School Me On mFTP (modeled FTP)

I’ve never paid much attention to it and honestly (even after listening to Hunter Allen and DCRainmaker) don’t understand how it’s derived, and how far in the future I could apply the value? Do you even apply it in the future or is it meant to supplement sFTP between tests?

The reason I’m interested is for an upcoming TT that I’d like to pace a little better. Perhaps the mFTP can be used in this way? I don’t know. Current sFTP is 304 (TR ramp) and mFTP is 332 which IMO is a lot higher! Why such a discrepancy?

Any advise appreciated.

1 Like

#1. mFTP is based on a rolling 90 day model of power data. It tells you what you’ve done, not what you can do. It’s a black box model based on a sample set of around 250 athletes that Coggan has studied, but the basic premise is that it’s using an inflection point as well as some other magic to derive a FTP at a particular time to exhaustion as well as your ability for anaerobic work capacity at Pmax.

#2. Your model needs valid maximal data in order to be accurate. You’ll notice in the offseason (or if you have taken a few weeks off) your mFTP will either be way off or you won’t be able to hit your numbers. You generally need a short (5-15 second), a medium (2-5 minute), and a long effort (35 minute and up). Use the “Coggan’s Normalized Residuals” chart to find points to hit informally or “Best Times for Informal Testing”.

Once you have a valid model, you can then use Optimized Intervals to target the adaptation that you are looking for:
Intensive FTP for increasing power
Extensive FTP for increasing TTE

I’ll use the Optimized Intervals numbers to tweak my TrainerRoad workouts slightly.

For pacing, you can somewhat use the PDC, but it’s not really great for that since again, the model is descriptive rather than predictive.

PM me with your PDC and your optimized intervals charts and I can help walk you through some other stuff.


ok check PMs. I added a note…thx!

@stevemz You clearly know WK04! Love that product! and always wondered (but never investigated) as to how mFTP was calculated and, similar to @Landis, I have seen wild fluctuations in it but didn’t know why. Thanks for your explanation!

FWIW: When I didn’t have an FTP test, I used 0.95*20min on PDC curve and knowing where/when that data point was derived, I adjust as needed. As an example, my initial Ramp test when I started with TR was way low (and I dug deep), so I threw it away and used the WK04 approach. Since then, I have gone back to the 20min test.

1 Like

Huge thanks to @stevemz for the help last night. I had 2 workouts in 2018 that completely messed with virtually everything you can think of displayed in WKO+. I knew there was a spike somewhere but, didn’t realize how much it could mess up all the data…he made a couple suggestions and through me fumbling around trying to figure it out I’ve learned more in one day than the previous number of years. The optimized interval chart is now something of value!


Nothing to add, just thought I’d throw in an interesting WKO4 article I found by Hunter Allen. Enjoy!

WKO4 Case Study: From Beginning Junior to Category 1 Racer


I also use wko4 and love the mftp. I questioned TR why they can’t achieve close to the same value with the ramp test and they don’t have an answer. Coggan’s work really isn’t that black box as it is mathematical modeling based on the inflection point that was mentioned earlier. The more good data you have, the better the mFTP will be. Also as noted, it is a reflection of what you’ve done. So if you’ve done a bunch of base work without maximal tests, your ftp will be underreported. If you’ve done a bunch of vo2 stuff, you’re likely to be more accurate. Realize this is also based on the ctl/atl/tsb rolling model. TR ftp is a 75% calculation on a one time test.

Nothing is prescriptive, but you should be able to derive a good guess based on what your current training load is and how to pace going forward. Best bike split is also helpful here in testing runs at different IF.

1 Like

@AndrewL great post! While using a PM and WKO+ for years until this fall all I can say for myself is I wasn’t really training and wasn’t using the software effectively. Basically, I was riding a lot of TSS with relatively little structure. If you can see my calendar, prior to November all that riding is LSD, racing and group rides. No structure. Same amount/strategy year over year going back to about 2004-05 with a brief amount of structured coaching in 2008-09.

Thanks to TR that has changed and I’m actually shocked at what I’m seeing. And huge thanks to @stevemz for helping me analyze my data in WKO+. Specifically, I had two bad data spikes that absolutely threw off everything I was looking at.

Thank goodness for this forum as I’d still be blissfully unaware and not making sound training decisions. Not that it’s the be all end all but, if I’m going to spend this much time and effort to collect the data might as well have good data to tweak going forward.

Anywho, thanks to many here and the TR staff for a great product. The proof will be in the puddling so-to-speak when racing starts.

I’ve looked at the WKO4 stuff but have not taken the plunge. I’m still using the single-point general FTP to frame my training. Out of curiosity, what are some (real life examples) of the more powerful & actionable tools WKO contains? Basically, if I start on WKO will I see a noticeable upgrade in my training results, especially w/o years of data to draw from?


That’s the real power of this forum!


@Captain_Doughnutman if you are a “data junky/nerd” then WKO+ is for you. But, I’m proof that unless you have good data (realize it) and can analyze it and finally adapt it to your training then it’s for you.

It’s almost endless the things you can learn. Whether they actually matter or not is another story. With that said, the power profiling, strength/weakness and now as I’m learning mFTP and optimizing intervals (to me) is useful.

At the end of the day, doing the work, will go most of the way for me. Analysis paralysis is something to be wary of with WKO+.

I hope others with more knowledge will chime in to give you more info.

I’ve found that if I have a proper length maximal effort in the PDC, the Ramp Test matches my mFTP very closely (within 5 watts).

My current system is to knock out a few sprints and a 1 minute effort a week before, do the Ramp Test, and then go do a long progressive effort at the FTP that TR gives me from the Ramp Test and hold until exhaustion.

At least for me, the numbers match up really well, and then I can fine tune some of the TR workouts with the Optimized Interval guidance (usually just swapping out workouts to have longer interval times for the same TIZ for sweet spot and higher powers for the Vo2 and Anaerobic work)

Top reply @stevemz

As you reported,I think that proper length efforts are the key. This is something that wko4 values over the last 90 days that aren’t factored in during the ramp test on TR. It’s interesting that you seem to be doing something to maximize your wko values a week in advance. Why do you do this? This is interesting as when I’m in the middle of my training, wko models a higher ftp than TR reports following ramp test. As a triathlete, I’ve long wondered if the TR ramp test is inherently biased towards short period high power versus lower steady state power. I seem to find myself at about 78% of the 1 min power versus the 75% that TR places me. This can be a difference of at least 10 watts or more.

I’m focusing on XC and plan on tracking the normal TP benchmark numbers over the course of the season since most are pretty relevant:

  • 5 second (for fun and technique)
  • 1 minute (relevant to XC)
  • 5 minute (relevant to XC, I usually just take the 5 minute number from the Ramp Test since it’s close to max)
  • 20 minute (mostly for vanity and because I have a high watermark number from last year. This is the least relevant tbh)
  • 60 minute (in reality I just focus on FTP+TTE since my TTE tends to be on the lower end so I’m trying to extend it out. Eventually I’ll get TTE out close to 60)

The reason why WKO4 models your mFTP higher than the Ramp Test is because as a Triathlete you probably don’t do a ton of max efforts in the short duration part of the power curve, so as a result your FRC reports artificially low and the curve is flatter.

If you search for the chart “Best Times for Informal Testing” and apply the 90 day right hand filter, it will give you some guidance on what durations to do some maximal efforts to help feed the model.


This chart is suggesting a 1 second, a 40 second, and a 40 minute effort as the largest deviations from the model. I have a 30 second recent effort and a 35 minute recent effort, so I’m pretty comfortable with where I’m at and don’t plan on doing any more feeding anytime soon.

1 Like

Whoa. I thought I was a data nerd… but I’ve never seen this one. I just downloaded it into my WKO4 Charts Library and am giving it a look. Yikes, there’s some meat in that one!

(Is this new for WKO? I don’t remember ever seeing it in the library before)

They just pushed it in this past week. It was one of the Facebook users that programmed it.

1 Like

love this chart, thanks.

1 Like

William seems to know as much about WKO4 formulas as Tim.

1 Like

This is an amazing thread. Appreciate all the great information. My question is this: when are you comfortable with your mFTP and your TR FTP? My mFTP is currently 333 watts but the ramp test that I performed today was 324 watts. Want number would you guys go with? My instincts say to go with the TR ramp test number but I don’t want to “miss out” on training by training too low.

333 and 324 are within margin of error basically. Call it 330 and move on with your life.