Incremental Power Measurement

[quote="Die fietsryer"][size=3][font=Times New Roman][font=Times New Roman]Is there any method available that measures the mean power for whatever interval the measuring device is set, multiplies that value with a MPC factor and sums all the values to give a representative way to compare two training sessions?[/font]
[/QUOTE]

Yes, see here: http://www.cyclingpeakssoftware.com/defined.html

The link does not really give any information.

What is the algorithm that is used to calculate Normalized Power?

dome said:

The link does not really give any information.

What is the algorithm that is used to calculate Normalized Power?

Algorithm goes:
-take power samples
-rase those power samples to 4th power
-take average of those 4th power power samples
-take 4th root of the average of 4th power power samples
-this is the normalized power

Example:
You have been riding
4 minutes with 250 W and
6 minutes by 300 W

Average power is (4*250 + 6*300)/(4+6) = 280 W

Normalised power:
250^4 = 3 906 000 000 = (a)
300^4 = 8 100 000 000 = (b)

Average of (a) and (b) = (4*a + 6*b) / 10 = 6 422 000 000 = (c)

4th root of (c) = 283 W
and this is the normalized power.

dome said:

The link does not really give any information.

What is the algorithm that is used to calculate Normalized Power?

For the complete details, go to the link at the bottom of the Power411 page, i.e.,

http://www.cyclingpeakssoftware.com/PowerTrainingChapter.pdf

Thank you sidewind.

Does this mean that the algorithm assumes that the MPC has a function of Power proportional to the inverse of Time to the power of 4?

What fietsryer is looking for is a measure of a training session relative to what the MPC shows is possible.
So, if we call the points on the MPC, say, Maximum Power Time Units (MPTU's), then in fietsryer's post 1000 mins @ 100W = 1MPTU ; 200 mins @ 200W = 1MPTU ; 10 mins @ 500W = 1MPTU ; and 1 min @ 1000W = 1MPTU. (And for any power output there is a corresponding time that that power can be maintained for.)
So, if the power measurement increment is 1 second, and at the end of a session we measured, say, 2000 seconds @ 100W, 500 seconds @ 200W, 200 seconds @ 500W, and 50 seconds @ 100W, then our total MPTU's would be 2000 seconds/1000 mins + 500 seconds/200 mins + 200 seconds/10 mins + 50 seconds/1 min = 1.24 MPTU's.
This does seem a much better way to compare training sessions.

sidewind said:

Algorithm goes:
-take power samples

-smooth power samples using a 30 s rolling average

sidewind said:

-rase those power samples to 4th power
-take average of those 4th power power samples
-take 4th root of the average of 4th power power samples
-this is the normalized power

dome said:

Thank you sidewind.

Does this mean that the algorithm assumes that the MPC has a function of Power proportional to the inverse of Time to the power of 4?

What fietsryer is looking for is a measure of a training session relative to what the MPC shows is possible.
So, if we call the points on the MPC, say, Maximum Power Time Units (MPTU's), then in fietsryer's post 1000 mins @ 100W = 1MPTU ; 200 mins @ 200W = 1MPTU ; 10 mins @ 500W = 1MPTU ; and 1 min @ 1000W = 1MPTU. (And for any power output there is a corresponding time that that power can be maintained for.)
So, if the power measurement increment is 1 second, and at the end of a session we measured, say, 2000 seconds @ 100W, 500 seconds @ 200W, 200 seconds @ 500W, and 50 seconds @ 100W, then our total MPTU's would be 2000 seconds/1000 mins + 500 seconds/200 mins + 200 seconds/10 mins + 50 seconds/1 min = 1.24 MPTU's.
This does seem a much better way to compare training sessions.

The TSS (Training Stress Score) calculation principles are defined on those pages already linked in this thread, I suppose. In TSS, one hour ride on the FTP level is defined as 100, and other powers and times are compared against FTP.

However, it should be noted that e.g. short intervals like 1000 Watts for 2 seconds are consuming other energy sources than 250 Watt ride of two hours. So at least I don't see why very different training types should be compared against each other at all. rather compare intervals to earlier intervals, and TTs to earlier TTs.

this is non-sense - i fear that anyone who reads this post will have a lower I.Q. as a result

dome said:

Thank you sidewind.

Does this mean that the algorithm assumes that the MPC has a function of Power proportional to the inverse of Time to the power of 4?

What fietsryer is looking for is a measure of a training session relative to what the MPC shows is possible.
So, if we call the points on the MPC, say, Maximum Power Time Units (MPTU's), then in fietsryer's post 1000 mins @ 100W = 1MPTU ; 200 mins @ 200W = 1MPTU ; 10 mins @ 500W = 1MPTU ; and 1 min @ 1000W = 1MPTU. (And for any power output there is a corresponding time that that power can be maintained for.)
So, if the power measurement increment is 1 second, and at the end of a session we measured, say, 2000 seconds @ 100W, 500 seconds @ 200W, 200 seconds @ 500W, and 50 seconds @ 100W, then our total MPTU's would be 2000 seconds/1000 mins + 500 seconds/200 mins + 200 seconds/10 mins + 50 seconds/1 min = 1.24 MPTU's.
This does seem a much better way to compare training sessions.

dome said:

So, if the power measurement increment is 1 second, and at the end of a session we measured, say, 2000 seconds @ 100W, 500 seconds @ 200W, 200 seconds @ 500W, and 50 seconds @ 100W, then our total MPTU's would be 2000 seconds/1000 mins + 500 seconds/200 mins + 200 seconds/10 mins + 50 seconds/1 min = 1.24 MPTU's.
This does seem a much better way to compare training sessions.

Uh oh, 1.24 MPTU's. That means we accidentally exceeded the MPC by 24% somewhere along the way. I guess our training is working, but now do we calculate our MPTU's based on the 'old' MPC curve, or the new one that we just created on this ride? And if our MPTU's are being referenced to a constantly changing MPC, then how do we compare one ride to the next? Oh dear, now I'm all confused. 😕

I guess I'll just stick to TSS. 🙂

How long does it take to recover from a MPC effort? Is is unrealistic to think that short duration sprints up to MPC level can be repeated in one training session?

TSS does average the training effort quite drastically, and if you have measurements of your effort every second or so, why not utilize that information by comparing each increment with a known performance measure such as MPC?

dome said:

TSS does average the training effort quite drastically, and if you have measurements of your effort every second or so, why not utilize that information by comparing each increment with a known performance measure such as MPC?

Because, quite simply, powermeters are capable of recording changes in power output far more rapidly than to which our bodies can respond. For example, I can pedal for 15 s at 500 W followed by 15 s at 50 W repeatedly for 20 min, and my respiratory, cardiovascular, neurohormonal, metabolic, and perceptual responses will be very similar to those resulting from pedalling steadily at 275 W - all that really varies are the motor units called upon to perform the task. However, stretch the on period just ever so slightly (e.g., to 20 s), and the situation will be quite different. To have any merit at all, any system for "scoring" a powermeter file must explicitly recognize this aspect of physiology - which normalized power, from which TSS is calculated, does, by applying a 30 s rolling average to the raw data before performing any further calculations.

dome said:

How long does it take to recover from a MPC effort? Is is unrealistic to think that short duration sprints up to MPC level can be repeated in one training session?

Depends on where on the curve you're trying to hit. A 30sec - 1min max effort can be recovered from within 10 mins, but a 2hr max effort may take a day or two. That alone illustrates why hitting the MPC curve at 1 minute should not produce an equal number of MPTU's as hitting it at 2 hours. How could you possibly compare the two 1.0 MPTU rides?

You really should consider reading through cyclingpeakssoftware.com...ningchapter.pdfcyclingpeakssoftware.com...ningchapter.pdfOpen ↗ , which gives a great explanation of the physiological factors involved and how they relate to the development of Andy Coggan's Training Stress Score (TSS) system.