How did the "This-Week" predictions do for Week 11 in Detroit? That depends: did Ryan Dungey really commit an infraction worth a two-spot penalty? And should Justin Brayton and Blake Baggett take a hit on their performance because of what seemed to be totally random flat tires? Let's look at it both ways. First, the official results, comparing the True-talent prediction including the injury adjustment:
Short Track: Detroit "This-Week" Predictions Review
Since Eli Tomac and Ken Roczen finished 5th and 6th--because of terrible starts--a lot of the results at the top are going to be thrown off. Add in that Cole Seely went off the track in Lap 1, and Brayton + Baggett getting sidelined by the flat tires, then you've got some big deviations in the middle, as well.
Chad Reed was predicted all the way down at 8th, but we saw a better Reed at Detroit than in several of the last few weeks.
Josh Grant did about as projected, which is actually a bit surprising considering there wasn't much to go on for his prediction.
Mike Alessi had the biggest positive variance, finishing 9th compared to his prediction of 15th.
Now, let's look at what really happened. For argument's sake, we'll move Dungey back to #1, and we'll put Brayton and Baggett back at the spots they held right before their flat tires -- 4th for Brayton and 7th for Baggett. That may not have been their final finishing positions, especially because Brayton was being chased by Reed, Tomac, and Roczen, but we'll just assume that's where they would have finished:
That looks quite a bit better, so let's go with that ;) Roczen and Tomac are still pretty far off, but Dungey, Marvin Musquin, and Jason Anderson are about right. Trey Canard had a crash in Lap 12, which left him 8th overall at the end of the race -- he's in 10th in this hypothetical because Brayton and Baggett jump ahead of him. In this scenario, Brayton and Baggett performed better than their original projections.
Josh Grant's still pretty close to his prediction, considering. Weston' Peick's 7th doesn't look so good when it gets adjusted to 9th, but still ahead of his projection. And now Justin Bogle, Mike Alessi, and Vince Friese are pretty close to their predictions.
As for the Location adjustment, we included 4 adjustments for riders that had enough history at Ford Field in Detroit that was different than their average performance away from Detroit:
Still a work in progress, the adjustment turned out to be more or less correct for Seely and Brayton -- both did worse than their projections. Seely's location adjustment expected him to be 0.9 spots worse than his normal prediction, and he finished 9 spots worse (because of his very costly crash in Lap 1). But, Brayton only was worse because of the flat tire. Peick had a much better day than predicted (5 spots better), which was very different from the location adjustment (expected 1.2 spots worse). Josh Grant finished one spot worse than his prediction, but it's hard to know what to make of that -- finishing 12th despite it being only his second race of the season is actually pretty good, so maybe that matches the location adjustment after all. Overall a bit inconclusive as for the Location adjustment's relevance and predictive ability, at least from what we saw in Detroit.
So that's the wrap on predictions for Detroit. In Santa Clara, I think the picture gets cloudier as Musquin and Anderson--who each have fairly short 450SX career data to rely on--start to mix in with Tomac and even Roczen for #2 & #3. But we'll see.