The computer started out its season with its best week of the year, turning a profit of $647, mainly won thanks to a huge victory by the Giants over Carolina. The Giants continued to deliver huge profits for the computer, hitting big during Week 1 ($431), Week 9 ($517), and Week 13 ($363). Unfortunately, the computer gave a lot of that back with tough losses in Week 10 (-$718) and Week 15 (-$148). The Week 15 loss against the Eagles (which featured Michael Vick’s astonishing fourth-quarter performance and DeSean Jackson’s heart-wrenching punt return at the end of regulation time) was the worst – it changed a profitable week into a losing week, and precipitated the computer’s limping struggle to the finish line, when it dropped a thousand dollars in just three weeks.
This season’s nemesis turned out to be the Miami Dolphins. Although the Dolphins delivered a few nice victories in the beginning of the season (including $151 in Week 6 and $265 in Week 7), they ultimately cost the computer a ton of money as they repeatedly failed to cover spreads as their season spun into a disastrous tailslide. (-$108 in Week 11, -$900 in Week 13, -$68 in Week 15, $-42 in Week 16, and -$282.22 in Week 17). The New Orleans Saints, Arizona Cardinals, and Minnesota Vikings were also pretty expensive for the computer this season, but were balanced out by solid seasons by the Chicago Bears, the Cincinnati Bengals, and (amazingly) the Washington Redskins.
The percentages for Gamblor’s different tiers were pretty consistent with its historical performance. Although the 90% tier took a big hit late in the season, the curve below shows that it was generally in keeping with theory. The hedge bets were disappointing yet again this year, which makes me glad that I switched the weighting in the beginning of the season, which severely limited the number of hedge bets that actually turned into bets by keeping them below the $5 threshold. It’s possible I’ll retire the hedge model entirely next season, since between Gamblor and Son of Gamblor there tend to be enough bets during any given week. Taken from a long-term perspective, this season actually ended leaving Gamblor essentially at the same place where it began its season. So obviously it’s nothing to rejoice over, but it’s nothing to be concerned about either.
Son of Gamblor, on the other hand, behaved very strangely. It started off the season brilliantly, hitting every single one of its peaks in the first week and maintaining a weighted win percentage of 70% through the first ten weeks of the season. Unfortunately, its remaining weeks were catastrophic, hitting a low point during Week 17 when it hit an abysmal weighted percentage of just 1.7%. In terms of confidence levels, it also delivered very strange results – performing its best at the 80-83% confidence levels and performing very poorly at levels above 90%. The graph below shows its performance compared to its historical performance – and the oddness of this year’s results actually changed the shape of its lifetime curve. Like Gamblor, the season ended up being something of a wash, with its lifetime total finishing pretty much exactly where it started.
Overall I finished the season with a profit of $222.34, off an initial investment of $1000.00. And my any measure, a 22% return is pretty nice. Add an extra $100 of the sportsbook’s sign-up bonus to that, and you’ve got a very worthwhile season of gambling. I’ll be back in a few days with my Superbowl pick for the year, and once again I’ll be tinkering with the model in the off-season to see if I can squeeze a few more percentage points out of the system.
|Week||GAMBLOR||SON OF GAMBLOR||Profit|
|W||L||P||Weighted Wins||Weighted Picks||Weighted Win %||W||L||P||Weighted Wins||Weighted Picks||Weighted Win %|