TSS versus Training Effect

I look at the TrainingPeaks TSS as well as the Training Effects score on the Garmin Connect after long rides.

A comparison shows that they are not coherent. Here is one frequent example: I did an endurance ride a couple of days ago, 3 hours plus, close to 4, with a TSS of 188. Garmin’s Training Effect score is 198. I don’t expect them to be the same, that’s fine, different measurement metrics.

Today I did an endurance session on my trainer for a little less than 90 minutes. HR average was 3 higher than the longer ride. TSS shows 92. Which is about half of the longer ride of two days ago of TSS’ 188: 49%

For today’s ride, Garmin’s Training Effect score is 120. Substantially higher than the 92 TSS, but the importance of that to me that it is at 64% versus TSS’ 49% of the longer ride. This happens often.

It is quite a bit of a discrepancy. Which should I rely on as a more accurate indicator of effort? I understand that TSS has been knocked a bit when it comes to measuring endurance level efforts.

Neither. It is math, not physics.

You can use both to quantify your progress in the long run.

You better sum your low intensity hours and make sure you increase them week by week. In addition, do one or two high intensity sessions.

Thanks. Your advice on training mirrors what Ryan suggested. I was doing too much. He suggested I cut back to a Monday session of 5 x 5s, leg weights the next day, then rest until Friday. On Friday I do 3-4 hours of endurance ride on Friday with 4 10 minute Tempo sessions at the beginning, and a shorter endurance-only ride on Sundays. I am 64, and my recovery was slower than my previous regimen.

It has been amazing what doing less has done for me. In 4 months, my watts at long endurance rides have increased by 20 watts, and my VO2 max per Garmin has increased 10%.


Depending on your fitness, at some point an endurance ride starts to become a workout. If I’m riding 5 hours I’ll start the ride at the very low end of my Z2 heart rate (60% of max) especially if it’s hot. If you experience cardiac drift earlier on perhaps shorter rides more often.

1 Like

I think that is the main problem with TSS. It doesn’t take ride length into account yet we can all understand that the training stress from one hour at 70% FTP is very different if that hour comes at the end of a 5 hour ride rather than at the beginning

Takes ride length into account but assigns each hour the same value I believe.

1 Like

yes, so the last hour of the same watts counts as the first hour, so it does not take in physiological wear.

To your earlier point, Steven, Dr. Sieler posted a while back a very easy endurance ride he did for a few hours at a very low HR, but showed where the uncoupling started to kick in toward the end of his ride, meaning enough stimulus for adaptation.

1 Like

Yes, exactly. But the physiological stress from the last hour of a 5 hour endurance ride is hardly likely to be the same as the impact from the first hour