While I already run down Luigi’s performance against both Bill Simmons and Bodog each week, I also get to read about a contest in ESPN’s insider section: the Hilton SuperContest. In short, people buy in and make 5 picks a week against Hilton spreads, earning a point for wins and half a point for pushes. At the end of the year the winner gets a big payout, a certain number of people are in the money, and there’s a bonus for being over 67%. I thought it might be a good opportunity to test out a piece of gambling advice I’ve heard in other places: the more you think the spread is wrong, the more you should bet.
The idea makes sense: if the Lions are only three point favorites but you think they’ll win big, shouldn’t you be more confident that they’ll cover? And if you’re more confident, shouldn’t you bet more? If you thought the Lions would only win by maybe 4 or 5, that’s close to the 3 and there’s some chance they won’t come through. You might not want to risk as much there. I already do this to a certain extent, advising that you just not put any money down if the prediction and the spread are within a point. I also looked at the results for my model, and there’s at least a trend for it to be more accurate against the spread when its prediction is farther away from the spread. So over the course of the year, maybe my fake SuperContest picks will be better than all of my picks.
Let’s travel back in time. I don’t make predictions for week 1 because there’s no data to go off of, but I do predict winners based on the previous season. So I’ll compare those probabilities to the lines that Bill Simmons posted to try and choose my five games. The Bears were underdogs but I had them as 63% to beat Atlanta, so that’s in (obviously any point difference prediction I made would have gone with the Bears, and been far away from them getting three points). Similarly the Rams were supposed to win, and the Redskins were home underdogs. I had the Raiders as a road favorite but the Broncos got the standard home -3, so that game is in. The last game is tough to choose, but I had Baltimore as pretty strong favorites but they were only -1.5 against the Steelers, so I’ll go with that game (alternative were Tampa but I wouldn’t pick them against my hometown Lions, or the Jets over Dallas but I dislike the Jets). So going with Chicago, St. Louis, Washington, Oakland, and Baltimore I would have gone 4-1. That’s a pretty good start.
The subsequent weeks work the same way but I have actual point predictions to use for picking the top five games, and Bodog’s lines to pick against. In week 2 I would have gone with Buffalo, San Fran, Detroit, Chicago, and Houston. Those picks went 2-2-1 for a total of 6-3-1. Week 3 would be Houston, Arizona, Buffalo, Indianapolis, and Detroit. Those picks went 2-3 for a total of 8-6-1. Week 4 would be Carolina, Washington, Buffalo, Houston, and San Fran. Those picks went 4-1 for a total of 12-7-1. And finally this past week the top 5 games were Buffalo, Cincinnati, Carolina, Tennessee, and Detroit. Those picks went 4-1 for a total of 16-8-1.
Since the Bodog lines are different from the Hilton lines (or certainly Bill Simmons’, and he might get his from the SuperContest), it’s possible my results aren’t the same as what would happen were I actually in the contest. But since I did better against Bill’s lines than Bodog’s last year, I’ll assume they’re at least accurate and possibly biased against me. So we’ll say that I stand at 16-8-1 after five weeks, which is good for 16.5 points. I would be tied for 15th right now and just below the bonus threshold at 64% accuracy (unless they drop pushes or go by points, in which case I would be right there). The last two years the winner has gotten about 55 points, so I’m on pace to be about there. Of course, the guy with 20 points who’s actually winning right now is obviously on pace for better. But the top 20 players get some money, so I would still be in that range at least.
The current result is a little better than Luigi’s pickable games overall, but probably not significantly better. I’ll keep looking throughout the season to see if using this pick confidence is helpful.