Showing posts with label Accuracy Index. Show all posts
Showing posts with label Accuracy Index. Show all posts

Sunday, September 6, 2009

Michigan and Notre Dame at the top of the Rankings

Week one of the college football season was full of surprises, including the Wolverines and the Irish taking an early lead in the inaugural Prognostication Index. With his shrewed pick of Alabama over Virgina Tech, Desmond Howard jumps out in front with a perfect rating. Granted, he only made one pick, so Notsodamus needs a bigger sample to see if this holds up. Of announcers with at least four gradable picks, Lou Holtz is the leader with a 0.75 PI. Lou accurately predicted wins for the Irish, the OSU Cowboys, and the Tide. His only blemish was his pick of OU over BYU. Out of predictions recorded by Notsodamus, no-one picked that game correctly.

Salsa Boy Kirk Herbstreit got off to a rocky start, going 1-6 with a 0.58 PI. His only correct selection was Notre Dame over Nevada, which was picked correctly by all announcers except Mark May. Herbstreit was also negatively impacted by not picking the Alabama/Virginia Tech game as he called the game later that night. This game was picked correctly by everyone. Salsa lovers and Sooner fans hope that their horses can recover from a terrible opening weekend to be contenders.


There were no observed pick changes this week, but Notsodamus didn't expect there to be in the opening week. The talking heads just made their picks for the year, and with only one week of football in the books, most picks are still solid. Notsodamus wonders what Mel Kiper will do, as he picked Oklahoma for the BCS National Championship.

For detail on the picks recorded during the week, click here.

Click here to see those picks recorded by Notsodamus which do not have a result yet. Florida, Texas, Oklahoma and Notre Dame have all been picked as BCS National Champions.

Wednesday, September 2, 2009

Notsodamus Is Here

Picture this. You're watching the Texas/Oklahoma game this year on ABC. In the first quarter, Brent Musburger announces that for sure Sam Bradford will win the Heisman this year. Two drives later, after a 50 yard touchdown pass to Jordan Shipley, he claims that with 100% certainty, Colt McCoy will win the Heisman. This plays out three more times throughout the game, flipping his pick from Bradford to McCoy after each game changing play.

I'm on a mission to rank announcers based on how Accurate they are in their picks, and how much Conviction they have in their predictions. That's why I've created the Prognostication Index. The Prognostication Index is composed of the average of the Accuracy Index and the Conviction Index. The Accuracy Index is simply whether the announcer's latest pick was correct. It is either 1 or 0 based on whether the pick was correct. The Conviction Index is the number of times an announcer changes their pick. If Musburger says that Bradford is going to win the Heisman, then McCoy, then Bradford, then McCoy, and then Bradford again, his Conviction Index is .2 (1 category over 5 selections). Assuming McCoy wins the Heisman, his Prognostication Index would be .1 ((0 + .2)/2).

I'd like to rank all announcers based on this criteria, but I need your help. I can't capture all the data I need by myself. I need you to click on the "Submit Prognostication Now" link and fill out the form for any predictions you hear from any announcers. I'll need to know who submitted the link (just create a handle for yourself), the date, the time (so I can sort through multiple predictions during a game), the announcer, the network, the category, and the pick. I'll fill in the winner. A category is basically any type of pick, whether it be national champion, Heisman winner, conference champion, or who will play in the title game.

Thanks for all your help getting this up and running. I'll be making adjustments along the way as I see how the data comes in. We won't know the answers for some of these things until the end of the season, but I'll update statistics on "gradable" predictions throughout the year. I hope you come along for the ride.