As you could probably tell from our coverage, we’re big fans of the LVH SuperContest (formerly Hilton SuperContest.) Last year, we tasked ourselves with the duty of creating a predictive model that would allow us to skyrocket up the charts so we could take home the $310,000 first place prize. Well, it didn’t happen. Our team (WM Index) finished 48-35-2 with 49 points — just two points out of the money¬†as we tied for 49th place out of 510 contestants.
The good news: we learned quite a bit during the process. ¬†As the season progressed, we made drastic improvements to our model¬†constantly¬†adding new data in and re-testing all weeks as we experimented to find the best mix.¬†The final model that we used finished strong — ¬†going 21-13-1 (61.7%) over the last 6 weeks as we jumped from 27-22-1 (55.1%) in week 11 to finish at 48-35-1 (57.8%.) Yes, that’s a small sample size and it would be unwise to read too much into that but it was solid progress.
We think that we’ve made even more significant improvements over the¬†off-season¬†and have our model primed to place at this year’s contest — and we want to share it with you, our loyal readers. This year, we will be releasing each of our SuperContest plays along with a confidence score for each — early Saturday morning (after the SuperContest deadline) at no charge of course.
Having to choose 5 games every single week can be challenging. ¬†Many weeks last year we only had 2 or 3 plays with strong¬†confidence¬†levels but were forced to turn in the full 5. Additionally, in some weeks we¬†actually¬†had more than 5 plays, but had to crop down our picks to hit that magic number. By opening up our data and showing our¬†confidence¬†score we hope to show everyone just how tough the SuperContest really is.
So every Saturday this year, check the blog to see our list of SuperContest plays.