Negative w'bal from tonights Cragg Vale Hillclimb

[Russell Brown]

Hi,

Just reviewing my file from the Cragg Vale Hillclimb and I see that at about half way my w'bal goes negative, and keeps going more and more negative, until the end.

What does that mean? Is my estimated w' too low? My estimated CP?

Cheers

Russell

[Mark Liversedge]

You could use the effort from the Hill climb as a test, do another duration and see what the model predicts :)

Mark

[Nathan Townsend]

Hi Russell,

What is your estimated CP and can you let me know what you best power values are for 12-14min, 7-8min, 2-3min?

If W'bal goes negative it could be due to the CP value being too low, but the model is not perfect yet either, so this probably needs tweaking.

[Russell Brown]

Hi Nathan,
I haven’t tested deliberately at those times, but I can give you the “bests” from Golden Cheetah.

Estimated CP (from Golden Cheetah CP chart) is 319w
12-14minutes is 340 watts
7-8mins 350-348watts
2-3mins is 390-378 watts

However, I’m not sure why, but the w’ estimate has lately gone up to 16.7kj, and my intervals no longer show negative w’bal, in fact, I am struggling to get w’bal below 11,000 now.

Cheers

Russell

> --
> _______________________________________________
> Golden-Cheetah-Users mailing list
> golden-che...@googlegroups.com
> http://groups.google.com/group/golden-cheetah-users?hl=en
> ---
> You received this message because you are subscribed to a topic in the Google Groups "golden-cheetah-users" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/golden-cheetah-users/4jLqKG4egFU/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to golden-cheetah-u...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Nathan Townsend]

Using the following values:

2.5min @ 385W

7.5min @ 349

13min @ 340

Using the 2-parameter nonlinear hyperbolic model gives CP = 330W and W' = 8276 J (Std error CP: 0.8857, W': 214.8 J).  Those standard errors are super low so these values are probably solid estimates.

So this means your CP is likely set a bit too low.  If you can hold 340W for 13-14min then you can surely hold 330W for about 20-25min.   A practical way to test your CP in reality is to do an over/under effort at 10w above and below your CP estimate.  If you feel as though you're recovering at 320 following 3-4mins at 340W then you know you're in the right zone. 

I had a discussion with Inigo Mujka about the nature of "threshold" last week.  We both agree that "threshold" is FUZZY zone, probably about 10-15w wide.  It isn't an exact figure to the nearest watt.  IMO this value also changes slightly according to fatigue.  I'm not certain but I even wonder if it can go down and then increase again during a recovery period within a race?  The implications here for modelling are complex, but the short answer is that CP has a much bigger impact on W'bal values than W', but there will always be some uncertainty about the exact value of CP within that narrow range of 10-15 W.

> To unsubscribe from this group and all its topics, send an email to golden-cheetah-users+unsub...@googlegroups.com.

[Pete from AUS]

Surely therefore the extent to which ‘work’ relative to CP depletes W’Bal is a continuum?

Well below CP the effect is negligible but as you approach CP the effect is greater and increases through and above CP.

[Russell Brown]

Thanks Nathan, appreciate the reply.

I think those numbers are skewed as they are from a hillclimb. On the flat I can’t get close to 13minutes at 340 (unless I’m just not trying hard enough.)

So far, since I upped the CP a little in GC, I’ve had no negative w’bal readings, and I seem to crap out on intervals when w’bal gets down below 8kj, so the model looks good now.

Would love w’bal on my garmin when I’m training to remind me that actually, I can do this next interval.

cheers

Russell

> > To unsubscribe from this group and all its topics, send an email to golden-cheetah-u...@googlegroups.com.

> > For more options, visit https://groups.google.com/d/optout.
>
>

> --
> _______________________________________________
> Golden-Cheetah-Users mailing list
> golden-che...@googlegroups.com
> http://groups.google.com/group/golden-cheetah-users?hl=en
> ---
> You received this message because you are subscribed to a topic in the Google Groups "golden-cheetah-users" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/golden-cheetah-users/4jLqKG4egFU/unsubscribe.

> To unsubscribe from this group and all its topics, send an email to golden-cheetah-u...@googlegroups.com.

[jkoch6599]

I asked in another thread, and then found this one so I'm asking again and I apologize.  Are you using the equation t = W' / (P - CP) ?

When I plugged these numbers in Excel, used that equation, and then minimized the squared error using Solver; I get CP = 328W and W' = 9165 J.  I can't figure out what I'm doing wrong.

[Nathan Townsend]

On Friday, 10 July 2015 18:58:30 UTC+3, jkoch6599 wrote:

I asked in another thread, and then found this one so I'm asking again and I apologize.  Are you using the equation t = W' / (P - CP) ?

When I plugged these numbers in Excel, used that equation, and then minimized the squared error using Solver; I get CP = 328W and W' = 9165 J.  I can't figure out what I'm doing wrong.

Yes but re-arranged:

P =  (W' / t) + CP

In this case P = y-axis and t = x-axis.  

Rearrange your eqn and try again. Does it make a difference?  If not maybe send me the excel file and I can take a look. There might be something to do with the way you're set up the curve fit or something??

[Tony]

Why not use Work = CP*t + W'. Y axis = Work and X axis = time. Y intercept = W' and gradient = CP. Work is of course Watts x secs.

No curve fitting required, just a straight line through two points e.g 3 and 15 minutes. Use 4 or 5 points as recommended by Nathan if you like i.e best efforts at 3,6,10 and 15 min.

[Ale Martinez]

Other mathematically equivalent linear model would be P = W' * (1/t) + CP, which also doesn't require curve fitting and can be solved by regression. The problem is they don't necessarily agree on the estimated parameters from the same tests, as shown by Bergstrom et.al. 2014 paper:

[jkoch6599]

I used this equation, fit using Solver in Excel, and now my numbers match yours exactly.

Now I need to think through why the same model fit using two different but equivalent equations would give two different parameter estimates...

[rjkflyer]

Hi, I too have an issue with W':

The Graph for W' on the Ride page shows a W' maximum of 20kJ.

The Model box on the CP page shows 13.5kJ

On a continuous ride, I can get W' down to -11kJ.

I have done the linear Excel line fit and my CP from that is close to what GC model (and all other tests I have done) says, and W' comes to ca. 20kJ, but even this appears not to make sense as I am going so negative??

Can anyone help me? Thanks!