for any geeks out there, open to page 100 of Dr Skiba’s Scientific Training for Endurance Athletes and big surprise, the TRIMPS HR intensity weighting factor is based on HR vs lactate curve. Big surprise because Coggan was inspired by Banister’s 1970s TRIMPS, and so the intensity weighting factor for power is based on power vs lactate curve.
No I don’t think so. Both Coggan and Skiba give reasons why its not. And why physiological cost is better approximated by power and ftp/cp (or pace and critical speed for runners), and its all built off the Banister TRIMPS HR based work from the 1970s.
There are a few variants of TRIMP, though the one you’re most likely to see in software is the exponentially-weighted form:
TRIMP = sum(t * h * 0.64^(k*h))
where t is time in minutes, h is fractional heart rate reserve, and k is a constant (1.92 for men, 1.67 for women)
This has some nice mathematical properties that TSS doesn’t. The derivative w/r/t time is just
dT/dt = h * 0.64^(k*h)
so individual bits of the workout just sum / integrate naturally.
You could certainly define a “normalized heart rate”. For an activity, compute the TRIMP for the whole activity and then find the heart rate N such that you would get the same TRIMP if your heart rate was a constant H for the whole workout. That is basically analogous to normalized power.
I never said that average HR did a good job of describing an interval session, but that doesnt mean normised HR is the answer and it will be the same as ave HR.
Also No, It really isnt that variable and isnt evidence for the need for Normalised HR (you might think it is but mathematically it is not.)
Just look at Normalised Power for the intervals Pwr/hr for the intervals. I would suggest this is what most people do and was commented on above.
What would you learn from Normalised HR? Nothing that makes more sense than established metrics.
If dont believe me take the HR data channel and Normalise it, I suggest you will learn nothing meaningful.
I hear what you’re saying but I still think it would be useful to have a sense of variability. We keep using the term “normalized”, and whereas I think that is overkill, I think the spirit of the OP question is give me a sense of how variable this ride was compared to other rides. (in terms of HR). You get that via power measurement, but that’s load, not strain.
Power measurement is fundamentally different than HR (not the least of which is the scale), so our use of the term “normalized” is what is bugging me the most. We don’t need normalized HR necessarily, simply because there are simpler ways of addressing the overall need/want.
So I guess I just disagree that it’s not “variable enough”.
I could imagine learning a lot. Avg. bpm is 141. Suppose normalized bpm is 150 (or even 145). Given what others have said - that the variability index will be close to 1 - small fluctuations are meaningful. That is, maybe as we used avg. vs normalized heart rate, we would learn that a 10 bpm difference for a ride means it was a wildly variable ride.
Now, suppose you are a TTer without a power meter. Could meaningful insights not be learned? One could argue to look at heart rate time in zone, but that might not represent what I did in a meaningful way. One hard effort that raises heart rate is different than going too hard up every hill on a course with rolling hills.
Scales are arbitrary. All this means is that small differences in avg. vs. normalized heart rate would have a different interpretation than small changes in avg. vs. normalized power.
This is simply false. I just created a descriptive metric that captured the variability in HR. Do I know if it relates to any physiological stress? No. But I know that it is possible to create such a metric which could then be validated through research. Happy to share some figures if interested.
You could certainly do some interesting things. You could normalise the heart rate into a scale 1-10. Then use that to give the session an overall RPE as well as graph the RPE second by second.
These 30-sec repeats focused on my anaerobic capacity and required days and days of recovery (my coach said a nice goal for the next time would be 10x600W ).
But compare 4 to the others if we define internal strain as Std Deviation and HRavg… using those two measures the Anaerobic Capacity workout is more similar to the 2 hour Easy Endurance than the others!
Your example wont happen (but if it did, using a weird formula not like the one for power then also see 3.)
There are already better established metrics which tell you more, NP amd VI etc already tell you if its a variable ride and you can just look at the TiZ graph and HR graph etc (like I said solving a problem that doesnt need to be solved)
The answer to the OPs question, as I have done several times is it makes no sense like it does for power
You can calculate it in WKO5 if you like if you reallly think it of benefit to you (but I cant find a chart where anyone has because I suggest it is pointless)
I dont have anymore to add to the discussion, except if you think it is of use just calculate it in WKO5, Golden Cheetah, you might be able to do so in intervals.icu. (I did years ago, when having a similar discussion and found it was a waste of time, hence I know you only get 1 - 2 bpm different to average HR, you might need to raise to the power of 8 not 4 like the NP calculation to get it to work)
I’m just saying there is no ‘need’ for normalising HR and that is why its not used, there are better ways of looking at it, many already established metrics, I don’t even think you need to put a number on if you can just look a the HR trace in real time plus the standard measures
So don’t do that. In fact, don’t use HR anything for an Anaerobic Capacity session.h
Find a hard 80 mile gravel ride with 11 hills that you either raced or went hard. Also, rest stops, mechanicals, train crossing.
Now find the solo training ride of roughly the similar duration. Only one nature stop. Same answer? If so, I’ll concede, but find a few examples where HR is actually informative.
Also, I’m not saying it would be a measure of strain. It would be a measure of the variability of strain (e.g. VI is not a measure of load)
I’m going to pull out some stuff from Intervals and compare. Intervals has a “HR Load” which is hrSS (normalized TRIMP), a form of TSS but uses HR rather than power.
I believe one of the primary reasons for TRIMP / HR TSS, and Power TSS, is to support charting out “dose and response” and taking into account the fact that activities with more higher intensity work should be given more load credit, at least in the context of dose and response.
Using the 4 workouts above, plus this week’s 90-min Wed group ride
Workout
Duration
TRIMP
HR Load / HRSS
Power Load / TSS
Work>FTP
W’bal Delta
2 x tempo + 20min TT
120-min
205
119
135
39kJ
21.3kJ
1 hour at 90% ftp
97-min
180
102
119
12kJ
5.8kJ
Easy endurance
122-min
163
90
98
14kJ
6.1kJ
Anaerobic capacity 2x 8x30-sec(180-sec)
116-min
175
102
161
126kJ
10.7kJ
Wed night ride
97-min
104
78
86
42kJ
8.9kJ
Edit: posted TRIMP the first time, adding a column for HRSS.
Going to be honest, the TRIMP score does nothing for me. Not much difference between easy endurance (163 HRSS) and anaerobic capacity (175 HRSS). Hmm. The HRSS more closely aligns with TSS, except with the anaerobic capacity.
Provided I copy&pasted that Intervals data correctly, it seems power based TSS does a better job at ranking my perception of recovery required. I believe one of the primary reasons for TRIMP and TSS is to chart out “dose and response”
Re: the unstructured Wed night ride… There was some wind and I mostly rode 3rd or 4th wheel behind two bigger guys.
Garmin rated the Wed night ride as “Impacting Tempo: this activity had enough effort at tempo pace (medium to high intensity) to improve your muscular endurance.” and benefit of “This activity enhanced your ability to maintain a moderate pace for a longer amount of time. It increased your high aerobic training load.”
Here is the 30-sec power with tempo zone in yellow:
where it clearly looks like riding at tempo for long stretches of road. Zooming into 31-minutes where it wiggles around the yellow, its 0.85 IF with 1.05 Variability Index.
In this particular case, I think Garmin nailed the analysis of that unstructured ride.
And FWIW here is the power and HR zones for that portion of the ride:
I’ve got 2 rides in mind, 3.75 hour solo ride with IF 0.87 and about 37-min of stops (2). And 6 months later a 3.5 hour group ride with 0.91 IF and about 15-min of stops (2). Both around 290 TSS. Both around the same ftp (260). Similar average temperatures in the upper 60s.
How are you suggesting I compare them? It needs to be something I can pull out of WKO or Intervals or GoldenCheetah. I’m interested in seeing if there is something HR related
They both felt hard, but in different ways. My two examples there is very little time at or above threshold HR. Neither are races as with your example. The question I responded to was about longer events. The solo ride had more time at higher HR zones, while the group ride had more time at higher power zones. But over 3 hours, either the power or HR stress scores are somewhat the same.
Going back to the original poster’s question…
I don’t think of 1.2 VI as very steady, but ok.
Above threshold times:
power: 18-min solo vs 9-min cyclocross race
HR: 0-min solo vs 37-min cyclocross race
I guess your point is that your ~43-min cyclocross race shows higher HR than the power would suggest, when compared to a ~60-min “seems like 11-min time in threshold power zone” workout.
Maybe that suggests racing tends to skew HR high for other reasons?
).
