Improved fitness decay estimation

A suggestion on improving the decay algorithm.

The “Optimal Decay” setting seems to reduce my fitness signature more than needed. I have experienced the following several times: Training camp, lots of hard climbing. Then I come back to the normal life and ride less both in terms of time and intensity for three or four weeks without digging deep. The whole signature declines, especially TP and HIE. Then I do a couple of hard rides again, and suddenly the TP and HIE is approximately the same as before.

Since this is a repeating pattern, maybe Xert can tune the decay parameters individually so that the fitness signature declines more appropriately for that rider? I know I can activate “No Decay”, but fitness will always be dynamic.

2 Likes

No decay is still dynamic… moves in line with load, just doesn’t decay relative to load

Why does the user even need to set decay at all? I find this to be a very confusing part of Xert and I don’t feel it is explained well.

1 Like

Decay is necessary to track your fitness continuously over time, whether you’re training or detraining.

An example of why this might be important/relevant… imagine you do an FTP test or get an Xert Breakthrough the day before 7-10 day vacation or illness. Would your fitness stay at that new value after a week (or more) off the bike? Not likely. If you were to return to training using that previous value, you might be performing workouts at an incorrect intensity. With Xert, you’ll see some decay of your fitness over that time. And if the decay is too much, then it should be (relatively) easier to generate a breakthrough and reset the signature.

The different methods work like this (in a very basic/dumbed down way):

  • No Decay - Training Load Matched: Xert predicts exactly what your signature is as your train & detrain (generally achieving a BT on No Decay is quite difficult).
  • Small Decay: Signature just below what Xert would predict since last BT
  • Optimal Decay: Signature lower than what Xert would predict since last BT
  • Aggressive Decay: Signature much lower than what Xert would predict since last BT
1 Like

Thanks for the reply. I understand why decay exists, I just don’t understand why there are different options that it is on the user to select.

Surely a “good enough” AI knows what my decay should be? Ultimately, each option of subjective and could (if I understand correctly) lead to an incorrect signature.

FWIW I’ve found that Small Decay seems to most accurately predict my fitness, both during the indoor training season and outdoors.

Everyone has different needs and their training regimen will vary season to season.
The Decay setting is another tool you can use to dial in your training.

Signature Decay Method – Xert (baronbiosys.com)
Xert’s Magic Setting – Signature Decay – Xert (baronbiosys.com)

If you’re a competitive rider or the type who likes to constantly challenge themselves, Optimal Decay will provide frequent signature validation on your XPMC chart. IE, clusters of BT event dots of varying size and color.

If you rarely generate max efforts, indoors or outdoors, consider Small Decay once you have dialed in your signature (which you can do with BT workouts).
If you specifically want signature changes to vary by TL only (once signature is established), then No Decay is your ticket. For example, extended Base phase training without any HIIT efforts.
You will be warned when no BT events are detected in over three weeks. Time to consider a max effort during a ride or a select a BT workout indoors.
In these scenarios stepped dots will appear on your XPMC profile chart or TP will closely parallel TL if No Decay is used.

In all three scenarios you need to generate max efforts at some point for signature validation. If not happening organically (group rides, hill climb efforts, sprints, ZRL races) you can force the issue with a structured effort in as little as 10 minutes.

If you have periods of varying scenarios, that’s what the Decay option is for.
The more you experiment with Xert’s knobs and dials, the better it gets. :wink:
As for the OP, Small Decay will likely provide what they are looking for.

Reference –
Breaking Through the Xert Way! – Xert (baronbiosys.com)

1 Like

People are different. Situations are different. That’s what settings are for. One size fits all rarely really fits anyone.

I consider No Decay the ‘correct’ setting because it most closely approximates the signature without maximal efforts. I hate maximal efforts. I’m not a competitive person and I’m a 0-1.5 stars none-athlete or really a couch potato who occasionally gets the urge to move for a while. Xert works astoundingly well even for me. Once I established a signature that I really don’t even fall below when I don’t move at all for half a year.
With the default Optimal decay my signature falls to zero or almost and that’s not even true for a couch potato and really tough to correct if you have not moved for half a year and need to start slow but sensible and fun. I also hate endurance workouts since I’m a sprinter by nature and very easily bored. So I like the workouts with some orange/red and some blue in there in shortish (few minutes) blocks and the XSSR blocks that get ‘easier’ when you get ‘groggier’.

If there was only Optimal Decay I would have quit Xert a long time ago since that is just not feasible for me, doing a maximal effort every few weeks and after a while of not riding at all having to do more than one with less then optimal outcomes just to start again with something resembling useful values. I hate slope mode on the trainer as well, so I try to do the few maximal efforts I do outdoors and that’s only possible when the weather permits and I am in the mood for riding outdoors: even more boring, no TV shows or movies to watch.

I’m strange, people are different, and I’d rather decide for myself what I like than have it dictated by an AI. Those are in my experience so far mostly very stupid and only cater to the least common denominator. Xert by the way does as far as I know not use any kind of AI, just clever algorithms.

My Garmin AI for example usually tells my around 1 or 2 PM with about 500 steps on the watch so far that it is proud of my for closing in on my step goal of 7000 steps. It does this nonsense because I usually don’t move all day (home office computer work) and then take a walk in the park after sunset for 6000 steps or so. So the not moving until evening is part of the pattern it detects and so it predicts by me not moving that I’m close to reaching my goal because that is what usually happens.
Really useless and :roll_eyes:

1 Like

Some decay works best since minor data quality issues or small errors could wreak havoc on your training since you may end up with a signature that is unattainable otherwise. Optimal and Small Decay work well since they give you the opportunity to reset or prove your fitness on a regular basis. This also helps catch possible chronic issues you might be experiencing that reduces your fitness outside of your training (think illness, chronic dehydration, etc.)

Optimal and Small are the best ones to use in general. No Decay if you know what’s going on and can manage things on your own or if you are planning on going through an extended period without attempting to get a breakthrough. Remember that leaving it on all the time could lead to an inflated signature if your data isn’t the best.

2 Likes

Thanks for the answers everyone.