Questions on Best Way to Monitor CP and W' Response to Training Using Golden Cheetah

[Paul Anders]

I have some questions about the best way for me to use GC to monitor my fitness and track my progress.

I've been using GC for about a year. I raced 20+ years ago as a competitive elite masters 35 in the SF bay area. Today, while I say I'm a recreational, non-racing rider, I ride with a competitive group of guys and I train during the week as if I were still racing, working on climbing, intervals, tempo, and endurance riding. I've been averaging between 9,000 and 12,000 miles per year since 2011, about 250 rides per year.

Starting this year, I've been doing all my training on two road bikes with power measurement systems, which I've cross-calibrated using a wind trainer. I also have considerably history from 2015 of other power-monitored rides, so I have a reasonable database of rides in GC.

What is the best way that I can monitor my CP60 and W' using GC, if I depend solely on the rides I do for power vs. time data, and I don't do specific CP testing? I do a wide mix of maximal efforts between 1 minute and 1+ hours in the course of my regular riding, and I do a large number of specific, continuous, maximal efforts at about 5 minutes (climbing) and about 30-35 minutes (flat). 

I'm using the Extended CP model with default parameters, with a date range of 3 months. What I've been doing is to take the fitted values for W' and CP from the CP plot for my ride on the first Saturday of each month, and enter those data (along with Pmax) into the Critical Power table under Tools/Options/Athlete/Power Zones, to set my CP and W' for the month. The number I'm using for CP60 right now corresponds very well with my actual 60 maximal power, perhaps a few watts higher I've actually achieved on a ride. On rides where I am doing repeat maximal 5 minute efforts (e.g. intense hilly group ride), I see W' Expended at around 110%, indicating I may have W' set a bit too low. 

Does the method I'm following make sense? Is there a better way to monitor my CP and W'? Is my goal of doing this without specific testing realistic, or do I need to incorporate testing, and let the results from testing determine my CP and W'? If so, what are the recommended test protocols and data extraction methods? 

I know it's a lot to cover, but I've reviewed many postings here and in the docs on GC and I'm still unclear on the best procedure. Thanks in advance for taking the time to read this and respond. 

[Christian Wiedmann]

Note: this is not an expert opinion. I thought I'd reply because I'm in a similar boat to you, except that I've converted to mostly doing indoor riding/racing (on Zwift). I usually do two online races a week and do Z2-Z3 riding three more days a week. I find that I often get to near-maximal efforts in the races, but rarely do efforts from fresh to exhaustion.

I find that the numbers that the extended model gives me given this distribution of rides are not a very good description of my fitness, but I use them to get a ballpark idea of what the numbers might look like. I then use the W' balance model to set my CP, assuming that my W' is not changing much, usually rounding to the closest 5. While this is not a very scientific approach, I find that it does give me numbers that feel reasonable to me.

As an example, since I had a break-through week this week I've just gone through the process of revising my CP estimate upward. At the beginning of the week my CP was set to 295 with a W' of 22 kJ. On Wednesday I did a short race that resulted in my using 127% of my W' balance, spending about 4 minutes at a negative W'. Based on this and GC's calculation of a 307 minimum CP, I set my CP to 305. On Thursday I did another race with a sprint at the end and this time I hit 111% of my W'. Given this, I'm inclined to increase my W' estimate to 24 kJ. The extended model estimates a CP of 297 and a W' of 20 kJ for me right now. Using these numbers would result in hugely negative W' balances in my harder rides.

Bottom line is I feel like unless you're doing real testing, the numbers from the models are probably not directly usable. However, the W' balance model allows you to test the numbers in hard rides (i.e. ones where you hit W' balance of close to 0) to make a reasonable guess at the right parameters.

As to testing, here is a post I saw recently on a CP/W' testing protocol: https://groups.google.com/d/msg/golden-cheetah-users/lhCAZoRvgw8/NiUhdro8AgAJ

[9606]

I just want to be sure I understand terms correctly.
CP is that power one can sustain for many hours? You can sustain 297 +/- for many hours?

[Christian Wiedmann]

This is probably a good place to start if you want to understand the terms better: http://www.goldencheetah.org/#section-science

To answer your question, my understanding is CP should be sustainable for 30-50 minutes until fatigue that is not part of the model starts to appear and pushes performance below CP.

[Mark Liversedge]

On Saturday, 21 May 2016 04:21:44 UTC+1, 9606 wrote:

I just want to be sure I understand terms correctly.
CP is that power one can sustain for many hours? You can sustain 297 +/- for many hours?

CP is the delimeter between heavy and severe exercise -- that tipping point between sustainable and unsustainable effort.

Whilst in theory you should be able to maintain that level of effort indefinitely it is a tough ask. Generally, most folks see a drop off from critical power at durations from 30 minutes out to an hour. The extent to which this drop off is physical or mental is an open question.

Mark

[massarob.info]

This could be helpful to have an idea of "typical" CP time-domain sustainability
https://www.researchgate.net/publication/235762602_Time_to_exhaustion_at_and_above_critical_power_in_trained_cyclists_The_relationship_between_heavy_and_severe_intensity_domains

[Nathan Townsend]

On Friday, 20 May 2016 22:35:06 UTC+3, Paul Anders wrote:

I'm using the Extended CP model with default parameters, with a date range of 3 months. What I've been doing is to take the fitted values for W' and CP from the CP plot for my ride on the first Saturday of each month, and enter those data (along with Pmax) into the Critical Power table under Tools/Options/Athlete/Power Zones, to set my CP and W' for the month. The number I'm using for CP60 right now corresponds very well with my actual 60 maximal power, perhaps a few watts higher I've actually achieved on a ride. On rides where I am doing repeat maximal 5 minute efforts (e.g. intense hilly group ride), I see W' Expended at around 110%, indicating I may have W' set a bit too low. 

If W'bal goes negative this can occur as a result of the following:

1. W' is too low

2. CP is too low (so what happens in this case is that modeled W'bal depletes faster and recovers slower than it does in reality)

3. The models are not yet robust enough to perfectly predict the point at which task failure occurs ie: when W'bal = 0 should be equivalent to the "limit of tolerance".

In general over a 60 min duration there is some fatigue which is not accounted for by depletion of W' ie: if you estimate sustainable power for 60 mins using the 2 parameter CP model, then this estimate is higher than actual power.  What that means is that you were unable to fully expend W' during the 60 min effort, or rather than staying 100% fixed, CP actually declines (or a combination of both).

Some literature on central fatigue during prolonged endurance activity indicates that type I fibers become less excitable.  If this is true, then CP would decrease in proportion to the decrease in motor drive to type I fibers.  Conversely, there is data in severe hypoxia which shows a direct effect of cerebral hypoxia on central fatigue.  This appears consistent with my own data that shows W' decreases significantly ONLY in severe hypoxia about approximately 3500 m.  Put together, these results suggest that if central fatigue is greater over longer duration than short duration, then a small decrease in both W' and CP is likely to occur.  

The uncertainty in how much fatigue occurs or what proportion of that fatigue affects W' versus CP means that it becomes a bit of a guessing game if you use CP60 to estimate "threshold".  You are better off to follow the guidelines I stated in the post which Christian linked to. At least we know that the CP estimate in this case has been physiologically validated as representing the boundary between heavy and severe domains.  Nothing else is validated properly. eg: FTP from a 40kmTT or the "extended" CP model in GC which AFAIK has never even been published in print. So you just can't be certain those threshold estimates are valid.

What I think is that if threshold is estimated using extended duration methods eg: 60 MMP or 40km TT, then it tends to slightly underestimate the true threshold and this is due to the effect of long duration central fatigue.  It makes the W' value look higher than it really is, but the decrease in CP has a bigger effect than we realise and despite the fact that W" is high, over time it still becomes realitively "easy" to drop W'bal into negative territory.  Some level of glycogen depletion likely contributes to decreased W' and the central fatigue probably also contributes to a reduced ability to fully deplete W'. 

Personally, I prefer to stick with knowns rather than unknowns.