Week 8 Lines

The record against the spread is only relevant to gamblers. As a non-gambling college football fan, I’m more interested in how your ratings fared straight-up. I know they miscalculated on the Texas-OU result, but how did they fare otherwise?

It’ll take a little bit but I can backcheck later.

Just FYI, my simplistic model is predicting another toss-up against Missouri. Of course, it doesn’t account for home/away or quality of opponents. And it does not weight recent games more heavily than early games. All of these factors favor Texas, so a properly adjusted model should predict a Texas win.

Forget about the computers. The way I see it, if Mizzou outscores Texas, Mizzou will pull off the big win in Austin. But if Texas outscores Mizzou, look out! - because the Longhorns will stampede their way to another victory.

Huck, is there a correlation between the success of your ratings and the number of games played? In other words, do your ratings peform better earlier in the year and then tail off down the road indicating sharper lines?

That Duke/Miami line just feels dirty. This definitely ain’t 2001.

OU pummels Kansas at home.

I agree on the Penn State line as well. Michigan is going into early hibernation.

Trips -

They stay pretty consistent when you consider the confidence numbers. Obviously later in the year there are very few picks with high confidence values (and lots of those are because of injuries, etc. that the ratings can’t take into account).

The 35-23 record in the low confidence range so far is flukey. Typically those have been in the 52% range. Not enough to make money after the vig. I would normally expect a 30-28 record over 58 outcomes.

Huck,

So what other factors do you take into account in determining which games to actually bet on? Or is this just your own little petri dish a la Vegas that you experiment with for fun?

I just use it as an objective input. I start at the top of the list and then decide if there’s any good reason not to bet that way. I usually stay away from huge lines for lots of reasons, and I also don’t think that the ratings method does that well with huge lines. Obviously true team strength becomes less of an issue in a blowout as the leading team goes into clock mode with backups.

But I don’t bet that often.

Okay, here are some tables with the performances for the ratings so far. The first three tables are for predictive performance beginning once the ratings had a reasonable amount of data. The final table is retrodictive, meaning that it takes the current ratings and applies them back over games already played to see how they perform. The left-hand column, of course, shows the predicted margin of victory for a team. These are not based on point spreads, just wins and losses and includes all college football games, not just D-1A.

10/4 Weekend W L WPCT Overall 249 82 0.752 21+ 87 5 0.946 14+ 48 9 0.842 7+ 60 31 0.659 <7 54 37 0.593 10/11 Weekend W L WPCT Overall 254 78 0.765 21+ 71 6 0.922 14+ 61 8 0.886 7+ 66 22 0.747 <7 56 42 0.571

            


Season Totals
W
L
WPCT

Overall 503 160 0.759 21+ 158 11 0.935 14+ 109 17 0.865 7+ 126 53 0.704 <7 110 79 0.582

            


Retrodictive
W
L
WPCT

Overall 1835 268 0.873 21+ 626 0 1.000 14+ 391 6 0.985 7+ 415 58 0.877 <7 403 204 0.664