Improvements to Fitness Signature Detection
One of the many reasons that athletes love Xert is the ability to track their fitness – Peak Power, HIE, and Threshold Power – without doing required dread FTP tests. Xert will automatically identify any maximal effort in your ride data, whether from an all-out Strava Segment effort, a maximal effort on a Time-Trial, a heart-pounding Zwift race, or trying to hold a wheel/breakaway on a group ride! If you’re able to exceed what Xert predicts your MPA to be, you will be awarded with a Breakthrough! and your fitness signature is automatically updated to reflect your improvement.
For more about Breakthroughs with Xert, check out this article, here!
In some rare cases, Xert’s fitness signature extraction algorithm might find a fitness signature from your maximal efforts, but due to possible errors or anomalies in the data, it may not be as precisely tuned as it could be. Some users may have seen their HIE too high and TP too low. Sometimes the opposite can happen. Users can be left scratching their head or contacting support for assistance.
In this latest update, the algorithm Xert uses to determine your fitness signature has been improved. Over the years, the team has recognized patterns in user signatures and have incorporated these patterns into the signature extraction process. Errors that throw off the algorithm happen less frequently so that signatures that may have deviated from the norm in the past, are much better today.
If you saw strange signature values in the past, this update will help with that.
Old Extraction Algorithm – The extracted signature (1112 W | 36.6 kJ | 259 W) does provide an MPA profile which matches the power data. However, an HIE of 36.6 kJ is highly unlikely for an athlete with a PP of ~1100 W, so the extracted signature above may not be the optimal signature for this athlete.
New Extraction Algorithm – Notice that the updated extraction algorithm calculates a different signature (1112 W | 25.7 kJ | 265 W), even when starting the extraction using the incorrect signature from above. The updated extraction algorithm provides an HIE estimate that correlates with their Peak Power of ~1100 W.