For the past four months in a row AI FTP Detection has moved my FTP up by ONE watt. I’m following a mid volume masters plan and it’s starting to become pretty frustrating. In reality, I feel my fitness has taken a real boost in this time frame, I’ve set numerous power PBs from 30 seconds to 30 minutes and seem to be riding stronger than ever but AI FTP Detection does not reflect this at all.
All of my rides are synced to trainer road with power data and I am responding to survey for every one.
I’ve gone from 231 watts to 235 according to TrainerRoad but my gains appear much bigger in reality. I guess I could be overestimating them
Anyone got any ideas what might be happening here? I suppose I should probably do a ramp test…
If you really want to know do a long 20-60 minute test or Kolie Moore test (these are available in TR as custom workouts) but not a ramp test. If you want to have an even better idea, do the long test and then get WKO5 and do your short and medium residuals after the long test
Doing a ramp test will likely just give you a bad data point.
Nice long warm-up. Start an interval about 20 watts below where you think your FTP is. Ramp the power up over a few minutes to what you think threshold is and keep on until you can’t hold that power any more.
If your interval lasts over 40 minutes and less than 80 then the average power for the whole interval should give you a good number for your FTP.
You already know the answer. For 3 months you’ve been frustrated with the result from an algorithm but haven’t done a manual test to see how your body performs? As others have said, don’t do a Ramp test, do a long form test like the Kollie More. Then you’ll know what you can REALLY produce. Light it up!
It’s possible to significantly increase your fitness without significantly increasing FTP. FTP is basically your medium length (35-80 min) power.
If you increase short term power (0.5-5 min) your true FTP might not budge (your ramp test FTP might get artificially inflated though) but you will sprint and get up short climbs much better.
If you increase your long term power (time to exhaustion / fatigue resistance / durability), you will be fresher after a long ride and more able to reach similar power numbers as at the start of a ride. This is often the biggest fitness gain you can make.
Also, you can have gotten fitter but you’re just too fatigued to really perform your best. If all the data the model (any model) is being fed is only from a heavily fatigued state or none of it was close to an all out effort, your fitness won’t show up in the model until you recover and feed it some max efforts.
So take some recovery if you need it, then do some max efforts of 5s, 30s, 5min and a KM FTP test. Also go for a really long ride at a race pace if it matters for your goals. (Spread these efforts out over several days). You will have a very good PD curve them and can really track fitness by making sure you have those kind of max efforts every 4-6 weeks.
What you don’t mention or give information on is if you feel the workouts being served up are appropriate or not. A 2% FTP increase in 4 months for a seasoned rider is not untypical (I would kill for 6% per year myself).
I my opinion, the only real value of FTP is to set the levels of your workouts appropriately. When you do an assigned V02 or threshold workout, do you consistently feel like you have more in the tank after, and are you answering the surveys appropriately? If not, if those workouts feel “right” then why look to boost the FTP number to a point where the workouts are unachievable?
If you’re feeling stronger and setting PBs, that sounds encouraging! As other athletes have mentioned already, FTP is only one measurement of your fitness – it certainly sounds like other areas of your fitness are improving based on your description above.
How does your current FTP feel? Does it feel about right for a power target that you could hit for ~60 minutes? If so, it might still be set properly for you. I like what @Chadvt mentioned – if your workouts still feel appropriately challenging, you might not need to change anything up.
If not, power testing is always an option – and it provides AI FTP Detection with more data it can use for your future predictions as well.
What I’ve found after tracking AI FTP vs performance for the last year is that AI FTP matches actual performance FTP (20 minutes outside after riding for at least 1:30) only when I’m not making large gains.
Perfect example was my latest FTP pump. AI FTP said my FTP was still 210 watts, while my 20 minute effort was 231 watts (both average and NP), and that came after 1:30 riding including a 6 minute effort at 223 watts average. So performance tested FTP was 220 watts or 10 watts higher than what AI fTP predicted.
I’ve been using the 220 watts + iLevels in WKO5 for 3 to 5 minute VO2 max repeats, and the 220 feels spot on
I don’t really get the 1-2W bumps in AI FTP either. I dont bother accepting them and just carry on. Its well within trainer accuracy %. I’m surprised that TR even bothers, maybe they should set some sort of % threshold which holds off FTP changes unless significant?
Don´t you think, with all the PB, that your training is right on track? With whatever number TR uses to create your plan
I, personally, find more joy in seeing PB´s than a number used to establish training zones.
I think it’s just part of the gamification of training that gets people addicted. It’s better for sales to increase people a watt a month than 5 watts every six. (I’m not suggesting this is unique to TR, btw)
Having said that, I agree that a watt or two is margin of error and I only change things in 5 watt increments. Too much hassle to change and memorize the new numbers in multiple systems every few weeks.
You have most likely increased your TTE (Time To Exhaustion) which is a great thing. FTP will usually hover around the same point while TTE improves. Later you will get more FTP improvements. You are probably doing very well at the moment.
