TSS: RPE vs powermeter


Hi there
I log my commutes on TR and have been doing so all year.
In the past month I bought a power meter. Also I recently (~5 weeks ago) did an FTP ramp test so my stats should be reasonably up to date.

Before the powermeter I was using RPE to estimate TSS and would typically get 25-30 TSS per commute (it’s not a long commute).

The results from my power meter (powertap P1s if it makes a difference) are now giving me a TSS of 45-50 for the same commute. I’m not putting in harder efforts or better times since buying the power meter.

In order to get anywhere near that TSS I would have to select an RPE of ‘All out’ for the same journey.

What’s going on here? I’ve no reason to doubt my power meter, but how can RPE estimates be so far off?

If I’d have had a power meter all year my TSS from my commute alone would have gone from 200 TSS/week to 400 TSS/week!

Any insight? Commute is only 11km so is it fair to say that RPE method is more inaccurate for short distances?



Are you commuting super hard? 50 TSS for 11 km is barely achievable; Would have to be an uphill all-out effort. (assuming a moving time of ~30 minutes?)

Assuming your commute is more Z2-Z3, the 25 TSS sound way more reasonable.
I commute a similar distance every day and TSS range from 18-25 usually (RPE, hrTSS and power meter TSS are all producing very similar numbers there).
Maybe it’s time to do the ramp test again? :slight_smile:


Hi there…intuitively I agree with you. My commutes don’t feel like ‘all out’ efforts.
However last night I hand-cranked my own TSS calculation and it came out with 47.

This was based on a normalised power of 296 and FTP of 267. I’ve just dumped the numbers in a spreadsheet and worked out that for my calc’d TSS to be 27, my FTP would need to be at least 350 which is clearly wrong.

Only other thing to play with is the moving time (which TR uses) vs elapsed time, but this takes the results in the wrong direction…ie TSS gets higher!

What am I missing here?


Is this coming from trainerroad? If not I’d check your FTP is set right on whatever website/device you’re getting the commute TSS from


Yes this is coming from Trainerroad but I’ve verified the calculation myself.
I’m going to check the fit file in Golden Cheetah to see if there are any spikes in power that would affect my NP.

If not, then the only logical conclusion must be that the ramp test massively underestimates my FTP…


Yes, the calculations seem to be correct, based on the numbers provided (as you verified yourself).

However, 296 NP on a 267 FTP is an IF of 1.11 - that’s more like VO2Max territory. In other words: super hard.
(For comparison: Bike leg of a sprint triathlon targets IF 1.05-1.10 for 20-30 Minutes)

I feel like “FTP set too low” is the obvious solution here - on what basis are you assuming an FTP estimate like 350 to be wrong? Just feel like the number sounds too high or do you have more specific reasons, like failing workouts at your current FTP setting?


An FTP of 350 would be 5.2 watts / kg…much as I’d love to be there, that definitely isn’t right!

As I said I’m set up on TR at FTP 267 which seems to be about right…I’m finishing TR workouts of ~ 90 TSS feeling fairly beaten up. Arguably I could go a little higher, but definitely not to the dizzy heights of 5.2 watts / kg. I’ll check the fit file for power spikes and let you know.

Might be relavant, but the wattbike i’m using for the ramp test doesn’t have an ERG mode. This doesn’t bother me, but I’m wondering if ERG mode enables you to push on that bit further given you don’t have to channel energy into controlling your power to match the on-screen profile…? just a thought

Please confirm Normalised Power calc

Can you retest using the p1 pedals on the wattbike?

Sounds to me like the use of two different power sources is likely to be the root cause of the problem here


Outdoor v indoor FTP calculation could be a factor too. For a variety of reasons (cooling/ natural motion/ motivation) my outdoor power is greater than I can achieve indoors, based upon how long I can hold above my indoor FTP test results outside in races.


@pirnie - very good idea - thanks. I’ll look into that. Wattbike is at the gym so will need to see if i can switch. If i can this would at least help verify the power outputs across the two systems.

@macy - yes totally agree, although I’d expect a 5% difference (maybe a 10% at a push), but not 30% (267->350)


I’ve just hand-cranked my own NP using GPX export from Strava and the formula from Coggan’s post here.
Result: 285 NP, vs TR’s calc of 296…
Now if my FTP was 295 I’d be down to a TSS of 35 which is more reasonable. I did go quite hard…

So question is: why is TR overcooking the NP? This should be a standard proprietary calc and shouldn’t vary.


Hi 5imon. I have pretty much the same issue - P1S pedals via Wahoo element bolt producing TSS scores that seem to be 30%-40% higher than when I’m not using power and simply taking the strava download into TR and estimating a 5 (tempo) effort for a long weekend ride. I test my FTP using the P1S, but I also use a wattbike a lot, so I know the comparison (max 5% difference) so that is not the major reason. My FTP scores are similar to yours - 270-300W in upright position and 250-280W in TT - depending on fitness. I’m pretty stumped and, like you, frustrated by this seeming incompatibility/inconsistency of various forms of tech we all set so much store by these days. Anyone from TR able to opine on this directly please?


I have p1s on one of my bikes too and they can massively overestimate if not calibrated. NP of 296 on a commute just doesn’t sound even remotely plausible.


Check the crank length setting on the P1S. Chances are it’s incorrect and is giving you dud data.

For a short while one of my P1’s defaulted to 170mm instead of 172.5mm and an all out sprint effort wasn’t even mustering 300w. It was frustrating and demoralising to think something was seriously wrong with me until I took a step back and thought ‘wait a minute’…

You’ll need the Powertap App if you don’t have it already.

Given you readings seem to be high I think perhaps the crank length might be set to longer than it actually is and the change in velocity/speed being rapid for that ‘throw’ is causing the over-read.

(Not completing a Zero-offset really shouldn’t give you wild readings, it’s not ideal to not do one and its quick and easy enough to do but I wouldn’t expect massive discrepancies)


Hi all.
I’ve been playing with my garmin settings and one thing that seems to make the results a bit more sensible is to turn off auto pause. Stages recommend doing exactly that.
Why does this work? Well it could be that the time value used in the calculation of TSS is the number of seconds recorded (not [end time] - [start time]). Secondly, and more importantly, the calculation of NP uses a 30s rolling average for which including all those 0’s when stopped is arguably the right thing to do.
This won’t affect the weekend rides as much since there’s relatively far less stopped time compared to moving time. On a short commute though, that stopped time is a higher proportion therefore has a larger impact.

So now I’m putting in commutes of NP around 245 rather than 300, giving me TSSs of around 35 rather than 50. Seems a bit more sensible, though still higher than I would expect.

@BennyC - I hear you on the crank length settings. I set this using the PowerTap app at the correct length, but it may have reset- will check again.

I’m also gonna stick the p1’s on a wattbike and check the outputs match.


Just wondering, are these single sided P1Ss, or dual sided P1s?


Single sided, but shouldn’t matter…I’m usually 50/50


I thought I’d ask as I think over BT the P1s (dual sided) measure one pedal and double it whereas over ANT+ they read both. I may be wrong on that though and its irrelevant anyway since you’re single sided


Absolutely this would have massively skewed your TSS, NP & AVG. Especially during a commute when presumably it’s a lot of regular stop/start where when pulling away you’ll probably be peaking 150-200% of FTP.

The efforts above threshold will drive NP as will erratic power output. You should be able to ride your commute for the same average power with a lower TSS by riding more ‘economically’.


Like you say, pauses at zero power should be included in the calculation of NP. Nice sleuthing to figure this out!