adamrice 

When I first got a smart trainer, I rode a 20-minute FTP test to establish my baseline, and have ridden them periodically since. It had been quite while since my last test—some of that because I didn't think I'd see much improvement, but mostly because the 20-minute test is just really hard.

I'd heard of ramp tests and did some reading about them (especially this article), and decided to give one a shot. I constructed a workout and rode it last night. End result was right about what I was expecting, and it was considerably less daunting than the 20-minute test, although my legs were gelatinous for the rest of the night. Anyhow, A+++, would test again.

rubiksoval

So you're estimating FTP based on a short-duration test that may or may not have anything to do with your FTP?

I genuinely don't see the point aside from simply getting in a workout.

Quiglesnbits

I am in the process of building back fitness after a long layoff, and I've personally found the ramp test to be a marginal barometer for my ability to actually maintain longer efforts (above 5 minutes) near my resultant FTP. This may just be a function of me as a rider, but I'd say that the ramp test gives me an "aspirational" result which is maybe 5% high. I'm sure it's more accurate for others, and maybe even measures low for some people.

Carbonfiberboy 

Yeah,, FTP is a lot more interesting than most folks think. I used to use the short tests. They don't tell the whole story, particularly they don't tell you anything about your aerobic endurance. That said, they are good enough for folks who want to put some numbers to shoot for in their anaerobic intervals, because short tests can have a large percentage of their result being composed of anaerobic power. Obviously an hour (or even 50') FTP test will be mostly aerobic power. If one wants to do the Hour of Power test, it's going to take a different training emphasis than just doing short intervals, but I suppose what one wants for a test will depend on who one is going to do with its result. So far, I like the effect on my cycling which working at getting better at the Hour of Power test is having.

Did I get that right, Rubik?

rubiksoval

Above five minutes is not anything like FTP.

rubiksoval

Yep. Unless you're testing 50-70 minutes, it's hardly even an FTP estimation so much as a guess (which often times seems quite optimistic).

Quiglesnbits

Correct

Carbonfiberboy 

That's what I've found. There doesn't seem to be a relationship at all, percentage or otherwise, between short tests and FTP. I admit to being dumbfounded to discover what a huge portion of cycling literature is just BS, probably originating in the desire of popular cycling coaches to produce more apparent progress with less effort from their clients. Doncha just hate how converts proselytize?

atwl77

I personally do not feel like ramp tests are accurate. A part of it is because there's the (contribution/dilution?) anaerobic power as mentioned by another poster above, but also due to the fact that the (shape?) of the ramp itself can significantly skew the result. Depending on the ramp's starting power, how much power each ramp goes up by, and the duration of each ramp, this can have a significant effect on the point when your legs call it quits. The only good thing about ramp tests is that they are pretty easy to do and produce repeatable results, compared to a longer 20-minute or 60-minute test where inexperience and/or mistakes in pacing can produce inaccurate results.

Having said that, though, I'm rather fond of using The Sufferfest's Half Monty for determining FTP and MAP values for interval workouts. Although it starts off with a traditional ramp, but afterwards there is an additional 20-minute heart rate-based section to fine-tune the results of the ramp test. While I don't know what sort of magic math sauce goes into their 20-minute test, but the numbers it produces (which sets the values for subsequent interval workouts) feels pretty... uh... (nice/good/not-too-painful?)... for a lack of better word.

Branko D

Personally, I just guesstimate it based on recent "full gas" efforts - eg, I'll take 80-ish% from a full gas 6 minute climb, or 95% from a 24 minute climb, or whatever in between (where the anaerobic contribution in terms is the same in terms of kJ). That, rounded, gives me about 265W as a recent estimate. So, then I go out to train and do, say, 4 x 8 minute climb intervals at around 108%, which is in the 285-290W range, and that's just about the most what's doable for me. Trying to kick them up a notch to 300W didn't quite work out and I started to fade. So, that's how I know the figure is roughly appropriate, at least from a training perspective. Hopefully in a month or so I'll be able to do them at 295W-300W and then I'll know I made some progress.

Would I be able to hold this power for a hour? That's a different question - probably yes, but I'm not sure if it's really relevant to whether it's appropriate for setting training zones and figuring out what power to hold for what interval duration. If 2021 sees racing return to something resembling normal I'll certainly try.

So, anyway, take the ramp test figure and just do the training based off it; if you can't do the higher intensity stuff properly, the number is too high. If you can do it with some power to spare, then try it and if it works out, mentally readjust the number a bit upwards. If you're not going to train, the number isn't even all that important anyway. If it comes to pacing a TT type effort, the worst thing which can happen if you use the number you've got is that you can fade a bit towards the end, and then you'll know that it was a bit optimistic.

In time you get a feel for what you can do.

Hermes

I have had several cycling coaches and none of them have used a construct to generate more apparent progress with less effort. Faster times with less effort, of course. They have all been excellent and have taken me to levels of results that I could not have done on my own.

When Coggan created the FTP construct, he did it to model human physiology. Do you want to know what FTP really is. Much like movie the Matrix, one has to decide which pill to take.

Hermes

I have done 20 minute tests and used hill climb and time trial results for setting FTP. I have not done a ramp test.

Today, I have many years of power files and results from races and events. I use an estimated FTP based upon looking at current results and then modify the workout i.e. revise the FTP based upon the day.

If I have an upcoming event or series of events then the training plan is prepared in support of those events. If I am going to race a 40 Km time trial, a one hour FTP measurement is more meaningful than a shorter test. The main reason is that one hour of constant power is mentally draining. The physiology may be capable of X watts for an hour but the brain and neurology may not be. And some athletes are better than others at constant power for longer periods of time.

rubiksoval

Me, too, except I don't even concern myself with the FTP aspect, just the power I'm aiming for in that particular workout. Can I hit it? If not, why not and/or what do I need to adjust for the next time, if anything?

Any particular number on any particular day just isn't that paramount in the grand scheme.

mr_pedro

The ramp test is not just a max test over a short interval to estimate your longer duration power. It actually takes more like 20 min to do and exhausts the anaerobic system in the process. It varies per person how it exactly correlates to FTP, but I have found it to be very consistent and good enough for setting interval intensity.

Anyway, FTP testing might soon become a thing of the past. Trainer Road wil be using machine learning to automatically analyze your progress in various energy systems using just the regular workout data, without the need for specific FTP testing.

surak

I've done ramp tests, 4DP, and gotten estimates from both hour-long race efforts on climbs as well as estimates from 20-min all out climbs during longer rides. I must be the rider those tests are all modeled after, because their estimates were within a few watts of each other when done around the same time and fitness.

ZHVelo

That sounds cool and might make me sign up.

mr_pedro

It is pretty cool. And it is not just about the interval intensities, but the entire training program that is determined by ML.
Automatically adjusting every workout depending on performance or if you skipped a workout.
Who knows how well it will work from the start, but it most likely will improve over time.

jpescatore

Agree with Surak - I tried the longer and short FTP tests, the ramp test, a long complicated multi-part test, etc. The classic FTP equivalent number was within +/- 5% across all of them - well within the real (not claimed) accuracy of most power readings done on different days.

I'm not a racer, so to me FTP is just a marker to check periodically to see if the riding and/or workouts I've been doing (or not doing) lead up to the marker going up or going down.

burnthesheep

That's the future right there. The machine learning.

5% can be make it or break it when it comes to doing sets of intervals. 5% of my 8min intervals can be ~15w. 15w too little and it's going to be too easy and not give good gains for the workout. 15w too much and you'll fail.

If you're smart you'll back off or give it gas during your workouts.

But, if you're doing that........what was the point in the first place doing a test? You could just guess at your intervals and adjust during the first workout. Then you're good and didn't waste days gaining some form to be able to do a test, do a test wasting a day, recovering the next day.......etc.....

This is where the nicer oxi meters and some machine learning could come in. Power and HR are still indicators of what's happening. The Moxy style stuff gets into the actual "thing". Pair that with machine learning prior to workouts, or even during sets to adjust loads based on prior workouts and current workout data.............that would be cool (and expensive).

I feel elites would gain something with that. But I think the richer low volume amateurs have the most to gain by optimizing the time they do have. I think even data nerds like me might be wrong about something we're doing and have 5% or so to gain just optimizing what we already do. That's more gain spending even $1000 on a device or $100/yr on a service than I'd stand to gain buying more bike parts.

gregf83 

There’s no way you can know that. Maybe you finish the workout a little fresher and can do more the next day. It’s just not that precise that you need to hit a target within 1%. Even it it was critical, you don’t even have to be smart to understand that if it feels easy this time, make it a little harder next time.

Wout van Aert doesn’t need machine learning to tell him to ride up the volcano 21 times. Most recreational riders just need to ride more.

burnthesheep

Wout is paid to ride a bike a lot. The old "amateurs just need to ride more" isn't a good argument. You can't invent time. If somebody is using the hour expenditure they have to spend......you leverage other things to be faster. Like optimizing the time you have. You don't sit around and go "I guess I'll be slow since I can only ride 6 hours a week". That's both mentally and physically lazy when it comes to performance.

I didn't say 1% either. I said 5%. I can do 90% for about 2 hours. Your 100% of ftp is well, 100%.....50 to 70min. A whole hour less. 105% is probably at best going to be a duration of around 20min, done once. Or sets of 9min under/overs or 8min each.

So yeah, 5% can matter. It matters more the shorter the duration of the interval.

Saying "just finish fresher for tomorrow" just shows you don't understand how you force adaptation during a workout. "Finish for fresher tomorrow" means you failed to give the body an effective enough stimulus to adapt to what you want it to. "Fresher tomorrow" may work if you're on a diet of lots of sweetspot or tempo trying to optimize how much TSS you can cram into a week at that power level. But not for threshold and VO2 work.

gregf83 

Sure Wout’s a pro but plenty of non-pros have extra Time they could use for training. Machine learning and 6 hrs/wk isn’t going to beat 15hrs/wk.

And 5% might matter a little in a workout which you could possibly prove with a long enough study but in a season or training block it’s not that important. It might make the graphs look pretty but if you’re doing a set of 6 intervals they don’t all have to match. If the first few are easy pick it up a little, too hard back off. VO2Max intervals are a function of time and power and again not that precise. You need to spend time at VO2Max but there are numerous ways to get there and they don’t require a power within 5%.

Hermes

I think you and I approach the setting of starting power for a workout about the same. Maybe visualized a little different but we are not a slave to a particular number for any given day. Assuming I am not getting a prescribed workout, I decide what the workout goal for the day will be such as endurance, threshold, VO2max, anaerobic threshold or neuromuscular. I then have to pick a course/route or maybe the trainer that supports the goal and which bike I will ride I.e road or time trial. Not all routes on the road are good for some workouts. Finally, I need power targets for the ride. After I pick the power target, I start the workout and may adjust the target as the workout progresses depending how I feel.

Back to testing ramp v 20 minute test. One of the coaches I used would have me submit my workout via a special Training Peaks account and in the narrative, I would assign a value of 1, 2 or 3 for how I felt making the power. 3 meant that the target power in the prescribed workout was harder than it should be. 2 meant that the effort felt about right for the power generated and 1 was a chainless day.

So if an athlete does a ramp test and reports a 1 i.e. it was easy and chainless, then one would assume he will score a high FTP and may be capable of even more for that protocol. I suspect the same athlete may rate a one hour test a 3 and say it was very hard to make power especially for the last 15 minutes.

In this case, OP is probably rating the ramp test as a 1 or 2 and the 20 minute test a 3. That could be an early indication of the type of rider OP is and what events OP would be good at.

The take away for OP is that the results of testing are a sign post on the road of what is ahead not a speed limit. And different test protocols match some athletes better than others hence some will seem easier. We are all capable of a lot more than we think we are so think big.

Branko D

That's all true and an argument against testing frequently once you know where your recent baseline is which you can adjust around. If you don't have a baseline then you are probably fresh enough to do a test to give you a starting point.
​​​​​
The way I see it, the ML is really going to track your HR and power and how you are completing your workouts and set what is an appropriate power to aim for in your intervals based on that. In many ways it's doing what many are already doing.
​​​​​​
The other side is suggesting trainerroad workouts which should help your fitness. That could work better than a canned plan, but personally I'm not interested because I really treat the smart trainer as means of last resort, prefer doing my workouts outside and last I tried the options for training outside weren't so great.

burnthesheep

Honestly recently with no event outside of a known future TT months out, I've just been optimizing my TSS per week with tempo and sweetspot. Then I'll take a day once in a while to hit the more neglected higher power stuff briefly. Then go back to optimizing my TSS. Usually a higher power thing to do is toss in the occasional practice TT between 10 and 30min long.

I just alter the power based on how long I have to ride by a hair.