Training load Fitness not right on Mobi?

Post a question about the software or find answers here.

Training load Fitness not right on Mobi?

Postby slacker2012 » Fri Oct 12, 2018 8:42 am

I don't think the training load and performance analysis under Health on is working very well. It's Mobi's implementation of a performance management chart right? Fatigue is equivalent to acute training load, Fitness is chronic training load and Performance is the difference between them aka training stress balance.

The time constant used for Fitness appears to be far too aggressive. Fatigue and Fitness track each other too closely. Fitness is acquired over a longer term, should have a much longer time constant and be less sensitive to day-to-day changes appearing smoother.

The problem is demonstrated in the first chart on Sportracks own page which shows the effect of immediate cessation of training.

First of all, Fitness should be smoother, not following exactly the same sharp day-to-day variation as Fatigue. Fatigue falls off rapidly, almost to zero within 30 days as you'd expect. However Fitness falls off at an almost similar rate, dropping by ~80% in the same period. You should only expect to see fitness drop by a half in a month.

A commonly accepted time constant for fatigue / ATL is 7-10 days, so you would expect it to fall to zero in a month. However Fitness is typically 40+ days, it should take up to 3 months to settle. Mobi seem to be using more like 15 days which is too short.
Posts: 3
Joined: Tue Nov 22, 2011 4:07 am

Re: Training load Fitness not right on Mobi?

Postby admin » Mon Oct 15, 2018 1:15 pm

The short answer: You can change the model parameters on your training settings page here: ... g=settings

Be advised that changing these parameters is an advanced option and your performance chart numbers will completely change if you edit any of these.

Longer answer:

You're mixing up concepts from two different models. Chronic/acute/stress - these are terms from the training stress balance model. Performance modeling in SportTracks uses the Banister model which has slightly different calculation and correspondingly different parameters, described below:

Fitness recovery: How many days does your fitness take to return to baseline after completely stopping training.
Fatigue recovery: How many days does your fatigue take to return to baseline after completely stopping training.
Fitness balance: A percentage (from 10% to 90%) that determines the model weight of fitness vs fatigue to calculate performance.

I'm not 100% sure how the default model parameters were decided when this was initially implemented but they probably need a revisit after 6 years, and your settings may depend on when you signed up. I can't recall the original paremeters Banister used in the research but if memory serves correctly it was around 50/10/35% ? You can play with these parameters if you like, but I'm not sure if they necessarily should match training stress balance.

EDIT: The current defaults are fitness recovery = 49, fatigue recovery = 11, fitness balance = 40%. This was changed at some point and we didn't change older users settings because it would throw all their performance data out of whack, so you may have the older settings. You can contact support for more assistance!
Get the latest info: Fan us on facebook or follow us on twitter
Site Admin
Site Admin
Posts: 2916
Joined: Tue Apr 05, 2005 8:52 pm
Location: USA

Re: Training load Fitness not right on Mobi?

Postby slacker2012 » Wed Oct 17, 2018 7:26 am

Thanks, I had no idea those settings were tucked away in there. Sure enough my fitness recovery was down at 20 days. Will have a play.
Posts: 3
Joined: Tue Nov 22, 2011 4:07 am

Re: Training load Fitness not right on Mobi?

Postby ksherman » Sun Oct 21, 2018 6:57 pm

Wow, me either. It explains a lot with those settings. So updating my values to reflect as advised will recalculate all my history?

Yes, it did.
And what a difference. Previously I simply could not reconcile the numbers, now they at least show a more logical pattern.I'm still not 100% on the fatigue never overtaking fitness. In 2016 I did a brevet over 3 weeks, approx 8 hours a day for 21 days albeit low intensity. In mobi with the updated settings I had peak fitness of 3,460 and peak fatigue of 2,375. That gives a pretty nice performance number of around 985, and of course it will increase over the following time, it actually peaked at 2,077. In reality I was buggered.

In the ST model I finished with (TRIMP based) CTL of 180 and ATL of 328, for a TSB of -148. By far the biggest hole I had ever dug. :) By comparison an Ironman finish had a TSB after of -54. And those big negative numbers feel right.

So my only remaining question would be how to evaluate over training using this model? In ST I happily push my TSB to about -25 before easing back a bit. I suppose I need to look back and see if there is a consistency in how it presents in mobi.

Any thoughts or inputs welcomed.
Keith Sherman
Garmin Edge 800, Garmin Fenix3
Trek Domane 2016
Specialized Roubaix SL3 '11
Specialized Epic '15
Posts: 243
Joined: Sun Jul 11, 2010 10:00 pm
Location: Australia

Return to Questions

Who is online

Users browsing this forum: No registered users and 1 guest