[stupac2]

I mentioned somewhere (we have so many similar threads that I forget where things are happening) that ESPN has us as an underdog against USC this weekend, which feels quite wrong. This is based on FPI, which makes generating spreads really easy since it's just "points better than an average team on a neutral field", so it's just subtracting the two, then adding/subtracting 3 for home (the actual home field advantage FPI uses isn't stated anywhere, but it's usually 2-3 points and I most commonly see 3).

So I decided to see if FPI has consistently overrated them and/or underrated us. I went through every week of data to generate a spread for the upcoming game (so I used preseason FPI for week 1's game, week 1 FPI for week 2's game, etc) and corrected for home/away. And (surprise!), it has! Over the whole season FPI has, on average, predicted that USC would be 3.6 points better than they have been, and it's predicted that we'd be 3.6 points worse than we have been. Right now it predicts us as being 0.8 points underdogs, so if you add the biases back in, we "ought" to be more like 6.5 point favorites. That's closer to Vegas spread (us by 4) than the standard line.

Interestingly, USC has only outperformed their FPI prediction in two games since Helton took over, and I bet you can guess what they are. (Pausing for you to guess.) Utah and UCLA! Turnovers are worth ~4 points, so if we factor those into those games, only in the UCLA game has USC outperformed the spread, and even then only by a point. (We don't gain many turnovers so the only game we have that turns from positive to negative because of turnovers is UCF.)



The takeaway here is what I've been finding every time I look into USC's stats: they're overrated. As long as we hold onto the ball we should beat them.

[stupac2]

I mentioned somewhere (we have so many similar threads that I forget where things are happening) that ESPN has us as an underdog against USC this weekend, which feels quite wrong. This is based on FPI, which makes generating spreads really easy since it's just "points better than an average team on a neutral field", so it's just subtracting the two, then adding/subtracting 3 for home (the actual home field advantage FPI uses isn't stated anywhere, but it's usually 2-3 points and I most commonly see 3).

So I decided to see if FPI has consistently overrated them and/or underrated us. I went through every week of data to generate a spread for the upcoming game (so I used preseason FPI for week 1's game, week 1 FPI for week 2's game, etc) and corrected for home/away. And (surprise!), it has! Over the whole season FPI has, on average, predicted that USC would be 3.6 points better than they have been, and it's predicted that we'd be 3.6 points worse than we have been. Right now it predicts us as being 0.8 points underdogs, so if you add the biases back in, we "ought" to be more like 6.5 point favorites. That's closer to Vegas spread (us by 4) than the standard line.

Interestingly, USC has only outperformed their FPI prediction in two games since Helton took over, and I bet you can guess what they are. (Pausing for you to guess.) Utah and UCLA! Turnovers are worth ~4 points, so if we factor those into those games, only in the UCLA game has USC outperformed the spread, and even then only by a point. (We don't gain many turnovers so the only game we have that turns from positive to negative because of turnovers is UCF.)



The takeaway here is what I've been finding every time I look into USC's stats: they're overrated. As long as we hold onto the ball we should beat them.