Improved W'bal implementation

[Mark Liversedge]

Dave Waterworth has reformulated the W' bal formula to remove the computational overhead associated with the integrals.

This means the calculation is extremely fast and completed in a single pass, it also means that displaying W'bal as you ride in the train view (or on a headunit via Garmin IQ) is now easily performed without needing significant CPU power. It also removes some of the loss of precision that my performance enhancements introduced (mostly means we see a slightly wider range of recovery and depletion).

We still offer 2 approaches; Integral (as per the literature with Tau) or Differential (no Tau, implicit from expenditure) 

Dave's reformulation is summarised here: http://markliversedge.blogspot.co.uk/2014/10/wbal-optimisation-by-mathematician.html

I've also added labels for the matches to make them easier to scan when reviewing a ride.

Both will be in 3.1 SP1 later this year.

Cheers

Mark

[Pete from AUS]

So that stuff about the uneven time steps could mean when you have ride time in sequential seconds but real time in a time stamp (e,g, iBike files) that ride pauses (traffic lights etc) could be handled neatly (ie factor some recovery) ??

[Mark Liversedge]

Hi Pete,

Yuck. I didn't know that iBike files were like that !

I hope we use the real time when we import them, do you have an example file with gaps I can use to test this ?

Mark

[Pete from AUS]

Cheers Mark, file sent.

In regards to the CP & W' models, has there been consideration of whether CP and/or W' would deplete during activities?

If CP is your sustainable power level, as you fatigue, is your sustainable level not reduced?
So if you maintain a power level above initial CP, as your W'Bal reduces, your CP is reduced and the rate of depletion would increase?
Eventually, if an athlete was utterly exhausted would their CP at that moment in fact be effectively zero?
The reverse would apply as W'Bal recovers, CP would recover with it.

In regards to W' mostly I have read about setting the initial value and the recovery rates.
However during longer exercise do the models assume that you can recover fully to the initial value or does the W' limit get reduced?
Given enough recovery (days) of course you get the full amount restored (or even more if the work pays dividends).

My understanding of recovery rates is that the model assumes we recover faster the deeper we dig into W'Bal.
But do recovery rates remain constant or vary (slow down) the longer an activity continues?
IE If each interval uses 75% of W'Bal do we recover as fast after the first one as the fifth, sixth etc?

Look forward to answers to help me understand it better.

Cheers
Pete

[Mark Liversedge]

On Monday, 13 October 2014 12:02:02 UTC+1, Pete from AUS wrote:

Cheers Mark, file sent.

In regards to the CP & W' models, has there been consideration of whether CP and/or W' would deplete during activities?

I have the same questions as you on this !

Whilst the different PD models we have implemented generally do not take into account fatigue. The Extended Model looks at longer durations and then adjusts accordingly whilst the Veloclinic models have 3 different approaches for accounting for fatigue. So far the Extended Model will underestimate longer durations if you don't have any decent efforts of a longer duration where, with most of the data I have seen, the veloclinic models always seem to underestimate.

For intermittent exercise, my understanding is that W' bal will reflect depletion / reconstitution of W' and tau may or may not take fatigue over the ride into account, since it is computed by looking at the entire ride (but, I don't really know). Also, since slow-twitch muscles do not fatigue as quickly as fast twitch I expect that by the time you're sustainable power (reduced CP) has dropped due to muscular fatigue your W' has also gone anyway (Again, I am speculating). 

Hydration and nutrition (bonking) I guess are assumed to be managed by the athlete ?

It would be so cool to identify rides that had impacted your ability to recover CP/W' during the ride (if that is even valid) -- and would they signal greater adaptations or fatigue systems so much they impeded post ride recovery (I have no idea).

We just need to wait for the clever guys researching this stuff to tell us !

Mark 

[Mark Liversedge]

On Saturday, 11 October 2014 18:55:15 UTC+1, Mark Liversedge wrote:

Dave Waterworth has reformulated the W' bal formula to remove the computational overhead associated with the integrals.

This means the calculation is extremely fast and completed in a single pass, it also means that displaying W'bal as you ride in the train view (or on a headunit via Garmin IQ) is now easily performed without needing significant CPU power. 

I've added W'bal as you ride to train view; there is a dial and it is also shown on the erg plot:

Cheers,

Mark 

[Juan Benjumea Pablo-Romero]

This is my first time writing in here, so thank you for having this open google group. I have been using for some years the Golden Cheetah software, but I am a bit lost with the changes. I am struggling with the W' and the recovery rate.

I know now what does W' mean and how to estimate it (conceptually) after reading some of the threads. However I am struggling to find how to measure the recovery rate against time and the fraction of power under CP.

I would like to try to do a workout that uses a certain percentage of the W', then recover during a certain period of time until I replenish the W', then do the next effort until I use the same amount of W' etc. ¿could it be possible?

Thank you in advance