victory, Elo looks more at wins and losses. For football versions it has been adapted to incorporate margin of victory to some extent, but the bulk of it looks at wins and losses.

Elo – What it Doesn’t Do: Elo doesn’t understand the concept of a “season” as chess doesn’t have seasons. As a result, the only way to adjust for a new season is to do something like 2/3*Last Year Elo + 1/3*Average Elo for All Teams = New Starting Season Elo. This is a pretty bad method as it doesn’t take into account factors like returning starters, coaching changes, etc. It also allows teams who were undefeated last year to sustain their high Elo rating, as long as they keep winning, regardless of their poor margin of victory this season. (For the record, BOL often uses this method the most for their opening lines).

FSU 93.13
UofL 80.19
HFA 3.15
Line -9.79

Under this method, Florida State is almost a 10 point favorite. Again, you can see why there is some huge disagreement on this one.

Sagarin Rating – The Sagarin rating is a combination of multiple handicapping methods, most notably the two listed above. As you can see, combining the two gives us a line of -4.34 which is around where the line has currently stabilized.

FSU 88.34
UofL 80.72
HFA 3.28
Line -4.34



Florida State is 1-6 ATS, with an ATS margin of -8.4. The only other team with a worse  record ATS is UCLA, but they only have a margin of -4.9. In other words, Florida State has been one of the most unprofitable team to bet on against the spread.

Louisville has been a pretty average 5-3, with an ATS margin of +3.4.

What about Thursday nights?

Underdogs of 3.5 to 10 points playing in Thursday night games have covered over 55% of the time, with a sample size of more than 700 games. There’s something about playing on a different night that makes favorites more prone to upsets. A 55% cover rate is about equivalent to 3 points on the spread.

Depending on what method you use though, Louisville may not be a 3.5 point underdog, or even an underdog, so it’s hard to really make an adjustment for that.

How would I handicap this?

Predictor is a better rating, but the Elo is so high that an adjustment must be made. Let’s take the predictor line as-is, and then let’s try to adjust the Elo for FSU’s poor cover rate. Then finally we’ll slant the line a few points in Louisville’s favor for the Thursday night game. Not 3 points because they might not really be 3.5 point underdogs, but mentally they’re underdogs, so let’s take a 0 to 3 point range and use that as our expected range for the closing game line.

1. Predictor = Louisville -0.7

2. New Elo = -9.79 – ln(8.4 ATS margin) = FSU -7.66

3. Combined Rating = FSU -3.5

4. Adjust for Thursday = [ FSU -3.5 to FSU -0.5 ]

5. FINAL LINE RANGE: FSU -3.5 to FSU -0.5

6. Win Probabilities:

High – FSU 65.3%, UofL 34.7%

Low – FSU 51.7%, UofL 48.3%

Week 10 Bets – Part 1

Georgia Southern -20.5 vs Troy – If you were a very keen observer of opening lines, you’d have found a gem at about 3 pm EST, when BOL released the line with the sign wrong. They released GSU +20.5 (oops!) and a few lucky souls got free money. And an intern probably got fired. Unfortunately I wasn’t one of the few who caught that in time, but I did get in on GSU -20.5 after that error as fixed. Wow, this line jumped quick. Within a few hours this line moved to 25, which is more realistic. App State rushed for 441 yards on these jokers from Troy, and Georgia Southern is #1 in rushing yards per game. So I bet GSU will hit the 500+ yard rushing mark for sure. Maybe they’ll even hit 600+ in back to back games. #SouthernNotState!

Wisconsin -11 at Rutgers: Rutgers hasn’t been the best at defending against the big bad run teams in the Big Ten (MOV -39 at Ohio State, -18 at Nebraska).  Gary Nova, the very good quarterback for Rutgers who is #4 in YPA, is still on crutches and highly questionable. Only one book has this line up…I’m not sure if they’re optimistic he will play or just don’t know he’s injured, but it appears they have made no adjustments for this. On top of Gary Nova, Rutgers also lost a WR, TE, and OL, and has another questionable LB, DB, and OL. Meanwhile, Wisconsin has no new injuries.

Purdue +24.5 at Nebraska – This looks fine on paper, but man that’s a lot of points! Purdue has shown that it can do just fine against strong running teams, and even though Nebraska should win this without too much trouble, I can’t seem to justify such a high spread. Looking back at Purdue’s matchups:

@ Notre Dame -16

vs Iowa -14

@ Illinois +11

vs Michigan State -14

@ Minnesota -1

Cincinnati -2.5 at Tulane – Gunner Kiel is injured, but if you can wash that opening game out of your mind, as Gunner Kiel has not sustained the type of performance that he did against Toledo:

Toledo 25 37 418 67.6% 52 6 0 216
Rest of Season 114 197 1631 57.9% 83 14 6 145

The replacement behind Kiel is not some freshman; it’s actually senior Munchie Legaux, who at once point thought he would never play again due to a career ending injury. He’s not new to the scene either – he has played in 20 games for the Bearcats, including 10 starts. In a heartwarming and emotional performance to beat USF, he went 14 for 15, at 8.0 YPA. On paper, Cincinnati is a 10 point favorite in this game, and I don’t see Kiel’s VORP being 7.5.

 Utah State pk at Hawaii – Utah State is an unlucky team. As if it weren’t hard enough returning the fewest starters in FBS, they have withstood so many injuries that they’re now on their 4th string QB. After their third string senior went down, they’re now starting freshman Kent Myers of Rowlett, TX. The kid put up good numbers in high school, but we basically have no data on how he’ll do in college. However, with all that said, Utah State is a 7 point raw favorite in this game, so it’s the VORP debate again as to how valuable this guy is over the third string senior. The line has since moved from pk to -3, so I’m considering either taking a middle with the spread or taking an arbitrage on the moneyline.

Duke +1.5 at Pitt: Did you know Duke is ranked again? And they are 6-1 with a loss at Miami that doesn’t look too bad. The lowest I can get to when I try to put together this line is Duke -2.5 so I’m still trying to figure out Duke being a slight underdog. To be honest I haven’t watched a full Duke game all year so maybe I should shy away from these, but I’ve watched Pitt enough to know that I don’t want to be backing Pitt.

Results After Week 9

Week 9 Results:

Overall: 37-33 (52.9%)

This week: 4-2 (66.7%)

Mississippi State -13 at Kentucky (W) – I believe this line closed at 14.5, and if that’s the case I’m glad I got in at the right time. I know Kentucky isn’t as bad as they usually are, and I know that it was on the road, but my gosh guys. This game should have been put away sooner.

Ole Miss -3 at LSU (L) – I kinda thought about getting out of this one after seeing that gameday crowd. Oh well. I’m still trying to figure out how these two teams combined for over 700 yards and scored 17 points.

Memphis -23 at SMU (W) – The only winless team in FBS. Memphis head coach Justin Fuente was a former TCU offensive coordinator, so no surprise that he had a little fire in his heart and put up some points against the ponies. his is the second game in which they’ve scored a touchdown, so congrats to them on that.

Arkansas -21 vs UAB (W) – If you needed any explanation for why I thought Arkansas would destroy UAB, please see my previous post regarding strong rushing teams against non-P5 spread teams. It’s just not fair to ask a team who has been taught to defend a spread attack, and has recruited in such a way, to match up in the trenches against a team like Arkansas, or Wisconsin, or Nebraska.

West Virginia +3 at Oklahoma State (W) – Part of me was worried that I was over-reacting when I took this line, but once I saw the line move from WVU +3 to WVU -2, I felt a lot better. When you adjust for home field advantage, WVU handled Oklahoma State very similar to how TCU did. I never thought I’d be saying this, but that West Virginia defense sure is good.

Wake Forest UNDER 14 (L) – I’ve got to stop playing these, they’re killing me. Wake was down 23-3 with 8 minutes to go, and then decided to score 14 points in garbage time to hit the over.

Do More Plays Equal More Penalties?

It has become well known that teams like Baylor and Texas Tech lead the nation in penalty yards per game. It is often attributed to their fast pace, and how many plays they run per game. However, if this were simply the case, we would expect to see a constant amount of penalties per play in each team. Instead, we see that teams vary greatly in the amount of penalties they incur per play. I have converted these to “chances of penalty in any given play”:

#1 Akron 7%

#2 Tulane 6%

#3 Marshall 6%

#4 Oregon State 6%

#5 Baylor 6%

#126 Air Force 2%

#127 Navy 2%

#128 Old Dominion 2%

It is very clear that plays per game is not a good enough descriptor. I ran a regression to confirm, and found that only 3% of a team’s penalty yards can be attributed to number of plays.

Regression Statistics
Multiple R 0.335476336
R Square 0.112544372
Adjusted R Square 0.109050452
Standard Error 7.586090412
Observations 256
df SS MS F Significance F
Regression 1 1853.731707 1853.731707 32.21149262 3.76335E-08
Residual 254 14617.387 57.54876773
Total 255 16471.11871
Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0%
Intercept 22.77018666 1.879675776 12.11389057 5.80503E-27 19.06845179 26.47192152 19.06845179 26.47192152
PEN 0.196230348 0.034574885 5.675516947 3.76335E-08 0.128140383 0.264320314 0.128140383 0.264320314


I then realized that passing plays on average tend to incur a higher average penalty yards per play. So with another regression, we found that 11% of a team’s penalty yards can be attributed to number of passing plays.


Regression Statistics
Multiple R 0.335476336
R Square 0.112544372
Adjusted R Square 0.109050452
Standard Error 7.586090412
Observations 256
df SS MS F Significance F
Regression 1 1853.731707 1853.731707 32.21149262 3.76335E-08
Residual 254 14617.387 57.54876773
Total 255 16471.11871
Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0%
Intercept 22.77018666 1.879675776 12.11389057 5.80503E-27 19.06845179 26.47192152 19.06845179 26.47192152
PEN 0.196230348 0.034574885 5.675516947 3.76335E-08 0.128140383 0.264320314 0.128140383 0.264320314


There still remain another 89% that is yet to be explained – other factors that could cause a team to incur more or fewer penalty yards than average could come from:

  • Aggression – Baylor, for example, often plays tight pressure on receivers because they know that they will not be called for every pass interference, and so playing that style is +EV.
  • Offensive line – Needless to say, false starts and holding penalties can stem from an inexperienced offensive line
  • Inexperience – aside from the OL, a more inexperienced team is also likely to incur penalties like holding, delay of game, substitution infractions, etc.
  • Thuggishness – Some teams just tend to get more penalties for unsportsmanlike conduct, targeting, etc.

Week 9 Bets – Part 2


Wake Forest UNDER 14 points vs Boston College

Wake has yet to score 14 points in conference play: Louisville (10), Syracuse (7), FSU (3). I can’t see them scoring more against BC.

Wake ranks 125th in scoring offense (13 ppg) and DEAD LAST in total yardage (176 ypg). Now they face a defense that is ranked #16 in scoring defense and #18 in total defense. They now face the 2nd best defense they’ll face all season (#1 being Louisville). Against Louisville, Wake scored 10 points – but one was a defensive TD scored on a fumble in the end zone. And the field goal came due to a fumble as well. In fact, the Deaks only gained 100 yards in that game.

Sure, they’re at home. But would you really be intimidated by this crowd? The total is a very low 40 points, and the spread is 12, so when Vegas opens team totals in Saturday morning they should put this up around 14.

Week 9 Bets – Part 1

This week is heavy in SEC bets. I don’t even cover the SEC that much so maybe I’ve just lost my marbles.

Mississippi State -13 at Kentucky: I have the line at 16.2 and I’ll definitely take the value on a Mississippi State team that’s done a great job of covering the spread this year. LSU drew up a pretty good blueprint for how to dominate UK last week, so let’s see if Miss St can put that to good use. UK fans thought College Gameday might have been coming to their game this week, which is hilarious.

Ole Miss -3 at LSU: The thing about LSU is…they tend to bully around all the weaker opponents, but fail to beat the rest of their SEC West competition. Case in point: NMSU 63-7, ULM 31-0, SHSU 56-0, Kentucky 41-3. But when it comes to the tougher competition, LSU has either played a close game, or gotten blown out.

Memphis -23 at SMU: My feelings won’t change too much for SMU this year. Those snobs have left their football team for dead, and as if attendance wasn’t bad enough, now literally no one is coming. I was not too sure how a shaky and inconsistent Cincy team would do covering the 13 point spread. Now I feel much better about Memphis.

Arkansas -21 vs UAB: The UAB vs Mississippi State game, which was a close game, will forever remain a mystery to me. With that said, Arkansas has a pretty good track record (as most rushing teams do) of running up the score on weaker opponents. UAB’s defense is really designed to stop offenses that are, well, similar to theirs. They managed to hang with Mississippi State due to 3 TD passes or 75+ yards. I’ll take Arkansas at home.

West Virginia +3 at Oklahoma State: I have the game handicapped at pk, and I think anyone who watched last week’s games will be all over WVU. I don’t want to give WVU too much credit here, because while they did beat Baylor and almost beat WVU, they also almost lost on the road to Tech. Oklahoma State showed that their young secondary has some work to do. If they struggled against TCU’s offense, so WVU should be able to put up some points as well. From a defensive standpoint, I’m not sure what to think. TCU implemented some changes in its defensive formations in order to push Oklahoma State to run rather than pass. With a younger defensive line, I’m not sure if that’s going to be WVU’s plan of attack here. Either way though, I’ve got to take WVU due to the offensive performances that they’ve shown so far this season.



Missed Out On:

Auburn -14 vs South Carolina


Results Through Week 8

Week 8 Results:

Overall: 33-31 (51.5%)

This week: 2-5 (28.5%)

Well, this one hurts. From the standpoint of who won and lost, this week made a lot of sense. But from a margin of victory perspective, this week was crazy.

Michigan State -14 at Indiana (W) – Zander Diamont was the poor soul who was tasked with taking on the Michigan State defense as a freshman. Stat line? 5 for 15 (33%), 11 TOTAL YARDS, for a QBR of 7.4. Think about that – he passed for 1.36 yards per completion. With that said, the run game actually held up very well.

Washington +19 at Oregon (L) –  Washington didn’t really come to play in this rivalry game, and Cyler Miles getting injured didn’t help much either.

Louisville -17 vs NC State (L) – Talk about a lackluster win for the Cards. I would say they look like they can’t beat FSU after this, but FSU played an even worse game against NC State.

South Alabama -17 vs Georgia State (L) –  This game was the second strangest game. How did a team that’s probably a frontrunner to win the Sun Belt almost lose at home to Georgia State, who is probably one of the three worst teams in FBS (hi there, EMU and SMU)? I don’t know because this game wasn’t on TV. But from the stat line, it looked like a very evenly matched game.

Florida -4.5 vs Missouri (L) – This game gets the award for biggest head-scratcher. Mizzou scored 42 points despite only gaining 119 yards. Think about that. Their yards-adjusted points were 8.5, yet they scored 42 thanks to multiple special teams and defensive scoring plays as Florida turned the ball over six times and allowed both a kick return and punt return. Here’s my favorite stat – in the past 10 years when holding opponents to 120 yards or fewer, 147 teams have won and two have lost. The two that lost were both Florida under Will Muschamp.

Texas A&M +14 at Alabama (L) – Had I known Kenny “Trill” was drunk on Thurday night at Chimy’s I wouldn’t have taken this bet. My big question is – was the entire A&M defense hungover too?

Nevada +11.5 at BYU (W) – One of the few games that made sense. To be honest, Stewart found his rhythm and did a lot better than expected. It was really the 3 fumbles that doomed BYU in this one.

Week 8 Bets – Part 1

Michigan State -14 at Indiana: Indiana starting QB Nate Sudfeld is out for the season, and now the backup Chris Covington is out as well. That means IU will be starting their third string QB against Michigan State. Have fun with that. Fair value is 12.5 with IU having their starting QB, so 14 at this point is a great deal. Sure, IU’s offense is actually run by this guy named Tevin Coleman…maybe you’ve heard of him. But without any threat of passing, IU’s run game becomes very one dimensional. The opposing defense will be able to stack the box, stop Coleman and dare the freshman QB to pass the ball and make a mistake. Oh, and did I mention the opposing defense is Michigan State? Yes? Ok. This one should be fun.

Washington +19 at Oregon: I had this line at 13 so I was really surprised to see it come out this high. Oregon is still getting experience on the O-line and while things appear to be stabilizing, there’s no reason to assume a massive 6 point adjustment for anything this week. Strangely enough, the line actually moved up to 21 so I guess I missed out on that.

Louisville -17 vs NC State: I have the game handicapped at a range of 14 to 20, so an average of 17 is fair. But this doesn’t account for the 7 NC State players that are suspended for the BB gun fight. It appears there was no real line move from this. I get it…they’re not the QB or anything. But the effects of all that add up. A new OL in the mix does have an effect on the running game. Playing a backup WR may put more pressure on a veteran WR who will now be seeing tighter coverage. A backup DB will likely be tested more on the long balll. Add all these things up and you get what I think should be a substantial move of at least 3-4 points.

South Alabama -17 vs Georgia State: I have this line at 20.5 without any adjustments. Add onto that the lesser-known fact that Georgia State withstood a lot of injuries in the last game and we’ve got a great bet!

Florida -4.5 vs Missouri: Not an over-reaction to Mizzou…I have Florida as a 10 point favorite or so here. I’ll be honest, I’m really not a fan of dual quarterback systems. But Florida doesn’t score that many points anyways, so how bad could it be?

Texas A&M +14 at Alabama: I have this game handicapped anywhere between 4 to 10 points, and plus this game has a bit of a rivalry flare. I don’t like the game being at Alabama too much, but even if A&M is out of the game late, we’ve seen them have some 4th quarter surges and that makes them pretty likely to cover here.

Nevada +11.5 at BYU: So Taysom Hill is gone. Obviously that’s a huge blow. Let’s take a look at backup Christian Stewart’s stat line from the UCF game:

Comp Pct 48.5%, 4.92 YPA, 3 TD, 4 INT

Raw QBR 18.4 , Adj QBR 23.6

That’s a huge step down from Taysom Hill who was having a great year. BYU only passed for 150 yards and somehow won the game due to 49 rushing attempts for 189 yards on the ground.

But wait, there’s more! BYU withstood a slew of injuries against UCF and will now be without many other starters:

Out: WR

Doubtful: LB, OL

Questionable: RB, DB, WR

That’s a lot of guys to lose. The line is completely fair at 11.5 with everyone healthy. But stack on all these injuries, and I’m going with Nevada.

Results Through Week 7

Week 7 Results:

Overall: 32-26 (55.2%)

This week: 5-3 (62.5%)

TCU +10.5 at Baylor (W) – This line was such a joke. So are those officials. I’ll be drinking myself into oblivion tonight. How do you drop a 21 point lead with 12 minutes to go?? 🙁 I’m still only halfway realizing that we actually lost this one. I’ll take the free $100 and hope it covers my bar tab.

West Virginia -2 at Texas Tech (W) – HA. HA. HA. Take that, Ryan Gossling! And take your free waffles with you on the way home!

Arkansas +10.5 vs Alabama (W) – Looks like cocky midwesterner Bret Bielema covered the spread after his bye week. But no win for him and his trophy wife. Still not as good as Bobby Petrino though. Bill Clinton must really miss living in Arkansas.

FIU +13 at UTSA (W) – I didn’t watch this game cause I was watching the 5.5 hour long TCU game, but it looks from the score like it went as expected. This game was the anti-TCU-Baylor game. They scored 29 total points, versus the 119 points in the other game. UTSA was supposed to have so much good leadership this year. I’m not sure if that leadership has turned to apathy or what has happened, but they are really struggling to put up points on offense.

Texas +14.5 vs Oklahoma (W) – Always take the underdog in rivalry games. Thanks UT for making it close…in the third quarter I was worried it would get out of hand. Now I’ve got 200 tickets for the State Fair tomorrow 🙂 my biggest worry was that Striker would be able to blitz a ton since UT wasn’t really going to spread out OU to keep him on the outside to defend tight ends. But hey, they did a surprisingly good job in the passing game, enough to force him out of ends box.

Temple -14 vs Tulsa (L) – Very lackluster performance by Temple today. I want to blame this on that missed field goal, but really it was the whole game. Tulsa really looked like they could hang with them. So I’m not sure how to mentally process that, other than to think there is a lot of parity in the AAC…more than I imagined.

Notre Dame -16.5 vs North Carolina (L) – What the…? I’m just going to go with the notion that ND played down to their opponent. Maybe they were tired from Stanford. But seriously, these guys hadn’t given up more than 17 points yet in a game. Where did 43 points come from? Anyways, they get s hall pass this week since everyone just wants to see how they do against Florida State.

Penn State +1.5 at Michigan (L) – They won in the Big House. It’s a miracle. Their isn’t even that bad, but by Michigan standards they think it’s terrible to have a half-empty upper deck. Or upper…rim. Whatever it is when you just have one giant level. Hackenberg had a good game, except for getting caught for a safety. Overall this one was a snoozer (what B1G game isn’t) and unfortunately Michigan was more explosive in the end, with their 57 plays. Woo, fun.