Another FTP discussion - 1.5 years on TR

Looking at these eye-watering numbers… I wish I had an FTP of >250W. Just for the fun of it, to get further on the bike trips. :pleading_face:

1 Like

Change to low volume and do 3 sessions a week on the trainer, not outdoors. Be consistent, be truthful on workout surveys, just follow the process - I usually get +5-7w every 4 weeks. Commutes don’t count, just go easy on those, no KOM chasing. One decent long distance ride on the weekends (~100km/4hrs). I usually also chuck in a 2 hour zone 2 outdoor ride, where I try to maintain z2 as much as possible.

I’ve gone from 300 (Feb-20) to 350 (Feb-22), but life has got in the way (currently 332 after being as low as 319). My body fat has been harder to shift than my ftp…


Interested by some of the replies above.

Particularly for my own situation (hope Op doesn’t mind) - I’m struggling to find a custom-generated plan that seems to fit the following template I have in my head

  • Quality >2hrs Zone2 three times a week
  • One intense ~1 hour session (trainer or group ride)
  • One long group ride (usually Saturday, hard, 3+ hours hilly)
    Total 10-12 hours.

Am I right in thinking that a lot of the people who responded above do their own planning on TR and essentially use it for the pre-defined workouts and maybe FTP auto-monitoring?


I’m on a standard Adaptive Training Plan Builder Plan and tweak it to suit my need. My plan has 3 HIIT sessions as default I drop the middle one to a Z2 session (through Train now usually) and ‘bolt on’ my weekend group rides. I’m not too familiar with them but I think what you might be after is a polarized plan. Find the one that’s closest to what you want and tweak it. I don’t know how they are made up but it could be a LV polarized plan gives you 1 HIIT session and 2 Z2 sessions (you just add a 3rd through Train Now) and you could bolt on your group ride. Or it could be a MV polarized plan gives you 3 Z2 sessions and 2 HIIT sessions, in which case you could substitute the group ride for one. My preference would probably be to go for the LV plan with bolt ons.

1 Like

I do my own sessions - when I joined 5 years ago I did SS base HV - the older really hard version - got me fit - and tired. Tried SPB HV - made me ill. I then did a hybrid of SS base mid volume - to much intensity. I now do my own version where I follow the SS sessions in the mid volume plan (including the under/overs) but only do 2 days of the intensity. - the o/u and one of the SS (usually the easier one!) I also do a gym session and a couple of runs/week as I was a runner. Then I switch to some sub threshold stuff once/week (95-99% FTP) in the spring and then the super threshold/VO2 stuff at 108% - like Red Lake and variants. I fill in the rest with the very consistent endurance rides that are 70-75% FTP - Cumberland/ Gibraltar/Phoenix -3 etc or ride outside if the weather is good at 60-70% FTP…if I have done an interval session the day before the zone 2 is easier - like Collins -1…I like the TR sessions but need 1-2 intensity sessions/week - rest zone 2. I haven’t noticed my FTP going up much - it’s about 270-275W @61kg for a road bike TT - (although I can get 300W on a ramp test but I don’t trust that - I just use my most recent race performance in an hour long TT)…but then I’m nearly 55 so not going down is fine by me! :laughing:


I guess philosophically, if you get to the point where you have the head-space to notice your prescribed plan is not ticking your boxes, it implies you have the capacity to self coach a bit :crazy_face:

1 Like

Thanks @PhydomiR - yeah I think this is where the issue is.



Yes that is a good point on rating the workouts harder leading to less burnout, thanks the point!

Yeah, I thnk I’m there now : )

Thanks for the praise Zack

However as much as “TR” is adaptive I still think it could be much better. I’m assuming you guys have many Petabytes of data that you can mine data through and with those insights provide better suggestions and line users up to specific goals.

i.e. I am currently at point X and I want to get to point Y. I suspect that you have many users in your database where the person is 41 years old, weighs 80kg is 5’11’’ and they have moved from point X to point Y (i.e. FTP 250 > 300).

What type of riding did they do? You guys have all our rides this can be analysed. i.e. obviously not all TSS is the same.
How much riding did they do?
How many months / years did it take?
How consistent were they?
Did they lose weight?

Say my goal is just pure FTP gains (Yes I know FTP is not everything) I am 40 with kids career etc, I just want to know how I get there. Surely TR would be able to plot my progress against a synthetic user (based on a population of real data) and tell me how I’m doing and where it thinks I’m going wrong OR maybe I’m doing better than the synthetic user.

Now I am not saying my marginal FTP gains are TR’s fault. It is just my realisation that I can’t do TR in ‘autopilot’ mode along with ups and downs of normal life. It seems I need to come to the fourm to get recomendations on where things are going wrong.

What I am saying is that TR could be a better all in one solution.


That the current model is a better revenue model for TR where plans just continue to roll on with no definition of reaching a final objective.

Anyways it would appreciated if you can pass this back to your Dev team. A fellow member passed me this interesting site If they want to review it.



1 Like

This is a crazy wish list for AI to be able to do. Search history and life scenarios of other riders, create a workout program based on that data, tell you why you’re not improving and then make adjustments based on how good or bad your day was at work. You’re asking a computer to be able to think like a human. Maybe this is the future but seems a lot to ask right now.

At least machines can’t time travel. Yet.

1 Like

This is hardly AI, creating a synthetic user that is based on basic information from a dataset should not be that difficult.

I dunno…I think it all seems reasonable, and along the lines of what what trainerroad themselves market the AI thing as doing right now. I’m sure not all the specifics have been touched on, but in general it has sure been marketed as a universal catchall alternative to one on one individualized coaching backed by millions of data points/rides/skynet/etc.

1 Like

IF riding were the only variable that mattered to results, then sure, that would be entirely reasonable and achievable, though it would certainly take a lot of analysis and data mining.

But given there is SO MUCH influence from other factors – build, fat level, training history, work type/shift/stress, family circumstances, genetic potential, nutrition, sleep, to name just TEN – I don’t think in the current reality it’s at all achievable to get that kind of analysis. In the future, sure, but also with a great deal more information on each rider. Not just the rides.

Right now, it’s nearly impossible to predict with any level of accuracy whether a rider will or will not reach a particular goal they set for themselves. Using ANY tool, unless you pretty much know everything about the person. That’s what coaches do.

Re the profitability of TR’s business model: IMO hardly anyone wants to just reach a given goal, then stop training. The reality is that most people who cycle and sign up for TR, don’t want to do so for a single endpoint in time. We all intend to train and ride for years and decades, so the training continues though the goals may change. It’s not about TR pushing plans, it’s simply the reality that this is what people want.


I understand this very differently. Right now, TR AI looks at just your individual rides over the last X weeks, and uses it’s analysis of a large dataset to try to guess whether you can handle a harder ride next time.

That’s it… there’s nothing in there about goals or long-term progression, as far as I know.

1 Like

The long term goals for TR & AT are driven by the training plan in use. It is largely subjective with workout selection & progression aimed at event types & their associated power level/duration demands.

It lacks specific objective targets like FTP, Power/Mass/Drag or other harder numbers, which some would like to see & use. I find those goals unrealistic and prefer the general specificity approach TR and other trainer programs use. We are usually training for event demands of some form, which are commonly dynamic in practice where those hard number values may not tell the whole story.

As such, I think a broader focus (per TR & others) serves a better purpose broadly speaking. Not to mention the inevitable issues if forecasting something that can be influenced by so many variables over the timescale of a typical training plan.

and works on changing the ramp of progression of workouts. Workout progression is prioritized according to the training zone goals of each TR defined training phase. You or Plan Builder are putting the TR training phases on the calendar.