Discussion: Skip analysis and the 2007 Billygoat

skip analysis

Nice skip analysis. It confirms my gut feeling that #18 was the one to skip.

i thought so too but then being in the lead I found that 20 is a better choice. looking back..probably I will still choose 20 based on tactics

I realized this morning that if a competitor missed a split time before or after the skip, the same analysis could apply, so I was able to use more data and analyze more skips (including #20). When I did that, skipping #2 became relatively better. Just goes to show how sensitive the analysis is because there isn't a lot of data. Still, I think it gives some idea of how the skip decisions turned out in practice.

(The link above now brings you to the revised analysis.)

I've been staring and staring at the Billygoat course and just cannot understand the rationale behind skipping 2. Can someone explain that please??

simple navigation, no hesitations, no crossing reentrants, less climb, smooth running. Check PG's route.

in theory it should be weighted even better as the difference should be lower early on in the course when people are fresher.

one year Hammer skipped first -just to be alone ( maybe) and he eventually won ( I was behind 110 sec and did not seen him the whole course ).it was a good skip but why would you do it so early in the course ?? hammer knows.

It's conceivable that the analysis for #2 turned out better than it should because the 2 people who skipped might have had speed-on -trails/speed-in-woods ratio higher than that for other competitors.

Still, I think it's a pretty darn good skip.

Here's an interesting story backing what Neil said about simple navigation:
When I got to the first trail junction in the woods (maybe a hundred meters from the control), other faster runners were arriving, but they seemed confused as to where they were and it took them a little time to figure out to follow Brendan (who had met up with me in the field) & me up the right trail.

Dean, did you take climb into account at all?

Not amount of climb, but quality of it -- I remembered a control on the #2 slope that I had physical difficulty climbing to in a previous race.

There were some people in my vicinity who boomed #2 (specifically, they were ahead of us, we took a different route, and they arrived after us). If those people are among those who posted splits, that makes going to #2 look worse, and therefore skipping it better (related to what Dean says about people confused by the trails). There was also an injury in the vicinity (stick in the ear?), and that may have delayed some people by a bit.

I don't understand the math, but could someone who does please come up with a way to measure the statistical significance of it?

p in ch(°f) s^a L(t)

I think Neil is overstating the significance (or perhaps he's using British units?) Here in the US I'd say it should be taken to be:

gr^a in(°f) s^a L(t)

My Father's Day greeting from son Kevin (3 Billygoat shirts) included a skip analysis.

That's hilarious!