What the model uses
V1 can translate one completed-race result or a saved history of up to five results with usable course facts. Each usable result is translated independently using distance, ascent, broad terrain, technicality, race format, navigation, support and expected night exposure. In saved-history mode, course relevance and evidence quality determine influence; materially discordant evidence can be capped, and disagreement widens the range. In either mode, lower comparability or weaker evidence keeps the estimate deliberately cautious.
What the model cannot know
Training, health, injury, fatigue, heat, cold, wind, ground conditions, nutrition, hydration, sleep, equipment, navigation mistakes, checkpoint queues and course changes can all move the real result outside the estimate.
Cutoff language
Comfortable, challenging, tight and cutoff danger compare the modelled range with the stored organiser cutoff. When an organiser explicitly publishes no fixed overall limit, the predictor labels the cutoff band “No fixed limit” and marks its buffer and cutoff outlook as not applicable. An unresolved cutoff remains “Unknown”. These labels do not measure safety, fitness or an individual probability of success.
Use it as one input
Use the estimate to ask better pacing and preparation questions. Do not use it alone to decide whether an event is safe or appropriate. Follow organiser requirements and seek qualified medical or coaching advice where relevant.