Can We Make Trail Races Easier to Follow?
Testing our Race Intelligence Layer at the Beast of Big Creek [HELP!]
First, huge shout out to Matt Walsh + Mathias Eichler for having conversations about what we’re building. I can’t recommend Trail Mix enough if you’re into this sort of stuff.
This weekend, Beast of Big Creek marks the return of the Skyrunner World Series to the United States. We’ll be testing what we’ve been building (in public) at Borderlands: Live on Course.
It’s an initial test to ensure we’re not overfit on synthetic observations from Cocodona 250.
We’re trying to answer a much more fundamental question: How do humans naturally describe what’s happening during a race and we need help from real people ein the field.
We’ll invite volunteers [runners, crew members, spectators, photographers, volunteers, anyone already at the event] to send short qualitative observations throughout the day whenever they want using what every language they want.
» Will you be there? Sign up here.
The goal is to see how a range of observers varying in quality take in and explain whats happening. In order for this to work we can’t rely on human nodes in the field with expert insights. We’d be too limited. We want anyone on course to be able to contribute.
We’re trying to build a corpus of real race observations and learn how people describe momentum, fatigue, atmosphere, weather, tactics, emotion, and the thousands of details that make a race legible to someone who isn’t standing there.
Eventually, LoC’s job will be to transform hundreds of these small observations into meaningful race intelligence in real time for video and audio commentaries.
We’re going to learn what people notice, what they ignore, how they communicate, and what moments feel noteworthy to them.
Our success metric on this is simply 5 people in the field making observations as often as they can. It’s a low bar. Every ensuing event will require more of the humans in the field.