Best FTP test - ramp or 20 min?

Training
,
41

Coach Chad has stated several times on the podcast that he believes that it’s not possible to overachieve on the Ramp Test. I also believe this. I think it’s much more likely that someone messes up the pacing of a longer testing protocol and get a lower estimate.

What’s implied in your statement is that you’ve done a Ramp Test and the FTP estimate you received you weren’t able to perform to. Again I assume you’re talking about failed TR workouts after a Ramp Test as opposed to something else like “there’s no way I could hold that wattage for 1 hour”. Before I blamed the Ramp Test for an over-estimate, I’d look at why you were failing workouts first. Mental toughness, nutrition, proper recovery, adequate workout preparation, just spending more time doing structured training and working on a weakness are all things that can affect performance. I personally trust the Ramp Test protocol and use that as belief that I can complete workouts at that level.

42

Overachieving on the test and overestimating FTP from the result of the ramp test are two different things.

It is a test to failure so you can have one of two outcomes: you stop at your absolute limit, or you stop before that. It is impossible to go beyond your limit, otherwise, well, it isn’t your limit.

The ramp test clearly overestimates, to a greater or lesser extent, for a number of riders just as it underestimates, again to a greater or lesser extent, for others. That’s the nature of bell curve distributions around a mean value.

Mike

1 Like
43

I should have said overestimate in reference to what Coach Chad has said previously on the podcast. However, where the goal of the Ramp Test is to go to failure, overachieve and overestimate are functionally the same thing since the FTP estimate is directly tied to how long you perform the test for.

For the population of TR users it’s likely that the overall results fit a normal distribution, but I don’t see how that has anything to do with whether the protocol over or under estimates FTP for an individual in a given test. Can you explain?

44

Well that got in depth! A rather more unscientific line of thought:

Leaving aside what exactly FTP is, we (mostly) have 2 purposes for this number:

  1. to anchor training zones
  2. to predict/set effort levels for racing, based on time. (subtly different from 1, but not much).

‘The best predictor of performance is performance its self’. Forgot where I read that… but it makes a lot of sense. The more similar your test is to what you want to predict, the more accurate it is likely to be, barring ‘other factors’. So the most accurate predictor for your best 5 min power is an all out 5 min effort, etc…

Logically it would hold that if you want to predict your (say) hour power, then going flat out for an hour is the best predictor, but here ‘other factors’ creep in. Namely it takes a lot of experience to truly pace that out properly, and the cost of a borked test is relatively high in terms of repeatability & effort. 20 mins is more achievable… the ramp less harder still to mess up… but each are progressively less like what you are trying to predict. I’d argue that the ramp is orders of magnitude ‘less like what you are trying to predict’ than the 20 min, really it’s looking at your maximum minute whilst fatigued from all the previous minutes…

Thing is, it’s also been stated multiple times around here (and seems generally accepted) that capability for work substantially over threshold is much more individual and harder to predict than work below/around. This to me is the ‘problem’ with the ramp. Once you reach the breakeven point, you’re doing a very different type of work…
BUT that is potentially offset by the fact that it’s pretty simple to execute, and those errors, for many at least are less significant than the errors due to poor execution of a more ‘similar’ test.

Of course a lot depends on what you’re training FOR. If you race crits and the like, you probably care a lot about those short/high power capabilities. It makes sense to anchor them accurately with a ‘like’ test. If you’re a long distance triathlete, the only reason you do those Vo2 intervals is to help ‘pull’ the aerobic capacity up. You don’t care so much about them (so long as they’re achievable), but anchoring and predicting your big steady state efforts is key.

At the risk of mentioning a dirty word, this is why we see things like sufferfest’s 4DP. Personally however I’m inclined to think life’s too short and I don’t need that kind of complexity.

As to martinheadon’s assertion that the ‘20min is simply wrong’… er really?? It has a lot more time and data behind it than the ramp! I’m inclined to think/guess that the ‘not adding up’ factor is down to NP being a rather imperfect measure, which weights suprathreshold efforts. There’s also (personally) no damn way I can get to 125% for 5 mins.
For whatever it’s worth, putting my numbers on, I ramp at 252, 20min (after factoring) 275, and did 276 (average, not NP) for ever so slightly less than an hour on a 40k TT. I’m a steady state guy, so I know which I’ll be continuing with - not to say it’s the best, just that it works for me. I’m also experienced enough to pace it out pretty well.

2 Likes
45

If you were to take the highest 1 minute power from a series of ramp tests and plot a histogram of that versus lab test results of power at MLSS for the same rider, you’d get a normal distribution with a mean value of 75%. Some riders would have a factor of less that 75% and some greater than that. That is how the ramp test can overestimate your FTP.

Mike

46

It’s definitely possible to “over-test” on every type of of shorter FTP test. I’ll explain why in my post.

1 Like
47

It’s an interesting theory and I agree that there will be variation, but my hypothesis would be that results would be more likely to be negatively skewed. Given the nature of the test I think it’s easier to have an “off day”, be fatigued or find it easier to succumb to a lack of mental toughness than it is to overcome it.

1 Like
48

Even if it’s more negatively skewed by fatigue etc, the highest values cannot be equal to the mean value, so it would still have instances where the rider would get an FTP higher than the lab test would give.

Mike

49

Much of this hinges on data that we don’t have, and are replacing with assumptions. We don’t have the data from the tests and evaluation that TR did when setting up the test. They started with a protocol that has remained the same (40% of FTP start and 6% increases each minute after the warm-up and repeat until failure).

The only variable they changed in the process was the percentage of reduction from the Best 1-Minute Power shown in the test. It started around 0.79 in the beta testing. That got reduced to the current 0.75 from their internal testing along with the results from the beta group. They looked not only at the test, but the subsequent workouts and stated feedback from riders.

All that is to say that I don’t think we really know where the 0.75 reduction falls within the typical distribution. Maybe it’s in the middle of the curve, but they may not have done that for any number of reasons.

They admit that it can’t and doesn’t work perfectly for everyone. It’s simply impossible for that to be the case as a standard distribution will exist and the predictable outliers along with that. So, there will be exceptions where the Ramp test is not a great tool or predictor for some riders. But the same is likely with any testing protocol.

I don’t know if it’s possible to predict or plan for the way that the test estimates FTP (over or under) in any meaningful way without some extensive data and research. I think that the best we can do is try the tests, recognize that variations may well exist, be mindful about the following workouts and determine what (if any) changes we need to make for our particular needs.

The whole concept of “best” is great, but unlikely to be reality. What may be possible is “best for me” with the proper evaluation and adjustment (if any) from the test. Despite it’s shortcomings, I still see the Ramp test as having some significant advantages over the other test methods.

As such, I keep using it and learning more about how it works for me, so I can make the best use of it for my training and evaluation. The other tests can and do work “better” for other people and that is great. Figure out which once is “best for you” and keep that as the measuring stick.

3 Likes
50

This is hilarious. The reality is it doesn’t matter. Your actual FTP fluctuates A LOT during a 4-week period depending on A LOT of variables. An FTP test is a way to create a guideline for a program. They are all accurate “enough” for the program to create an appropriate training stimulus.

The ramp test is “good enough” for most people who are doing the program and more accurate than other tests for MOST people.

5 Likes
51

my understanding is that this protocol has been around for 20 years, and the range is 72% - 77%. So the TR 75% implies there will be some above and below. And TR podcast has long reinforced you need to pay attention to FTP results and training, and modify as necessary to optimize your training.

When TR was in beta on the ramp test I did some research. Pretty sure I found the original published research on ramp testing and correlation to FTP, just looked and couldn’t find it. However I did find an interesting discussion on ST here: TrainerRoad New FTP test: Triathlon Forum: Slowtwitch Forums

3 Likes
52

Good point on the testing that has been done outside of the TR universe. That probably helps define the distribution indirectly since the TR is around the middle of the range from the outside testing.

53

Great post from @Bryce in the comment section of the TR blog announcing ramp test:

"While we did not do actual lab testing to arrive at the FTP calculation that we are using today, other researchers such a Ric Stern have. The Ramp Test that we created uses your VO2 max as a limiter, and we then use a calculation to relate your VO2 Max to your FTP. Of course, this will be different for different athletes, but according to Ric Stern, this relationship places your FTP at 72-77% of your max one minute power when using a ramped testing protocol.

We did not simply use this calculation though. When we began Beta Testing, we had the adjustment set at 0.77 times your max one minute power. We then had thousands of Beta testers try this test, go onto complete workouts at their new FTP, and provide feedback. Overall, this estimate was too high for too many people. We reduced the adjustment to 0.76. After a few thousand more tests, we assessed that the estimate was again, too high. We then reduced it to 0.75 which, based on our experience with the test and the feedback from the testers, yields an FTP estimate that allows users to successfully complete their workouts. And as it happens, that value falls right in the middle of the 72-77% range.

As for the “accurate estimate” phrasing that we used in the press release, it is important to note that every FTP test that takes place outside of a lab environment is an estimate. The 8 and 20 minute tests have been around so long that they are a staple in every cyclist’s training, however, that does not mean that they are anything more than an educated estimate at the end of the day. While they all aim to estimate your Functional Threshold Power, every cyclist is different and as a result, there will be outliers who lie either above or below the average that is used to calculate the test. The Ramp Test is the same way; we have come up with a formula that will provide most people with an accurate estimate of their FTP, but in a less time consuming and stressful way. And since the goal of the test was to estimate your Functional Threshold Power, it would be counter-intuitive to name the output anything other than FTP."

And that first paragraph explains a couple of things I keep posting:

  • why my FTP estimate from ramp test can be unusually low when my vo2 is detrained (due to both age and being naturally more of a diesel)
  • why I think ramp test is first and foremost a great way to estimate max 5-min vo2max power, and I’ll continue to use it for that reason as its non-intrusive (actually @stevemz posts convinced me of this along with using threshold effort to estimate FTP ala the TP ‘Physiology of FTP and New Testing Protocols’ article)
5 Likes
55

Farquhar is a good FTP check ride.

56

This is a really good point. I have been amazed by how the TR workouts since my first ramp test have been just within my reach but only just - ideal from a training perspective I imagine but not from a suffering point of view! :rofl:

57

“FTP” is really more of a proxy for lactate threshold work i.e. sustainable work for more than 5-8 minutes. for most people, especially those whom are not highly trained a 60 minute test is not realistic. it likely will result in training zones that are too low for that person.

58

A general rule is that a ramp test is very good at estimating your VO2 max training zones and for most people good at estimating sweet spot zones. However, this is not true for everyone. If your 20 minute and ramp test FTP are quite disparate (more than 3-5% or so) than you may need to use 20 minute test to set threshold and sweet spot zones and ramp test to set vo2 max zones.

1 Like
59

agree 100%

for most this is unlikely to be true.

61

Uh but outdoor FTP test results are typically higher than indoor?? Could your friend test again on the trainer? Just ain’t the same.

Sorry it’s been mentioned already, didn’t see it anywhere🤷🏻‍♀️

62

He has, and the trainer 20min protocol is a little lower, but within 3-4 watts. Not the 10% diff to the ramp. I guess I’m just a snowflake :wink:

1 Like