In case you missed it last week, we'll again have a "this-week" ranking (further explanation here). It's still very basic -- a much more involved ranking is in the works -- but it at least provides a data-based prediction.
2016 Supercross Predictions - 450SX Update going into Week 7 at Arlington
Not hard to guess the current state of affairs: Ryan Dungey is on top, with a growing lead over Ken Roczen at #2. Jason Anderson keeps performing well, and he has crept past Chad Reed. Trey Canard rebounded in his first week back on the track, but missing two weeks has handicapped him enough that he could only jump up two spots in terms of his season-long prediction.
Brock Tickle's injury prognosis improved slightly, combined with a poor result from Christophe Pourcel round out the major changes in the bottom half of the chart.
If we factor in injuries that riders are dealing with (that we're aware of), a couple of them stand out:
**Keep in mind this is more of an indicator that the rider was recently injured than an actual, precise prediction of how that recovery from injury will affect the rider.**
Eli Tomac could slot in below Jason Anderson, depending on how much you think his shoulder injury is still affecting him. For Trey Canard, his hand injury might mean he should be predicted to finish right behind Marvin Musquin rather than right ahead of him. Again, for Canard that depends on whether his injuries are still bothering him -- he rode in the Top 5 at San Diego 2, which is good, however he was unable to hold his position in 3rd for the entire race, which suggests he may not be at 100% yet.
Actually, let's look at last week's predictions first, to see how they fared. I don't expect that much, since they're not the deepest from an analytical perspective. But, they did pretty well. Here's the ranking by "true talent", taking into account the injury adjustment, compared to where they actually finished, along with how many spots they were above or below the prediction:
That top 10 looks remarkably good, if I do say so myself (there are only 19 riders listed because Broc Tickle was not in the race). Chad Reed finished way off his prediction, but the entire rest of the Top 10 was predicted within 2 spots of where they finished. One caveat is that San Diego 2 was a bit of a weird course with very little passing, so I'm not sure it's representative of a typical Supercross race.
On to this week:
We'll just have to wait and see how the predictions fare in Arlington!