Ramp test - does feedback loop positively skew FTP results?

First time poster here. I’ve been enjoying TR for a few months now and on completing by third ramp test this week and taking a closer look at the numbers I had a realisation:

  1. As your FTP goes up, the starting power and the size of each step also goes up.
  2. This means you arrive at the peak power of your previous test sooner and with fresher legs
  3. Consequently, you can exceed your previous power, even if your fitness has not improved
  4. If FTP=75% of highest 1-minute power then this will result in a higher FTP being awarded
  5. Return to step 1 and repeat

Isn’t there a positive feedback loop going on here? Even at a constant level of fitness, each time you take the test it will get a little easier so you will get a higher FTP so it will get a little easier the next time and so on?

For reference, take a look at the last two ramp tests:
(1) 11th Dec, 339 FTP awarded (318 going in) https://www.trainerroad.com/career/rupertb/rides/68184105-ramp-test
(2) 4th Feb, 345 FTP awarded (339 going in) https://www.trainerroad.com/career/rupertb/rides/72442429-ramp-test

[Ignore the final minute in (1): I got out the saddle (I know I shouldn’t have) but this didn’t affect the result as average power for the final minute was less than for the preceding minute. FTP awarded was bang on 75% of the penultimate minute at 452*75%=339]

In ramp test (1) I start at 146W and each step the power goes up by 19-20W.
In ramp test (2) I start at 156W and each step the power goes up by 20-21W
Therefore in (2) by 1-min power peaked at 20:00 whereas in (1) it peaked at 21:00. In other words in test (1) I had to fight for longer in order to reach a lower power than in (2).

Now I’m thinking that the small FTP bump of 6W from my last test may be illegitimate. Let me know your thoughts. Would love to hear @chad and @Nate_Pearson have to say, too (huge fan of the podcast, guys)

EDIT: sorry if this has already been covered before - let me know and I’ll remove the post


I guess the growth in power can be modelled as an asymptotic function: at some point, you reach your maximum potential. So I would assume that, incrementally, you more quickly reach peak power in a ramp test, thus incrementally lowering the delta between peak power and initial power where you start the ramp test.

So, yes, I would expect there is some “feedback loop” present. Of course, if the true power curve is asymptotic, there can be a lot of error variation around this curve (as in statistical regression) dues to “test circumstances”: sleep, food, consistent training in-between previous and current ramp test, etc.

1 Like

I think the phenomenon you ID is real, but would be surprised if it was at play in your case given the very small difference in ramp steps (19-20 vs. 20-21 watts.)

When I first started I estimated my power based on the questions TR asks when you sign up (age, weight, experience etc.) It greatly underestimated my FTP, so my first ramp test had 12w steps; that meant it was 26 minutes from the first increase in power to failure for me, which meant a lot more total time above FTP. It slow-cooked me.

That first test led to a ~50w increase in estimated FTP, but I’d guess that was still low due to slow-cooking phenomenon. So when I re-tested a few weeks later TR gave me 19w steps in the ramp, and that led to another 26w increase in estimated FTP. I don’t think my FTP actually increased that much; that was due to a ramp test that was predicated on an approximately correct FTP. At this point, after a few more tests, I think it’s actually pretty close to correct.

Remember your FTP is calculated based on your best 1-min power, not the highest step you complete. So even if you only make it 20 seconds into the last step before failure, that influences your FTP calculation.

Personally I view FTP number as a useful training tool by which TR calibrates workouts, but definitely not an immutable value judgement on fitness. I wouldn’t worry your FTP is “illegitimate,” I’d only be concerned if you can’t finish workouts pegged to the new FTP. If so, reduce your intensity a few percent until you catch up.


Along with a higher starting point this adds up to significant amount though - in test (1) the power stepped up to 452W at 20:00. In test (2) it stepped up to 481W at 20:00.

I can see that the ramp test needs to adjust to each user so that a beginner isn’t cooked after 10 minutes and a seasoned cyclist isn’t still whistling after half an hour but I wonder if it wouldn’t be better to fix the power profile of the ramp test from your second test onwards, once your ballpark fitness has been ascertained. I’m realising as I write this that you could do this artificially by reducing your FTP to the result of your first ramp test right before loading up each subsequent ramp test throughout your training…

Add 20w to your FTP and retest. You are not considering the fatigue difference that will show it’s head. This is not a simple math problem that means a higher start equals a higher test result.

The guys did extensive testing with FTP set too high, too low, and “right”, and got very close results in all cases. The reality of more or less demand relative to your “actual FTP” will lead to a different time to finish. But the power and FTP calc will get you to the same spot within reason in each case. Differences may exist for people with varied strength in some energy systems. But this is not really a gamable test like suggested. You will not always pop at the same time in the test, it varies with the FTP start relative to your current ability.


Can you explain what you mean here? I don’t quite follow but I’m keen to understand if I’ve missed something.

I think Chad is stating that if you artificially bump your current FTP by 20W then you’ll have bigger steps during the test and therefore get to the point of exhaustion/failure a little quicker.

I’ve done three Ramp tests in this training sequence, it was the 8min and 20min tests when I was last using TR. My points of failure occurred at 18:04, 19:04 and 19:07. From other threads I believe that POF at 18:30 would give you your current FTP, i.e. no change.

1 Like

This does not seem to be true of my tests:
In the first the average power for 17:30-18:30 was 404W. 40475%=303W but my FTP was already 319W.
In the second test if I had cracked at 18:30 my max 1-min power would have been 430W. 430
75%=323W but my FTP was already 339W.

Because the profile of the ramp test is scaling upwards each time as I describe surely if you last the same duration each time then you will be finishing at a higher power each time and so always increasing in FTP?

  • Close. It is at 19:30 in the ramp, where you match the starting FTP (assuming compliance to the power target steps.
  • Yes, this exactly.

  • If you do a ramp starting with 200w FTP, and you pop at the 19:30 point, you match that 200w starting FTP.

  • If that same rider instead performed a ramp with a 250w FTP setting, they would fail much sooner than the 19:30, because the total power and steps are higher than the 200w Ramp test.

Remember that the Ramp steps are scaled to the starting FTP

  • The start is at 46% of the start FTP, and each step adds 6% of the start FTP.
  • So, a test at 250w start FTP will get to higher wattage much sooner. And a rider with 200w FTP will crash well before even seeing 19:30 to hit that higher FTP setting.

I am just finishing a tough workout, and might have too much foggy brain to be explaining this better. :stuck_out_tongue:

1 Like

You’d have to look through the other threads about that time, @mcneese.chad should be able to confirm or refute it.

My FTP went up by exactly one Watt between the first and second tests. During the next set of workouts I felt that I’d tested low so bumped my FTP by eight Watts. When I did the third test (this morning as it happens) my FTP went up by three Watts. I’m using a wheel-on trainer so the power at each step isn’t as accurate as it would be using a smart trainer in ERG mode. Just looked at the normalised power for the minute from 17:30 to 18:30 and multiplied it by .75, this gives me my new FTP value. Moving that minute along to be the last minute before I blew up gives a normalised power just half a Watt more. That would seem to support your point about increasing FTP if I lasted to the same time point but the last two minutes my power was 10W higher than the target, adjust by that “offset” and I end up with my original FTP.

Edit: Just looked at the data figures from the test. For all the steps prior to the final two I was within 2W of the target power, then things got a little haywire! The next to last step I was 7W over and the last step I was 16W over. It’s really hard to stick to a target power when you are at your limit!

We’re homing in on my point here. In your 250W example you reach the peak 1-minute power of the 200W test sooner because, like you say, the steps are higher. Because you reach it sooner you have less fatigue in the legs and can press on for, say, an extra 30 seconds. Although you are still popping sooner than in the 200W test you have nonetheless reached a higher peak 1-minute power and so are awarded a higher FTP.

This is exactly what happened to me - in my second test I cracked a minute sooner yet was awarded a higher FTP.

In other words, the higher your FTP is before the test, the higher the FTP you will be awarded from the test. The output of the test affects the test itself - which affects the output of the test, so on and so forth.

To take the example to an extreme - imagine that you could set your FTP so high that your test started and your max 1-minute power and you cracked after 60 seconds. Look at your lifetime/annual/season power curve - I bet your FTP is a lot lower than 75% of that value. If my FTP was based on that 1-min power then it would be 417W - clearly too high! You can’t reach that number in the ramp test because by the time you begin your final 60-seconds you’ve already got a whole lot of fatigue in the legs. Yet the higher your FTP rises, the harder the test starts and the closer to this situation you get.

1 Like

Or to take another perspective, imagine that your starting power and the incremental jumps are the same every time you take your test ie. your ramp test has the same profile every time. Your tests when then be clearly comparable because you could see that, say, yesterday you managed to last a whole extra minute compared to the test you took two months ago. The duration that you last would be the only measure of performance (assuming you are on ERG mode/sticking to target power) with no other differences to muddy the waters. Wouldn’t this more clearly show changes in your fitness over time?

1 Like

What you seem to be saying is that if you “game” the test then you’ll get an advantageous result. The point of the test is to follow the protocol which is designed(?) to give you about 13 minutes of steps (and therefore fatigue) before you get to the period where you’ll improve your FTP.

For most people that will be more or less correct but as ever there’ll be some outliers. Your first ever Ramp test will be a bit different but once you’ve a figure that’s in the right ball park, say +/- 5W then it’s going to be reasonably realistic and consistent.

My point is that the test ‘games’ itself. I only give the example of falsely setting your FTP to a super high value to demonstrate my logic more clearly by creating an extreme example

I don’t think that’s the case. The starting FTP is 46% of your current FTP (we’ll assume you have one) and you increase by 6% every minute. Since the test takes 75% of the final minute then your current FTP will be reached when you get to 133% (100 / 0.75). That gives you a range of 87. Divide that by 6 and you get 14.5 which is the number of minutes after the five minute warmup when you match your FTP. That calculation is irrespective of the actual value of your FTP.

So looking at your original post the actual values appear correct for both tests. In test 2 your starting FTP is 21W higher so you’ll get a little bit more than 1W increase in the size of the steps (19.08 to 20.34). What you are ignoring is that in test 1 your FTP went up by that 21W rather than 6W so you had to work harder. It’s nothing to do with the absolute value but the percentage increase over your current FTP - test one you improved by 6.6% whereas in the second test you only increased by 1.7%.

I had some similar thoughts to this previously. I keep getting the same results in my Ramp Tests, despite pretty consistent training I always seem to fail between 19mins and 19:30mins leading to a decrease in FTP. For my last Ramp I upped the starting FTP figure by 10w and saw a 2w increase in my FTP. Not statistically significant but yet again I failed before 19:30 which seems to be my limiter on my engine. I suspect that if I whacked up the starting FTP by 30w I’d see an FTP increase at the end of my Ramp, but then again like others have said I’d be gaming the system.

Don’t know for sure, but seriously doubt 6W difference (less than 2%) is meaningful (or statistically significant) in any way when your FTP is in the 300s. It would just be noise to me.

Add to that, no FTP test should be viewed in isolation. How were the subsequent workouts and were the RPE and overall feel of the workouts as intended.

We can all get high, low, or “perfect” results on any given day for any given reason.

Recognize that this is a number intended for setting training zones most importantly, and as one means of tracking progress as a secondary item. It is fallible and should be viewed in the greater context of training.

1 Like

I see the math point you are trying to make.

But how does this have anything to do with the actual real application of the Ramp test. If you are taking the test every month or two, your new fitness level will come into play.

You are also assuming that you didn’t drop any FTP. It doesn’t matter if your steps are +/- 1W difference than before, the test will estimate your current FTP.

Furthermore the team has a protocol that is simply making an assumption about your threshold. If you want razor blade accuracy you need to go get a lab test done where they are taking blood samples.

Any other protocol which is estimating your threshold is going to have ways to ‘game them’. So why all the consternation on a protocol that estimates your FTP and is going to have some +/- variance.

The users who use the protocols to the best of their ability monthly will get consistent results and will therefore get fairly accurate measures of their fitness. With all the variables that are going on with our bodies during exercise I simply do not understand why we need to often be covering this well covered ground.

1 Like

The canadian/british MAP ramp protocol involves 3 min steps, as well as a lot of the literature. There is no agreed upon protocol and to think they would give the same results is optimistic at best. I would definitely skew on the lower initial ftp setting, and extra few minutes sub threshold would be better than a few minutes less in terms of end result especially if used for setting FTP and not just determining MAP