Sufferlandrians using Xert to track fitness?

My thought is that if you pace yourself according to your model, then in the middle of the MAP or FTP effort, you can adjust your pace depending on how you feel. Conceivably, even the AC effort might be done better than do as hard as you can and try to hold yourself together.

I cannot try this yet because my next full frontal is in a month, although tomorrow is Half Monty.

I finally modeled the FF with Xert, and it approximated what my latest results on FF and HM were. I plan on using it to pace my next FF effort. If you are interested I can give you the details.

1 Like

Yes, I am interested to have your details :face_with_monocle:

There are two parts to the model. The first part is fixed and only has to be done once, the second requires adjustments because your fitness signature changes over time.

Fixed Part:

  1. Create as many intervals as there are in FF. Each interval is a no-repeat, no rest interval. The length of each interval is the same as in FF.
  2. For the two NM intervals I set the power to 100% Peak Power (PP)
  3. For the two AC intervals I set the power to the limit of reserve power - 0% Reserve MPA.
  4. I initialize all the warm-up, recovery, and cool-down intervals to their FF suggested values. They may be modified in the second part.
  5. I do not set anything for the MAP and FTP intervals. They are set in the second part.

Variable Part - Set the pacing depending on the current fitness signature. This has to be done before every FF, presumably the day before so you have the most up to date values.

  1. Adjust the power on the intervals before the first NM interval to ensure you have full MPA available for the interval.
  2. Adjust the power on the intervals before the second NM interval to ensure you have full MPA available.
  3. Adjust the power on the intervals before the MAP interval to ensure you have full MPA available.
  4. Adjust the power on the MAP interval so that at the end of the interval the output power matches the MPA available, essentially a breakthrough effort.
  5. Adjust the power on the intervals before the FTP interval to ensure you have full MPA available.
  6. Adjust the FTP interval so that at the end of the interval the output power matches the MPA available, essentially a breakthrough effort.
  7. Adjust the power on the intervals before the AC effort so that you have full MPA available.
  8. You then note the resulting power levels for each interval and use that as your guide for FF.
    You have to do the steps in the second part in order because the changes propagate through the model.

The only real variables are the MAP and FTP levels. I suppose you could lower one so that you could get higher values on the other. It would seem based on the coaches’ advice you would not do this.

1 Like

Yeah if you deliberately sandbag your MAP interval, then you will likely over-estimate your FTP. Unlike a traditional 20 min FTP test, 4DP uses your full 20 min average power without the usual 5% reduction because you have just smashed yourself in the MAP interval before. Clever idea as long as you do actually give it everything on the MAP section.


For what its worth, I use Xert to set my SUF 4DP metrics, and haven’t done FF in years - instead I get regular Xert breakthroughs from racing (and A Very Dark Place) and set my 4DP from the Xert power curve as follows:
FTP = 1 hour power
MAP = 5 min power
AC = ideally this would be 1 min power, but I use 1:30 at the moment. I did try 1 min, but this took me to a breakthrough at the end of the first set of intervals in Do As Your Told, and I couldn’t complete the second set.
NM = 5 second power

I find that these values keep me nicely dialled-in on the suffering, which I have honed over the years since the early days of doing Downward Spiral on a fluid trainer by RPE.


My plan is to do a couple of more FFs, and see if it matches my model. If it does, then I will decide what makes sense.

As you saw with AC, it may or may not match reality. Since I neither know the algorithms, nor the data sets that are behind Xert, I want to validate the model for my purposes. Also, I am new to Sufferlandria, and 4DP power.

Except FTP and 1hr power aren’t necessarily the same thing.

If you don’t want to do FF, I’d probably still test for the FTP number even if you’re using other values from Xert.

That’s super cool actually!

I did FF yesterday using my pacing plan based on my Xert model with mixed results.

The plan worked well enough for the NM/MAP part of the test reaching MPA = Power at the end of the MAP segment. I was almost at MPA when the MAP segment started. The NM estimate was pretty close, the MAP was overestimated.

The FTP pacing did not work out. There seems to be a relationship between MAP and 20 minute/FTP power that Xert does not capture. I was exhausted after the FTP segment, but the power level was nowhere near MPA.

The Xert estimate also did not come close to MPA = Power at the end of the AC phase, even though I had a 25% improvement there on the test, and my stomach was telling me to stop. I ignored it.

1 Like

I find Xert to rate FTP higher than The Sufferfest. During the 20 minute segment, were you holding the Xert FTP setting? If you were under the Xert threshold, then this would not result in a drop of MPA (by definition). This was my experience.

In my case (from my May 2020 Full Frontal), I had a breakthrough on my MAP, but was nowhere near it on my NM and AC segments.

1 Like

I had major breakthroughs on NM and AC (about 25% on AC, a little more on the NM). This was significant because it removed a major weakness from my last FF which was a low AC due to inability to recover from repeated efforts.

I could not hold my FTP setting according my modeled pacing plan, but on the test I did better than Xert’s estimate for threshold power. My MPA only declined about 30% during the FTP segment.

I now have a new weakness - sustained efforts. This result is not surprising despite my excellent MAP to FTP relationship. The training plan I was on did not emphasize MAP or FTP.

The problem with power duration curves as I understand them is that they use your best value (in a given time period) for a given duration. The FF is a classic counterexample. You can do far better on a standalone FTP test than on FF.

1 Like

Interesting… and well put.

I usually end up with an AC weakness on FF tests and have been frustrated with my slow progress in the past year, so I’ve been considering doing a NM/AC building blocks plan.

@Heretic What was your training plan like that it helped you to train your way out of your AC weakness? Would you mind sharing?

1 Like

Thank you.

I used the Couch to Crusher plan :The Sufferfest

It used to be called the Fitness Kickstarter, but there were other plans with that name.
It was the plan that Dan Llyod used on GCN.

Here is the link to the last video, but I believe there were others in the series: How Fit Did Dan Get In 10 Weeks? | Zero To ...? - YouTube

1 Like

I honestly don’t see the point of using a xede for testing purposes, especially for NM and 1min efforts. Your sprint is your sprint… you just rail it as hard as you can.

For the 1 min, you need to hit it as hard as possible, then sit to keep pushing, then get up and stand and push some more.

I guess I can see the utility of using it for the 5 and 20 min pacing efforts, but it’s a test. I think you should be using previous data for the 20 min to negative split if you need a target. The idea of Kohlie Moore’s FTP test is good because of this as well. (as posted above)

As an experiment, and a test of Xert’s modeling capability.

If it is a good modeling tool, it should be predictive. After all, that is ideally (as Hunter Allen, said in "Training and Racing With a Power Meter" what you want. After all that is how we made progress as a civilization by learning how, in the right situations, to abstract reality.

For fun, the other day I modeled Kohilie Moore’s FTP test that you have previously put forth. After my FF, I thought about trying it in the future.

In the end, the old saying applies: “If the map and territory disagree, believe the territory, not the map.”


I read that book for the first time over a decade ago at this point.

For me the hold up during testing is that you want to be able to suss yourself out based on how you’re feeling as well. The model is only as good as what you put into it. Unless you have lots of data including hard efforts during things like hard group rides or races, it’s unlikely it’ll know your true potential.

Like I said, I don’t see the point of using something like this for testing, but that doesn’t mean you can’t! It’s still an interesting idea in theory but for me it’s important to learn to pace. Additionally, I have a feeling that the data is going to underpredict if you don’t have a ton of hard info going in. My ‘feeling’ is opposite to what you encountered where the prediction was too high, but you still ended up with an improvement.

Still interesting.

Edit: The other thing is that I don’t think xert derives all the info from one ride. That’s what makes FF different and makes the comparison hard. You’re doing all these efforts in one go when doing FF vs data you’ve compiled from previous rides where you likely didn’t do that.
The model should allegedly be able to adapt to that but…

The ideas in the book have evolved since that first edition.

I do not necessarily disagree, I still plan on doing FF, and putting all my cycling sessions into Xert. I would rather do the Sufferfest workouts than the ones in Xert. I trust the current state of the Sufferfest training plans more than the Xert advisor.

You always have to test a model, but eventually the cycling models will improve. At some point, Xert may have enough data about me. What I do not know is how good the model is, and what their improvement plans are. Nonetheless, compare sports analytics now with what they were even 10 years ago.

Of course there are things that analytics will not be able to model in the foreseeable future: motivation, pain (physical and mental) tolerance, etc.