Where road racing
meets data.
procycling.tech is a suite of tools for the sport — reading each race, drafting the story, and modelling the likely top ten. Built on machine learning and a newsroom that has covered cycling since 2009.
Data in, story out — and predictions that stay honest about being predictions.
Everything here rests on evidence and on years of covering the sport. The model weighs the odds; the newsroom keeps the judgement.
Five tools, one race-day workflow.
Each tool does one thing well. Together they run from the result sheet to the published story — and the club that ties it together.
CycleFlow
The stage ends — and the story starts writing itself. CycleFlow pulls the results, reads what the reporters on the road saw, and drafts the race report the same evening: three takes, one fact gate, an editor with the final word. Battle-tested through a full Vuelta. Next: the same report, native in six languages.
CyclePredict
Every rider, every stage: the calibrated chance of a top-ten finish — built from 72 signals across form, terrain, race history and the GC battle. Frozen before the start, scored against the result, reasoning shown rather than hidden. 28 stages on the record, the Tour de France among them. Next: teaching it breakaways.
TopTen
A full fantasy game on top of the model. Ten riders and a reserve under a budget — favourites cost the most and pay the least, so the glory is in the riders the model doubts. Lock before the start, then meet Allan, our resident analyst, playing the same budget. Played all summer, from the Tour to the Vuelta.
procyclingAI
The craft under the prose: cycling's own vocabulary, a language editor, a hallucination judge and a fact contract no sentence may contradict. It's why the output reads like cycling, not like a template — Norwegian first, because it's the hardest.
PCC — Procycling Club
Where it all comes together. TopTen already wears the Club colours; next come leagues against your friends, full seasons and history — with the predictions and the deeper coverage in one place. The inside line on the sport.
Better with numbers than with legs.
Allan Iversen is procycling's analyst — virtual, and the first to say so. Dry-witted, calm, fluent in probabilities. He turns the model's numbers into forecasts you can argue with, and when a call goes wrong he doesn't bury it: he explains it.
“The result judges the machine. The reasoning judges me.”
It reads the climb before the riders do.
Every stage is broken down into what actually shapes a result — form, terrain, history and how races tend to play out. CyclePredict turns that into calibrated top-ten probabilities, and shows why each rider is there. A prediction, never a promise.
No more forgettable
cycling sites. Build the race.
procycling.tech brings the data, drafts the story and models the finish — so the newsroom can do what it does best. Built on AI, engineering, and a newsroom that has covered the sport since 2009.