For Once, the Fastest Rider Didn't Win
VOL. 01 // FIELD NOTES · POST TWO
At Loudenvielle the order on paper and the order on the day came apart. The winner was only fifth-fastest across the weekend, and the two quickest men finished third and twenty-third. Everywhere else this season, the paper order has held.
Jordan Williams was the fastest rider on the mountain at Loudenvielle. He topped the opening qualifying session. Add up the quickest version of every sector he rode across the weekend, timed training included, and his best possible lap comes to 3:27.039, faster than anything anyone else showed all week. He finished third.
Loris Vergier's best possible lap comes to 3:27.043. Four thousandths slower, across three and a half minutes of racing. Vergier finished twenty-third, 12.2 seconds down.
The rider who actually won, Luca Shaw, had the fifth-fastest weekend on paper.
Every weekend, the sport produces more speed than it pays out.
The measure here is the composite lap. Take every timed run a rider completes across a race weekend, training runs, qualifying, the final, and keep their fastest version of each of the five sectors. The sum is the lap they proved they could ride. Set it against the lap they actually delivered on Sunday, and the gap between the two is a number the results sheet never shows: pace owned but not collected.
Race analysis usually quotes a different number, the perfect time: the fastest version of every sector by anyone in the field, stitched into one impossible lap. Useful, and entirely theoretical; nobody gets to ride the field's lap. The composite is the same idea made practical. Every sector in it was ridden by the same rider, on the same bike, on the same weekend. It is not what the race could have been. It is what that rider had already done.
Shaw collected nearly all of his. His composite is 3:27.515 and he raced 3:27.637, a difference of 0.122, all of it in one sector. Four of the five sectors of his perfect weekend lap were set in the winning run itself. Benoit Coulanges, second on the day, delivered to within 0.31 of his paper lap.
The podium, with one exception, was a list of riders who arrived at the final with nothing left to prove and proved it anyway.
The exception is Williams, and his case is the strange one. Only one sector of his composite comes from his race run. The other four were set on other days, in training and qualifying. In the final he gave back 0.3 to himself in the opening sector, 0.7 in the second, 0.1 in the third, 0.1 in the last. There is no crash in that list, no moment a camera would replay. He was simply a tenth or three slower than his own benchmark almost everywhere, and the increments add to 1.2 seconds. That is the full distance from third to the top step, paid in instalments too small to see.
Vergier's run is the same loss told the other way round. His race lap contains two personal-best sectors, set at full racing commitment, and one sector twelve seconds slower than his own benchmark. The timing sheet does not say what happened there, only where. He lost the race in one place. Williams lost it in no place at all.
It would be tidy to conclude that paper pace is worthless. The women's race, run on the same hill the same afternoon, refuses the conclusion. Vali Höll was the fastest woman on paper and won. Her composite lap and her winning run are the same lap: all five of her fastest sectors of the weekend came in the final, borrowed from nowhere. She has now done it at both rounds of 2026, fastest on paper, nothing borrowed, and the win, at Yongpyong and again at Loudenvielle.
She is comfortably the most decorated of the three riders in this story, 25 career World Cup wins and 45 podiums to Williams's 5 and 15 and Vergier's 10 and 33, and she did this while leading the 2026 standings in defence of her overall title.
And round one says she is the rule, not the exception. At Yongpyong the men's paper top five and the day's top five were the same five riders, in near-identical order, with the winner, Asa Vermette, borrowing nothing. Seven of the women's eight fastest composites that weekend were delivered whole in the final. No winner this season, men or women, has borrowed more than 12% of their best lap. The chaos in this post is not how downhill normally works. That is what makes Loudenvielle's men worth the post.
The field-level numbers say how rare Höll's afternoon was. The theoretical perfect time for the elite men, the fastest sector by anyone in any run, comes to 3:23.884, owned by five different riders and 3.8 seconds quicker than the run that won. In the women's race the same margin is 1.0 second, and Höll personally owns two of the five field-best sectors.
Five men left Loudenvielle having already ridden, sector by sector, a lap faster than the time that won. Only Shaw collected it. The chart below is what that did to the order: the paper top ten on the left, the day's top ten on the right. Six ribbons hold. Four fall out of the ten entirely, and the four places they leave behind are claimed from outside the paper ten by riders who had shown less and delivered more.
The result sheet at the bottom of a World Cup hill is not a ranking of speed. It is a ranking of speed delivered, in one run, at the one moment it counts. In every other race this season the two rankings have agreed. Shaw won with the fifth-fastest lap of the weekend, delivered whole. Höll won with the fastest, delivered whole. Williams and Vergier carried the most pace down the mountain and left it spread across a timing sheet nobody reads.
Gravity isn't the obstacle. It's the input. Pace is the same: showing it is the entry fee. Delivering it is the race.
How these posts are written
Gravity Lab is a personal project, written by me, Sam Cave-Penney, with generative AI. Every post is built in two halves.
The human half
I choose the subjects, do the research, pull the race data from the source and run the analysis behind each piece. The composite-lap numbers in this post come from my own pipeline over the official Chronorace timing for both 2026 rounds: timed training, qualifying and finals. No timed-training data was published for Yongpyong, so round 1 composites use qualifying and finals only. I make the final call on every paragraph.
The AI half
Generative AI drafts the prose, working from that analysis and source material. The split exists for two honest reasons: AI writes better sentences than I do, and it is what makes a series possible rather than a one-off.
Nothing here pretends to be something it isn't. The analysis, structure and editorial judgement are mine; the prose is collaborative. Rider career records are taken from UCI World Series profiles and combine junior and elite results.