- Peak thresholds for bike power, run power and heart rate trended over time. Each activity is analyzed to determine the peak thresholds. The thresholds from each activity are stack ranked against the last 8 weeks to determine if you have an achievement 1, 2 or 3 with 1 meaning a breakthrough. We capture all the available context (e.g. cadence, temp, altitude, velocity, gradient, etc) from each peak threshold to support the Threshold Factors algorithm mentioned below.
- Run Best Efforts for 400m, 800m, 1600m, 5,000m and 10,000m. We record the pace in meters / second and context for you hitting the peak pace for all of your run activities.
- Segment Effort Watts/kg achievements and leaderboards. Pivots Strava’s PR from time to watts/kg using your weight and power from power meter. This feature is only active for Rides and Virtual Rides done with a power meter when your weight has been configured in Strava.
- Compute and maintain a trend of the typical Cycling Power Metrics (Normalized Power, Intensity Factor and Training Stress Score)
- Bike Prediction: Predicts which bike you rode on an activity and alerts you when activity wasn’t recorded on that bike.
- Threshold Factors: Assess the impact several context variables are having on an athlete’s peak threshold achievements [bike power, run power and heart rate]. What variables (e.g. cadence, temp, altitude) are influencing them up and down and to what degree are they impacting the result.
- Best Effort Factors: Assess the impact several context variables are having on your best efforts across the 5 distances we capture for running best efforts. What variables (e.g. cadence, temp, altitude) are influencing them up and down and to what degree are they impacting the result.
- Threshold and Segment Effort Anomaly: Assess peak power for watts/kg and all segment efforts with power for anomalies (e.g. too high or low) in hopes of helping spot PM calibration or battery issues. This will also be used to help us establish when a users weight or wattage is materially wrong when they consistently have segment performance anomalies fast or slow for watts/kg.
- Image Classifier: The cover image on activities are put through Google’s Inception v3 algorithm to identify what objects are in the photo. This information is used to maintain an image trending list showing what the top 10 things people add pictures of to their Strava activities.
- Training Partners: Cluster users by their watts/kg for both running and cycling in their geography to help people find others who can ride / run at the same pace. Currently only supported for Bike Watts/kg and doesn’t as yet factor in geography. We’ll add both Run Power/kg and geography when we exit beta.
- Segment Performance Prediction: Predict performance (moving time) on any segment using the athlete’s Peak Power Thresholds. For example, predict if you could achieve a PR on a ~5min segment with your current 5 minute power threshold. Eventually this will be used on segments with a goal attached.
- Segment Performance Factors: Calculate variables impacting Segment Performance (moving time) across all strive.ai efforts across a segment. This feature brings the same “Factors” approach we do to your Peak Thresholds to each Strava Segment helping athletes understand what the keys to success for a top placement are on a particular segment leaderboard.
- Bike Power Meter Calibration: Allow you to learn if your bikes have an impact on your peak power curve. For example, does one power meter read higher/lower than the others, or do your virtual rides read different from real rides.
- Apple Health Kit integration: Bring in additional context data into strive.ai to improve performance factors insight.