So a few weeks back I stood in front of a camera and told you my Garmin was wrong about me. Then, because I couldn’t leave it alone, I spent a week going through thirteen years of my own race results, over a thousand races, trying to actually prove it. I couldn’t. This video shows you exactly where my own argument fell apart, and it’s more interesting than I expected.

The idea was simple enough. Fit an exponent using only my shorter races, 5K up to 10 miles, hold the marathon back, and see if it predicts what I actually ran over the distance. It doesn’t. In 2022 and 2024 my short race exponent was identical to within 0.0009, as close as makes no difference, and yet my marathons in those years were 4 minutes 48 seconds apart. Same engine, completely different result.

Dig into it further and the pattern’s clearer still. Between 2017 and 2024 my 5K time moved by 4.6%. My marathon moved by 9.2%. If the predictor worked the way I wanted it to, those two numbers should move together. They don’t, and the gap tells you the marathon isn’t about the engine at all, it’s about the block of training you put in to get there.

I go through all of this properly in the video, the maths, the years without a PB, and the test that finally killed the idea I was chasing. It’s not a comfortable conclusion for someone who likes to think he can out-argue his own watch, but it’s an honest one.

Valencia is on the 6th of December. My watch says 2:30. I think I’ll beat it, and after everything above, I genuinely can’t prove that I will. Have a watch of the full video, it’s a good one if you’ve ever argued with your own predictor.

Just keep running.

---

New videos go up on the channel regularly. Subscribe so you don’t miss the next one.