How do race predictions work?

Race predictions use AI to estimate your finish times across different distances based on your recent training data. ## How it works Our model analyzes your training activities from the **last 90 days** to build a picture of your current fitness level. It looks at factors like pace, distance, elevation, heart rate, and training consistency to generate personalized predictions for common race distances. Because predictions are based on a rolling 90-day window, they adapt as your fitness changes — improving as you train more and adjusting if you take time off. ## Tips to improve accuracy The more representative your training data is, the better your predictions will be. Here are a few tips: - **Mix up your training.** A combination of easy sessions, tempo efforts, intervals, and long workouts gives the model a fuller picture of your fitness. If you only log easy efforts, the model has less signal to work with for faster race paces. - **Include some intensity.** At least one or two harder sessions per week (tempo efforts, intervals, or race-pace workouts) helps the model understand what you're capable of at higher intensities. - **Be consistent.** Regular training over the 90-day window gives better results than sporadic activity. Even a few sessions per week is enough — consistency matters more than volume. - **Log your activities.** Predictions can only use data that's available. Make sure your training activities are being tracked and synced to Stridee. ## Things to keep in mind - Predictions are estimates, not guarantees. Real race performance depends on many factors like weather, terrain, nutrition, and race-day execution. - If you're just getting started or coming back from a break, give it a few weeks of consistent training before relying on the predictions. - The model works best when your training reflects the kind of effort you'll be doing on race day.