If you have Cycling Analytics set up to send your files to Strava, it’s always been possible to have Cycling Analytics also set the title on Strava. Two new options to set details on Strava have been added on the linked accounts page: for setting descriptions and commute status. When checked, these will set the appropriate details on Strava when entered in the “new rides” tool (to the left of the calendar) or the file uploader. Commute status depends on the purpose being set to “commute”. Titles are only set when changed (there was recently a bug that meant default titles were being sent). The title and description can optionally be sent to Strava too when setting them on a ride page. Sending rides to Strava from the Strava status page can also include the three details (the title isn’t included if it looks like a default title).

The other change today that could do with some explanation relates to how activities with only heart rate data are included on the training load chart that uses power-based training load. By default, this scales TRIMP values to estimate an equivalent training load. It previously used a linear fit on your activities with both power and heart rate, which generally worked fairly well, but not forcing it to go through the 0,0 point meant activities with very low TRIMP scores could have surprisingly high training loads (activities with 0 TRIMP were considered to have 0 load). This becomes more of a problem when less intense forms of exercise, like short walks, are included.

After exploring many options, this has now been replaced with a hyperbolic fit that goes through the 0,0 point, and, depending on your data, will be somewhere between a linear fit and a moderately sublinear fit (so doubling TRIMP will result in less than double the training load). From crunching the numbers, this results in a considerably better fit for activities with low and high TRIMP scores, and marginally better results for moderate scores where the old linear fit was already very good. Everything I looked at visually also supported the idea that this is a better fit. A chart for this can be viewed at the bottom of the training load page.

There are a few other little changes. This also does a better job at getting ride types on import. A “gravel” ride type has been added. And, about a week ago, a lot of pages become a lot faster to load for those with many rides in the system.