Summary: I am betting Over and if it looks like Desormeaux will be 100%, ULL minus the points
For the folks with more idle time,
ULL has an incredible running game with QB Desormeaux and RB Fenroy being about the most productive tandem in the country. Because of this, the passing game clicks right along and has provided at least 1 TD each game this year. The FIU defense they?ll be up against is quite similar to the Sun Belt defenses they have already faced this year, particularly Arkansas State?s. The big question is whether Desormeaux will play (he sprained his knee a couple of weeks ago). He missed the Arkansas St game, but the coach is hopeful he?ll be back and he?s listed as Probable (they are coming off of a bye week). Arkansas State?s passing and rushing defenses aren?t much better than FIU?s, so while Desormeaux would be more productive assuming he was healthy, back-up QB McGuire should be able to turn in a performance similar to his Arkansas State game if called upon. With Desormeaux running things, I?d predict about 63 plays and around 600 yards of offense. That equates to about 45 points for ULL this year. If McGuire plays, or if Desormeaux really can?t run, we?re probably looking at about 68 plays for around 500 yards and about 40 points.
On the other side, FIU hasn?t really faced any conference defenses like ULL?s, and I don?t mean that in a good way. ULL isn?t very good at stopping the pass or the run. That said, FIU has shown a tendency to run if they can, regardless of whether passing would be easier. Of course, when the need to catch up is there, they turn to the pass and do a pretty good job of it. I suspect that they will be able to run very well on ULL, but that scoring pressure will cause them to move more and more to the pass over the course of the game. If they perform similarly to how they have done in conference games this year, I would think they?ll have about 70 plays, with around 40 coming in the second half. This leads to roughly 130 yards rushing and 260 yards passing, with 4 TDs and an interception ? say about 30 points.
This leads to a direct calculation of the expected spread of 10-15 points.
The Sagarin Predictors for ULL and FIU are 65.23 and 64.47 respectively for a predicted spread of 1. Throwing in the generic Sagarin HFA of 2.68, we get a spread of ULL -4. Since this is very different from my direct calculation, I tried to figure out what is happening. At first glance, it seems that FIU?s predictor is too low, if anything, as they opened with 4(ish) strong out-of-conference games at Kansas, Iowa South Florida and Toledo while ULL opened with a weaker Southern Miss, Illinois, Kent State and Kansas State. This is confirmed by FIU?s Strength of Schedule ranking of 67, as opposed to ULL?s ranking of 121. Since conference play began, both have played a roughly similar mix of teams. Now, I have to admit that I don?t know exactly what is happening in Jeff Sagarin?s Black Box, but I am fairly certain that Predictors weigh recent performance heavier than performance many games back. This only makes sense since the whole point is to predict point spreads as of now. On the other hand, Strength of Schedule counts all opponents equally. Since I do not use the early, out-of-conference games at all in developing my projections for the game, this seems to jibe.
Looking strictly at the last 3 games of ULL and FIU, ULL has averaged scoring 44 points and giving up 29. FIU has averaged scoring 32 and giving up 21. Since their conference schedules have been similar in difficulty (about 4.5 predictor points harder for FIU, but that is all from Troy, to whom they lost by 10), it seems that we can assume that the differences in recent schedule can be largely ignored, giving us Margin of Victory as a proxy for the Predictor. The average MoVs for ULL and FIU are 15 and 11 respectively. Since HFA is built into these numbers already, we come up with the same 4 point difference. Now that I can understand the Sagarin spread a little better, I am going to see what happens when I look at the teams and their conference performance from a slightly different angle.
The correlation between ULL?s total yards and the number of points they give up is nearly perfect. This makes some sense in light of their lack of ability to stop anyone ? they score, and then their opponent usually scores. Fortunately for them, their D is just good enough to get a few more stops than the offense experiences. Based on conference play, using this correlation leads to ULL giving up 31 points, right in line with the earlier projection. The correlation between FIU?s points scored and their opponent?s average points allowed is nearly as good and leads to an estimated FIU score of 32 points. This does have the problem of using average points allowed that includes out-of-conference games of the opponents, but there are only so many calculations that I am willing to do. Based on all of this, I am willing to go with 30-32 as my estimate of FIU?s score.
The problem is that I cannot do the same thing on the ULL side because of the injury to Desormeaux. The points from the Arkansas State game (McGuire at QB) were earned in a completely different way from the games under Desormeaux. Furthermore, Arkansas State was a much better opponent than the conference opponents that Desormeaux faced. Using the very limited number of conference games played this year is already pushing the limits of using the stats as a projection device, so I cannot break them down further and do the analysis separately for each QB. The best I can do is to assume that the offense will perform roughly as before when it has been faced with similar opponents. If Desormeaux plays, I am quite confident they will. He?s put up about 1800 yards of total offense so far this year and is on pace for 1000 yards rushing again. You can count on Fenroy to rack up the rushing yards, so I feel good about the 45 points predicted for this combo. Redshirt freshman McGuire has exactly one game under his belt, so who knows how he?ll perform, but the defense he faced from Arkansas State is better than what he?ll see from FIU. Another encouraging thing about McGuire is that he had to bring them back from behind twice before they finally pulled away at the end. The problem is that he cannot run as well as Desormeaux, so FIU would be able to focus on Fenroy more, so rushing would be less effective? Maybe 40 is the right amount and maybe not.
So I am comfortable having very different results from the implied Sagarin spread in the situation that Desormeaux starts for ULL AND IS 100%. I am not comfortable assuming this, so I have to hedge. Planning for the worst case that McGuire plays the whole game, I am going to assume that 40 or less is the right number.
So to sum up, with Desormeaux, I put my spread at 13-15 and with McGuire, I put it at 8-10. Clearly, the actual, bettable spread available of 7.5 is based on McGuire playing (or Desormeaux being less than 100%). This makes it a pass for me unless I find out that Desormeaux is 100%. As far as the total goes, since I?m coming up with 68+ points regardless of who starts, I think I have a good bet on Over 62.
For the folks with more idle time,
ULL has an incredible running game with QB Desormeaux and RB Fenroy being about the most productive tandem in the country. Because of this, the passing game clicks right along and has provided at least 1 TD each game this year. The FIU defense they?ll be up against is quite similar to the Sun Belt defenses they have already faced this year, particularly Arkansas State?s. The big question is whether Desormeaux will play (he sprained his knee a couple of weeks ago). He missed the Arkansas St game, but the coach is hopeful he?ll be back and he?s listed as Probable (they are coming off of a bye week). Arkansas State?s passing and rushing defenses aren?t much better than FIU?s, so while Desormeaux would be more productive assuming he was healthy, back-up QB McGuire should be able to turn in a performance similar to his Arkansas State game if called upon. With Desormeaux running things, I?d predict about 63 plays and around 600 yards of offense. That equates to about 45 points for ULL this year. If McGuire plays, or if Desormeaux really can?t run, we?re probably looking at about 68 plays for around 500 yards and about 40 points.
On the other side, FIU hasn?t really faced any conference defenses like ULL?s, and I don?t mean that in a good way. ULL isn?t very good at stopping the pass or the run. That said, FIU has shown a tendency to run if they can, regardless of whether passing would be easier. Of course, when the need to catch up is there, they turn to the pass and do a pretty good job of it. I suspect that they will be able to run very well on ULL, but that scoring pressure will cause them to move more and more to the pass over the course of the game. If they perform similarly to how they have done in conference games this year, I would think they?ll have about 70 plays, with around 40 coming in the second half. This leads to roughly 130 yards rushing and 260 yards passing, with 4 TDs and an interception ? say about 30 points.
This leads to a direct calculation of the expected spread of 10-15 points.
The Sagarin Predictors for ULL and FIU are 65.23 and 64.47 respectively for a predicted spread of 1. Throwing in the generic Sagarin HFA of 2.68, we get a spread of ULL -4. Since this is very different from my direct calculation, I tried to figure out what is happening. At first glance, it seems that FIU?s predictor is too low, if anything, as they opened with 4(ish) strong out-of-conference games at Kansas, Iowa South Florida and Toledo while ULL opened with a weaker Southern Miss, Illinois, Kent State and Kansas State. This is confirmed by FIU?s Strength of Schedule ranking of 67, as opposed to ULL?s ranking of 121. Since conference play began, both have played a roughly similar mix of teams. Now, I have to admit that I don?t know exactly what is happening in Jeff Sagarin?s Black Box, but I am fairly certain that Predictors weigh recent performance heavier than performance many games back. This only makes sense since the whole point is to predict point spreads as of now. On the other hand, Strength of Schedule counts all opponents equally. Since I do not use the early, out-of-conference games at all in developing my projections for the game, this seems to jibe.
Looking strictly at the last 3 games of ULL and FIU, ULL has averaged scoring 44 points and giving up 29. FIU has averaged scoring 32 and giving up 21. Since their conference schedules have been similar in difficulty (about 4.5 predictor points harder for FIU, but that is all from Troy, to whom they lost by 10), it seems that we can assume that the differences in recent schedule can be largely ignored, giving us Margin of Victory as a proxy for the Predictor. The average MoVs for ULL and FIU are 15 and 11 respectively. Since HFA is built into these numbers already, we come up with the same 4 point difference. Now that I can understand the Sagarin spread a little better, I am going to see what happens when I look at the teams and their conference performance from a slightly different angle.
The correlation between ULL?s total yards and the number of points they give up is nearly perfect. This makes some sense in light of their lack of ability to stop anyone ? they score, and then their opponent usually scores. Fortunately for them, their D is just good enough to get a few more stops than the offense experiences. Based on conference play, using this correlation leads to ULL giving up 31 points, right in line with the earlier projection. The correlation between FIU?s points scored and their opponent?s average points allowed is nearly as good and leads to an estimated FIU score of 32 points. This does have the problem of using average points allowed that includes out-of-conference games of the opponents, but there are only so many calculations that I am willing to do. Based on all of this, I am willing to go with 30-32 as my estimate of FIU?s score.
The problem is that I cannot do the same thing on the ULL side because of the injury to Desormeaux. The points from the Arkansas State game (McGuire at QB) were earned in a completely different way from the games under Desormeaux. Furthermore, Arkansas State was a much better opponent than the conference opponents that Desormeaux faced. Using the very limited number of conference games played this year is already pushing the limits of using the stats as a projection device, so I cannot break them down further and do the analysis separately for each QB. The best I can do is to assume that the offense will perform roughly as before when it has been faced with similar opponents. If Desormeaux plays, I am quite confident they will. He?s put up about 1800 yards of total offense so far this year and is on pace for 1000 yards rushing again. You can count on Fenroy to rack up the rushing yards, so I feel good about the 45 points predicted for this combo. Redshirt freshman McGuire has exactly one game under his belt, so who knows how he?ll perform, but the defense he faced from Arkansas State is better than what he?ll see from FIU. Another encouraging thing about McGuire is that he had to bring them back from behind twice before they finally pulled away at the end. The problem is that he cannot run as well as Desormeaux, so FIU would be able to focus on Fenroy more, so rushing would be less effective? Maybe 40 is the right amount and maybe not.
So I am comfortable having very different results from the implied Sagarin spread in the situation that Desormeaux starts for ULL AND IS 100%. I am not comfortable assuming this, so I have to hedge. Planning for the worst case that McGuire plays the whole game, I am going to assume that 40 or less is the right number.
So to sum up, with Desormeaux, I put my spread at 13-15 and with McGuire, I put it at 8-10. Clearly, the actual, bettable spread available of 7.5 is based on McGuire playing (or Desormeaux being less than 100%). This makes it a pass for me unless I find out that Desormeaux is 100%. As far as the total goes, since I?m coming up with 68+ points regardless of who starts, I think I have a good bet on Over 62.
