How I Learned to Stop Worrying and Love the RAMP

If I was writing clickbait for this amazing audience I’d call this; "How I added 100W in two weeks” … but I’m sure you all know there’s no such thing as a silver bullet.

Instead, I’ll tell you I’ve been a Trainer Road user for about two years and until the last week of 2018, I’d done all of 5 FTP tests in that time (just one RAMP).

I’ve now done 12 in two weeks (starting Dec. 26th), and I thought I’d share what I’ve learned along the way in the hopes it would help others. I didn’t try to crush every test, just worked my way up based on my gearing. I’m effectively “back where I started” (more on that later) in terms of raw numbers, but much better off in so many ways.

Here’s a picture of what it’s looked like;

I was really spurred on by comments in the Facebook group (thanks John.P, Alvin.R and Alex.S among many others in that thread) that hitting 19:30 would result in you testing to whatever FTP you started with initially. Thank you all so much for that information, and inspiration if effectively allowed me to tell myself I could gut out at least that time and forced me to practice my pacing.

I’ve always assumed my lack of pacing went hand in hand with an understanding of gearing ratios (gear inches) only slightly beyond basic. That’s why I decided for my new years resolution I wouldn’t shift during the test at all, not once - not ever. Maybe it’s obvious to others, though for me it sounded like a dumb idea, but it’s been anything but in terms of helping me get comfortable with the test.

So in one of those “what’s the worst that could happen” moments I decided to look up my gearing ratios (in terms of gear inches), order them from smallest to biggest, and continue to take RAMP tests - accepting any of the suggested outcomes - until I topped out.[1] Now’s probably a good time to admit I don’t have a smart trainer, but my Kinetic Road Machine has been solid through the years, and I put my money into the power tap P1 (left and right) pedals.

I’ll spare you any sob story, my previous (painful) RAMP test was 173, but I haven’t trained much since July, and I’m not a racer[0]. I’ve done exactly one Sprint Tri, way too long ago, which I loved and tell myself I’m training for more, but I keep letting life get in the way. Really, I use Trainer Road because it’s cheaper than a gym, helps me hurt my buddy on rides, and you all (and the podcast) are super motivating.

Basically, I try… but I hope I’m not the only one who wishes we were shooting for a high kg / W instead. So, perhaps I’m something of the target n00b for the RAMP test, and certainly I was dreading doing a “real” one to start the new year.

So looking back at my first RAMP test I think my pacing strategy went something like;

  1. Starting out thinking I’m crushing it…
  2. then I figured I should start to “spin up” a big chain ring,
  3. then I ended up getting buried and,
  4. jumped back down to my little ring to try to get cadence,
  5. then I gave up on finesses and tried to muscle it out again with a big ring.

Yea it was as terrible as it sounds…. I tried to estimate my gearing from that one “untrained” RAMP test and think it went something (per interval) like;
34x19, 34x21, 34x19, 34x17, 34x17, 34x17, 50x21, 50x19, 34x15, 34x14, 50x17, 50x17
50x16, 50x16 … Tip: don’t do that

Here’s what my pacing looks like now; pacing_chart

As you can now imagine I started my first test in the “granny gear” and tried to keep going. Of course that sucked, but then I took that initial number and have been working my way up ever since. [3]

Here are a few ideas I’ve had as a result of this experience;

  • Having a fixed time target really helps, at least for me. I’m sure my “improvements” would have been more drastic if I took some of the initial tests longer, but then I’d have been far more demoralized when the latter tests couldn’t go as long.

    • Knowing I wasn’t going for some “maximal number” really made the multiple tests a lot more interesting to me, seeing anywhere from 2-24% “improvement” (excluding the first three I’ve averaged about 5% each test) even if it’s less physiological and more just measurement based has been very motivating
  • So far, every “gear inch” has been worth ~2.5 Watts of power on my FTP - though I expect that to degrade once I get closer to my current limit.

    • Gearing does matter, but not near as much as I thought. I think from here on out I won’t bother shifting. Something I now assume all you “smart trainer” owners have known for a while. [4]
  • Generally speaking my HR ramps really consistently. I know HR can fluctuate and we don’t use it for training, but the trend lines suggest that HR could help you figure out if it was a “bad” test
    - The slop is consistent because the work increments consistently, but I noticed a large delta based on gearing (e.g. higher cadence implies higher heart). Again, not surprising but something for me to think about in the context of testing.

  • I am getting way better at pacing (maybe within a watt or two as you can see in the graph above) and after about 5 “easy” rides (working my way up ) the initial trend line was able to predict my next FTP really closely.

    • this is really the result of being “locked” into the protocol vs. some amazing mathematical insight and I’m sure the curve will “break” soon once I hit closer to my physiological max.
    • I’d say at most I have another 30% improvement (50 Watts) I could go if I keep “running the rings” all the way - which isn’t going to happen but gives me a good set of targets to train towards
  • I started hitting an FTP >= my eventual FTP on roughly the 10th step, which means I only had to suck it up for 5 more minutes. However, in some cases I hit it a minute earlier in the big chain ring. So that may factor into people’s experience and outcomes.

    • I suspect some of the “sprinting FTP vs. endurance FTP” I’ve read in this forum may have to deal with people’s gearing selections / preferences - e.g. if you’re more comfortable muscling a higher gear combination for the test vs. longer workouts.

Overall, I started hitting an FTP >= my eventual FTP on roughly the 10th step, which means I only had to suck it up for 5 more minutes. However, in some cases (e.g. in the big chain ring) I hit it a minute earlier, so that may factor into your experience.

I have lots of charts, and can tell you way more about what I started to track if anyone’s interested.

Recommendations for other RAMP takers:

  • don’t look at your number or the line - flick the middle screen to the “numbers” and watch the power per interval (i.e. the overall average). I have 6 sec of power smoothing and even then watching that number too much causes more problems than it’s worth
  • don’t try to “jump” it - I had a natural (e.g. ~10 W) jump when the clock changed over even when I tried not to, this will then drift back down, possibly under the power target… that’s ok. (see next point)
  • give yourself 20 seconds to “settle” into the new target and think about “dragging” the number up (if you’re low) or “slipping” it down (just slightly less cadence). obviously don’t make it an obvious over/under interval but don’t freak out, there’s a whole minute to average your numbers
  • if the interval wants to be “too easy” - LET IT, they’ll get hard later. I thought I was pacing more consistently by say trying to hold 100W for 3 intervals vs. 80, 90, and 100 W respectively (aka don’t try to be a hero too soon).
  • pick a single gear - maybe that’s not actually the best advice, but it’s worked well for me and I’m certain when I do need to use multiple gears in testing now I know how to break it up much better.

Contrasting my “Power per Gear Inches” metric vs. the “Cadence per Gear Inches” numbers I suspect I’m not being as efficient (and consistent in applying pressure) so I have a few more potential elements to explore. However, I think I’m finally ready to start up my training block and take a RAMP test “for real” and see what I can do about increasing it.

Hope that helps some of you who may be feeling a bit more trepidation!

  • tatlow

PS here are a few things I’d recommend to Trainer Road to consider;

  1. When you pause the test it’s confusing that you have to “Return to Ride” to end it; e.g. “Pause” -> “End Workout” ends up with a greyed out “Save & Close” and the “Return…” option is hidden at the bottom of the screen

  2. The test should automatically stop when it knows you can’t reasonably make a target - e.g. if you’ve done 2 intervals at 60% - otherwise the test can drag on in the worst kind of way where you’re not sure if you’re supposed to keep going

  3. I’d suggest considering a fixed duration of the test vs. having it so open ended, though I did love reading Coach Chad tell me they’re going to hang my picture up

  4. “copy / paste” from the website is regrettable, I had to transcribe data (i.e. .FIT files aren’t very friendly to spreadsheets) so maybe give an optional format to download in, as well as an option for bulk export of rides during a certain timeframe. What I’d really love is for you to consider mirroring the Golden Cheetah API;

  5. Mentally the phrase “8 minute test” sounds awesome, and sucks in practice. “RAMP Test” sounds scary, especially when I tell you I took it for 20 minutes, but for me the method proved far more doable and trainable

[0] I have a stock Raleigh Militis 1 (other than saddle and pedals) that could really use some love - aka I have some slight mechanical issues, including a slightly bent derailer, hat cause drag

[1] I assume everyone knows that sometimes a small chain ring can be harder to push than the big one, depending on the cog size but I’d never explored that extensively. In addition, l know I feel more comfortable at some gear ratios and cadences but have never really tried to puzzle out a relationship beyond staring at Golden Cheetah charts and scratching my head.

[3] As it turned out (because of my first ride) I always ended up stopping my effort at 20 minutes, waiting 30 seconds at ~0 power and then ending the test. I’m sure it’s not quite what the test expects, but it’s worked reasonably well as a consistent protocol for me, and it’s something that I know I can make myself gut out now.

[4] If I’m lucky enough to really get my FTP up then I’m sure I’ll need 2-3 gears, but I won’t try “running the links” or anything like I used to assume was needed


I am at a loss for words… :anguished:. I’m thinking people should just push really hard until failure in ramp tests. Probably better to not give it much more thought than that and put all your energy into the actual plans.


@MI-XC I certainly didn’t mean to stun anyone, I was trying to inspire more hope!

The realization I came to was that it’s better (for me mentally) to get a 5% bump every 4-8 weeks (with a very small bit of stress from the test) than suffer once and never want to do it again and then train less effectively.

If I can encourage you to just try 3-4 of them, I bet you’d have plenty of words (or at least data). You shouldn’t even have to start at the “bottom” like I did, take Nate’s advice from one of the pod casts and try something more like a binary search. Just start on your low ring and about 3 cogs “down” and see how it goes - i.e. just try to follow the line.

The worst that can happen is you discard the result, but I bet after doing it once or twice that way (don’t go till failure, just go till 20 minutes) you’ll see that the RAMP is viable!

1 Like

I’m kinda excited for my next ramp test on Saturday, but with a 1x12 MTB (and a KKRM Trainer I have to shift) its a little different than your 2x setup. I like my cadence around 85ish , So when I start on the first step Im in whatever gear that matches the power target with my cadence around 85 and I will not shift for the first 5 steps. I will hold that same gear till my cadence tops 100 then I will shift up a gear and repeat till the end of the test until the end when I’m killing my Quads.


@dlove - that makes sense, especially with a 1x - I could see using 2ish gears for a “real” RAMP and I’d just over estimated the level of shifting required.

It’s interesting to me (and feels validating) to me that you don’t try to reach for a 3rd gear towards the end. I’m curious (a) how long your test generally runs to and (b) what your sort of “final cadence” is when you reach the end.

Nice work, Dr. Strangelove


Not sure if I’m understanding this right. You are on a dump trainer (so no ERG mode) and you keep in one gear for whole test… surely you must either start with v low cadence or end with crazy high cadence (or both!)??? I usually change gears every 2-3 steps to keep cadence in my usual range


Agreed. Confusing to me, considering he is on a standard trainer.

What I think he means is not shifting the front chainring, and only shifting the rear cassette, but I may well be wrong?


OP, I think you might be overthinking things a little.

Your FTP is just a representative number to help guide the structure of your training, taking the test a dozen times just so you can map out the perfect approach won’t help you with your training. It’s a ton of impressive data you have collated there but unless you take that methodology and apply it to every workout you are going to struggle IMO.


That’s the very definition of over thinking something that is basically very straightforward. Ride, then ride but harder and harder until can’t ride anymore.
My brain actually aches after reading the post.
But, each to their own and more power to you. But I just don’t understand a thing you’ve said.


and @mcneese.chad sorry I wasn’t clear - I don’t shift at all neither chain or cassette.

As you mentioned I used to change gears every few steps (and end up bombing my gearing choices) so I was surprised by the results too - but I haven’t found that it’s been necessary.

1 Like

Totally agree, this was really designed to help me get over the problem of training with structure, but a poor representative number. As I mentioned, I’d effectively had 4 tests over 2 years cause I didn’t want to retest when it was time again. Mentally it was just too much of a barrier for me.

So I figured before I started a new training block (going back to SSB1) I’d just conquer my fear of the FTP test itself so that when I do use that representative number for initial training I’d be ready to retest to an accurate new number.

That’s how I started to stop worrying if my “not shifting” gave me “the number” (i.e. something I was going to be stuck with for a year because I was too scared to retest) nd realize that it was “a number” and that if I wans’t scared to test so frequently than I’d find enough small percentage gains over time to “zone in” on the right number with a whole lot less pain overall (i.e. small pain repeated frequently vs. mostly avoiding one big pain).

@SJB definitely - to each their own. I just wasn’t able to ride and ride harder until I gave myself some sort of marker to know when that was… that’s why for me mentally setting my test limit at 20 min seems to work. When it’s the open ended “ride the RAMP till you crack” … well I just don’t feel mentally strong enough to have that open ended goal.

So what I think I effectively did was use the “level up” gearing process to set some “way points” for me to know what’s reasonable and reachable.

Impressive analysis! You clearly are driven and are willing to put in the time to improve, and I think that’s awesome.

If I could make one recommendation, it would be to remind you and everyone that the goal of the ramp test is to estimate your Functional Threshold Power, which at the time of each test has a “correct” value based on your physiology. Therefore, the goal of all your analysis should be to find a way to take the test that is repeatable AND gives you an FTP result that is as close as possible to that true physiological FTP. Over-unders are a good way to evaluate the “correctness” but I think over time you start to get a sense for what riding at threshold feels like.

So, if you feel like there is a mental block preventing you from digging deep enough during the test, then by all means do what it takes to overcome that. It looks like that’s what you have done! But also remember that an optimal value in this ramp test quest is not necessarily the highest number, but the most useful number for training.


thanks @madman2 I agree on the recommendation. I think the “don’t stress over the gears” has really helped me find that repeatable way to test - basically I can manage a gear at the right cadence to get me to and through an FTP level, and now I’m starting to see where that progress is tapering off (e.g. I’m not too stressed over the difference between 169 FTP and 172, my next predicted number) so I feel like I know I can go do a training block and be ready for the next “retest” (both at the start of the block and at the “end” of it, i.e. the beginning of the next).

And if I reach 172 then I know I’m still trending higher but if I significantly exceed that predicted value than I can feel very confident that the training was solid and it wasn’t just that my previous number was too low - because my previous number was reached through a repeatable protocol (not that the RAMP isn’t but implying that I was able to figure out how to estimate my FTP consistently vs. floundering)

Couple of questions,

What were your maximum power figures per test in date order, or is it shown above and I’m missing it?

How did you factor out the training benefit (FTP bump) from your results?

1 Like

In my 8 minute tests my FTP was always around 200 (maybe 198-202) and I think I was really stuck by the mental block. You can see the FTP numbers from this “test of RAMP tests” in the first chart column G, but in order they were; 62, 75, 91 ,113 ,122 ,132 ,138 ,142 ,149 ,158 ,163, 169

Some of the tests I had a 1 day rest before hand (usually I’d do 2-3 tests then need a day off) and I’m sure I could have got past the 20 minute mark but I was trying to keep my own personal protocol consistent.

To answer your second Q (about factoring out the FTP bump from training) that was easy - I hadn’t trained for any of the FTP tests. These were all in the span of two weeks, so the only training was the tests themselves. I’m certain they gave me some fitness, but I can’t imagine it’s anything but marginal in that span.

@mcneese.chad and @RobertSims - I was thinking more about your cadence question last night and am still not sure gear shifting is required based on the slope of the RAMP curve (at least not yet for me - maybe when I’m +250).

Here’s my cadences graphed out;cadence_chart

As you can see, in almost all cases they’re able to track right in line with the RAMP expectations, though you can see some like the 34x24 was really an outlier.

I’d be curious for you or others to try a “fixed” test and see if it was more cadence familiarity or actual gearing requirement. I can totally understand wanting to keep a consistent cadence and using gearing to do that, I just realized (from my first test) I wasn’t able to do that yet, thus the experimentation.

1 Like

Ugh, the exported graph is a little ugly - you should be able to access the spreadsheet directly via the link;

I think you slightly misunderstood, I start at a baseline of roughly 80-85 cadence at a gear that matches with the expected wattage target, keeping with the wattage targets till my cadence tops 100 then I will shift which generally drops my cadence back to the 80-85 area then builds back up following the wattage targets till I hit a 100 again, shift repeat process till I just cannot match the wattage targets…complete test…

When your cadence is topping over a 100 in most cases you are forcing your heart rate way up, same thing when you are smashing your quads attempting to grind out wattage targets instead of using your gearing to making it easier on yourself…

1 Like