EliotJackson

@mtbikemccoy: As you use the brakes, residue from the pads, dirt and other stuff build up on the rotors (and pads). When you put new pads/rotors on you have to go through the break-in process all over again and the brakes, inevitably, don't ever break in the same way twice. When you're at a world cup and you only have a few runs, it sucks to have to use one of them to get the brakes bedded in/adjusted to where they should be and no matter how much you ride through the car park with them on, it really takes a couple of turns for them to really feel right. By using the heat gun to get rid of residue (which is why you would usually change pads st a world cup), John is cutting this process out.

@scott-townes: " Pushing hard with your front leg while unweighting the rear leg is the most effective way to pedal." I ummm. I think uhhhh.... Well I mean.... Are you? hmmm.

@CaptainBash: Yeah, for sure! And that's the correct way! I was just saying that maybe we shouldn't even try to prove anything if it doesn't look like it would be correct. Just torturing the data at that point and we are really likely to overfit.

@joel4: It does for sure! We just need enough data for those scores to be meaningful, which we don't really have :)

@CaptainBash: Yeah extracting the data from the PDFs is extremely annoying haha. I understand what you are saying, and that would be totally valid if we were trying to answer the question of did the times got tighter. But we’re not. We’re trying to answer if it is more difficult to get a top 10 or not. If we did an analysis and we said that from one season to the next the times get tighter by 1.3 seconds with a p valid of 0.04. That doesn’t tell us anything about it getting more difficult. If the tracks are all really difficult and the we go and race on a flat road, the times will all be within .5 all the way back to 50th place. Like I said, 1 second at one track is not equal to one second at another track. I.e making up 2 seconds at the top of fort William is waaaaaaay harder than making up 2 seconds at the second split at Val do sole. The other thing that A simple averageing doesn’t take into account is that you’re not measuring the same phenomenon. It would be like doing an analysis on Olympic sprint times and measuring 100m one year and 50m the next. It’s just a bit more complicated than it looks at first glance :). I hate it when people use p values and statistical techniques to make an article look more scientific or credible when the assumptions they used for those test make the test invalid anyway. That’s why I didn’t do any of that.

The original question is it getting harder. Not are the times getting tighter. If all the tracks are getting easier then the times will be closer but the difficulty of getting those times will be the same. You can't compare from season to season because, as you said, everything is different. Closer times =/= harder to get results. It's a lot easier to be within 3 seconds at the top of ft william than it is to be within 3 seconds in the second sector at val di sole.

@dbc3: For sure!

We thought of that but then you would have to control for the elevation of the track affecting tire pressure and the riders ability to breathe :P

Yep.

That's why we didn't provide P values :) Because you can't make any conclusion based on the data. We have 179 data points. Even just using the variables you mentioned you are already overfitting. The box plots you're talking about with confidence intervals would be useless once you move to the next season because the tracks have changed, the riders have changed, the conditions have changed. You aren't measuring the same phenomenon, which means P values don't apply, or any conclusions for that matter.