Gnar Weasels Promotion Report/Strava Segment Foolishness

So last year we ("Weasel Enterprises, LLC"*) put on a mountain bike race and it went GREAT, which means no one got an ambulance ride and we didn't lose money.  We attempted to repeat that feat this year.  Here is a collection of semi-random notes from race promoting.

In case you haven't noticed, mountain bike race scheduling is STUPID, and we're part of the problem.  There's only so many Sundays in the year, and there's not a single open Sunday in the months of May, June, or July... and maybe even August, although I'm too lazy to check right now.

Oh, but Colin, why don't you just put your race on a Saturday?  It works for road racing and cyclocross!

Oh, don't I wish.  Unfortunately, our #1 obstacle to successful race promotion is parking, and the motorcycle dealer who graciously lets us park at their place only lets us do it on a Sunday.  So if other promoters with more flexible scheduling want to move to a Saturday...that would be great.  But we can't.

Adding to this issue is the fact that Thom spends most weekends of his summer driving all over the country covering 100-milers and number of available dates we have to put the race on is something like four Sundays in the entire spring/summer.  So we basically look at those four Sundays and try to decide which will be the least horrible choice for New England bike racing.

This year we ended up on Memorial Day weekend, and initially I was like "that's perfect and tons more people will come,"  but as time progressed it became apparent that we had conflicts with a Canada Cup race, the Transylvania Epic, and Coyote Hill (Root 66) in Vermont.  And it was a holiday weekend, which for some people means "going places" and "doing things," despite the fact that Foxborough is a place (I think) and racing bikes is definitely a thing.

Aaaaanyway our prereg numbers were not looking amazing at any point and I was forced to use my promoter balls of steel to calm down Thom at one point.

("Balls of steel" means I sent an email that said "if we lose money I'll take the hit, but if you post a single whiny tweet about prereg numbers I will drive to Boston and blow up your minature van")

In the end, 200 people showed up to shred some Gnar and it was great.  I'd love to fix our Canada Cup and TSE conflicts for next year because I like the idea of having the most competitive non-UCI  mtb race in New England, but really as long as we aren't losing money and people are smiling it works for me.

The Enduro Segment

Last year's course saw some minor revisions, removing some trails that may not have actually been legal to ride (um) and shortening the lap a bit.  On a scouting mission back in April with DA LOCALS (Rich Pirro and Shawn Mottram), we rode down the Rock Ridge Trail and it was AWESOME.  It was also SKETCHY, but I was on a long travel bike and I've been racing mountain bikes for 23 years (ask me about the 1991 Mt Ste Anne kids race, baby) so I started lobbying Thom to put it in the course.

He will probably deny this and/or have a case of selective memory, but let the record show that he initially thought it was not a safe idea and it was me, goshdarnit, that had faith that the expert riders of New England would both like and not die on that section of trail.

Actually, I think I said everyone should ride it, but after Kevin (holder of one USACycling Pro MTB license) struggled on his first run at it with a hardtail, we realized that it might be slightly unreasonable to have beginner class riders hitting it blind.

It ended up working out better than any of us could have imagined, as no one went to the hospital (for the second year!!) and having a legit experts-only section was wicked fun.  At least for us.  And the only thing more essential to a race's success than racers having fun is the promoters having fun.

Adam Szymkowic, "not lacking in commitment" according to spectator commentary.  He also may have clipped a tree seconds after this photo.  [ Tam Lynn ]

Alex Grabau claims he didn't go over the bars here.  [ Tam Lynn ]'s outtakes from the final rock obstacle.  Looks like about a 50% success rate for lap one of the expert/elite race, which is just about what I'd want it to be.

Strava vs. Enduro Segment

Thom is an unapologetic Strava junkie (unlike me, who is at least embarrassed that I pay attention to Strava sometimes) so in no time flat he had the Rock Ridge Trail and its approach made into a Strava segment, and a sponsor lined up (Mavic) to give a sweet jersey/baggies combo to whatever man/woman could ride it the fastest in the race.  

But first, we went out and tested it, because as washed-up elite racers all Thom and I can do competently these days is pin it downhill on a trail for a few minutes at a time, and then spend ten minutes talking about it while catching our breath.

On the final course marking day, Thom somehow knocked an amazing THIRTY SECONDS off my record time on the segment, a feat which I refused to believe.  So when I should have been printing reg forms and assigning bibs, I was instead staring at Strava trying to come up with an explanation for why I had gotten so badly schooled.

And that's when I found it.  The enduro segment has a small buttonhook at the start, so after about 25 seconds of riding, you're pretty close to the start of the segment.  Throw in a little GPS drift, and Strava will occasionally decide you're basically "back at the start" of the segment, and restart the timer at zero.

So then we had to delete and recreate the segment the night before the race trying to make a version of it that wasn't susceptible to this "buttonhook" problem.  We only had about 8 test GPS tracks to work with, so when I got something that matched all 8 runs, I figured (prayed) it was good enough.  Remember this was the night before the race, so "good enough" is a sliding scale that is lowering with every passing moment...

...and then race day happened, and BOTH of our fastest times still had the buttonhook issue in them, because an Iphone is not military-grade GPS.

I documented this with some screenshots, so if you're a big time Strava junkie you should probably know these things next time you're losing sleep over a 1-second KOM miss.

Figure 1:  Raw Segments from the top four times on the leaderboard

Notice how both the first (Dr Wuss) and the third (Eric Bascome) don't turn blue until AFTER the little loop at the start of the trail, while second (Big Bikes, aka Thom) and fourth (Josh Anderson) turn blue before the loop (as we intended).  So those guys got a free 25 seconds on their runs.

Note especially how disparate Dr Wuss's 4 gps tracks are near the bottom of the enduro segment -- this is how bad Iphone GPS and thick foliage is!

Luckily, the "analysis" button on the left side of the Strava ride page lets you select an arbitrary subsection of a ride, even if it's not a segment, and gives you the stats on it.  So for Dr Wuss and Eric, we can manually attempt to time them on the segment:

 Remember how much drift the good Doctor had at the end of the segment each lap?  Even if I could perfectly select the start/end points of the segment (which I definitely can't), these times are basically only good to within 5 seconds, I would say...

...which is a problem because now we have all 3 guys doing the segment within 5 seconds of each other, and we only have one jersey to give away.


In the end we decided to give it to Josh Anderson, because he's 15 and we like kids on mountain bikes.

Also he was the fastest on the bottom section according to Strava, and that's my favorite part:

Figure 3:  Rock Ridge KOM
Meanwhile on the women's side, we were lucky that the top two women (Karen Potter:  3:56, Stephanie Baker:  4:00) both had legit GPS tracks so we didn't have to play this game with their data, too.

Oh wait, did I say up above that you shouldn't trust Strava on anything less than 5 seconds?  Hmm.  Crap.

I guess next time we have to find a way to time it for real.  Laser timing systemsVolunteers are cheap, right?

Whatever, we still had fun!  See you next year!

* this doesn't exist at all


Anonymous said…
Something similar to your laptop video at the finish might solve the segment issue. Maybe a gopro at the start and another at the end. You could use the time stamps from the video to get everyone's relative time thru the segment. You'd just have to tell people to flash or call out their numbers to the camera as they passed

Popular posts from this blog

A letter to everyone's parents about Coronavirus

Sam Anderson Cheats at Mountain Bike Racing

Do-It-Yourself March Cycling Blog Post