Comparing Garmin FTP estimate with Ramp Test?

I did a 40min effort, doesn’t that count?

I’ve seen auto FTP estimates from doing a 52-min effort, and 33-min effort, a 67-min effort, etc.

Right, so I’m not sure why people are telling me to do a 20min max effort.

1 Like

Maybe you need to add a VO2max effort in there? I’ve noticed my Garmin FTP move for both long efforts and more frequently for 3-5 minute efforts.

The Auto FTP calculation varies by Garmin model, with the older Garmins having simple estimation that requires steady 20-min efforts. Which Garmin do you have, and what are you using for HR?


even if it uses a longer duration than 20 mins, unless it was max effort a previous max effort of a shorter duration may still override the new “test”

Sure, a couple months ago I did 3 x 2-min at 125%-133% and received an FTP update. And that was not a max effort.

I’ve a 530 and currently using a Polar Verity Sense, but the FTP estimation was equally junk when I was using a Garmin HRM.

Sorry to hear that. Same model as mine, so I’m assuming you either have issues with clean heart rate variability data, or enough steady efforts in order to get a good VO2max estimate. Or you need to update personal information, in particular Activity Level and HRmax and possibly Lactate HR.

Did you recieve an V02max update for that too?

Cycling VO2 Max Estimate

If you are having difficulty obtaining a cycling VO2 Max Estimate, try riding a course where you can maintain a steady effort. You can also try using an indoor trainer.

  • Cycling activity must be 20 minutes in length or longer without stopping
  • Heart rate data from either a built-in optical heart rate sensor or from a chest strap
  • Heart rate must be elevated to at least 70% of your maximum heart rate for at least 20 minutes continuously
  • Power meter is required

Still trying to get a handle on how garmin is doing these calculations since I don’t see any documentation on their website since the firstbeat white paper was published

I only take post-ride pics of the FTP updates. Looking at Garmin connect my vo2max bumped up from 42 to 43 the week of May 3rd (week of the 17-min climb). And WKO had estimated vo2max absolute go from 3.990 to 4.003 on the 17-min climb. Its easier to see vo2 trends in WKO (and workout level % vo2max), so I mostly ignore the vo2max estimates in Garmin.

And to be fair, I see updates to FTP estimates without vo2max changes.

1 Like

I read all the firstbeat stuff and googled and came up empty.

My 530 always wants me to do more high aerobic for optimal training but I think that pleasing the Garmin gods would kill me.

1 Like

A little high aerobic goes a long way :wink:

My auto FTP rarely updates.

But it does like to auto update when I do a ramp test, and it is usually pretty close, 5-10 watts of TR. The targeted efforts seem to line up with my efforts as well (I have a MASSIVE base, so that blue bar is pretty long).

Currently my VO2 and FTP seem a little low, but my last few hard efforts were done at 6000’ higher than I live. I’m not going to worry about them though.

Found it, I’m guessing they are doing something similar to this:

Relative VO2Max = [(10.8 x W)/M] + 7


W = watts
M = cyclist weight in kg
VO2Max = mL/(kg x min) (ACSM, 2010)

Thus FTP(W)=(VO2max)-7)*M/10.8

This gets me in the right ballpark for my FTP from my estimated VO2max. Silly that the 520 can’t do that calculation or the garmin connect servers

1 Like

WKO puts me at 42.1, Garmin puts me at 43, and that formula underestimates at 36.5. A pretty big gap.

There is a more detailed formula but it requires an estimated efficiency.


yea IIRC golden cheetah has a pretty comprehensive calculation as well but I don’t recall where it is. I think the efficiency is incorporated into the constant in the formula so you could probably tweak it to match your physiology

GoldenCheetah uses Critical Power as far as I know from looking at the source code. CP is based on an all-out max effort of 2-3 minutes.

I’m probably misremembering mark livermore doing those VO2max calcs with efficiency in one of the other forums, guess they never implemented that in GC and went with CP instead, its been years since I played with the software

Looks like they use 2 values, one short between 2-5 mins, and one long 20-30mins

You can configure the intervals used. We recommend using multiple tests of between 2-20 minutes.

But the user can set and do this however they want.