From recent efforts 801
The median on the curve would give you a W’ of 15 Kj with that PP. But, you need to be sure it’s really your best all out effort…
Hello David
Thanks for that reply, just check this ride earlier it had significant dips in graphs where efforts are high something similar to HR graph in inverse, but now it’s different i.e. more or less flat.
https://intervals.icu/activities/3702621853,
If check this activity, it’s a continuous climb and i was totally empty when i stopped. however the W’ shows me almost fully recovered…
https://intervals.icu/activities/3712148582
Thanks
Not fully clear the value of “-. 6.6”! If this is - 0.66% or -0.66kj and a hard effort, it’s probably not worth worrying about, similarly if - 6.6%. If - 6.6kj then agree with those saying you probably do need to update W’!
1 Like
Looking at those you only spent a little time over FTP on the first which is why your W’ didn’t drop much. The 2nd looks more like it. Your W’bal drops to 16 kJ (from 30). You should probably just go with 15 kJ for W’ as that would have been almost gone on the 2nd activity which matches your riding experience.
Sorry, I did see your message, didn’t mean to ignore you. Have edited to 19500 now.
Sorry for the interruption. To save some potential headache for folks who accidentally want to learn from this, the question was:
It certainly isn’t. {} in an expression is an object initializer, not an array initializer. Spreading an array instance into an object initializer is possible and does create numeric object indices for the object values alright, but the length property of the object is undefined, and any accidental existing length property on the object instance (i.e. by prototype pollution) would not be incremented.
In other words, it’s not iterable. This is a modern syntax version of a classic mistake. 
The idiomatic ECMAScript >= 2015 way to clone an array is
data = [...data]
Thanks. I didn’t know that. Just went to check I didn’t replace the slice() with {…data}
Sorry, jumping in late on this topic, but I stumbled upon this thread and thought I’d share a comment.
This is because W’ is only a capacity, and the CP/W’ models (2 parameter and 3 parameter) assume that the capacity is not rate-limited. In other words, according to the CP/W’ model, if your W’ is 20kJ, then the model predicts that you could sprint at 20,000 W above CP for 1 sec, which isn’t physiologically possible… you’re limited by some physiological Maximal Power. Because of this limitation, it’s often stated that the CP/W’ model isn’t valid for short durations (<2 min)…because they can’t account for how your W’ is rate limited (remember that power is a measure of a work rate).
At short durations, you’re not limited by the capacity of your W’, but the actual rate at which it can be expended. This is why you only expect to utilize 16.8 kJ at 2 min, but a full 19.2 kJ at 8 min. Something Xert has worked on is not only calculating your capacity above threshold, but also how you can expend it (via the metric Maximal Power Available - MPA).
Hope that makes some sense! Cheers!
10 Likes
@david I’m not sure if I’ve missed this or doing something wrong, but is there a way to auto populate the W J into settings, as this is constantly changing figure on power view (Mortons 3 CP); 42 days or this season.
2 Likes
I have that on the todo list. The best thing to do now is use the “For workouts (90d)” curve with the default model and look at the W’ for that. That’s should be a good number.
1 Like
@david previously you said to use Mortons 3P, Using W balance - #39 by david
The default is Mortons 3P plus fastfitnesstips?
A little while ago I added W’ estimation to the eFTP algorithm so you should use that. That wasn’t the case when I wrote the post you linked.
1 Like
Late to the discussion but saw this popup on my feed.
The question “is it useful” is the central question with any variable.
My personal experience has been - yes. I do a lot of Zwift racing and we know all Zwift races are hard, fast start. Often I have video recordings of the full race. In a recent race, I fell apart after the first 7 minutes from a fast B group on the most sadistic course made - Innsbruck KOM. The data in intervals.icu for W’ showed 0 for W’bal at the start of the 7th minute, meaning from 0-7 minutes, it was all downhill in terms of physiology. The video was nice because I could correlate with the facial expression, the in-game dramatics and everything
The races are not going to favor my physiology, but perhaps I can try to challenge the body to be better prepared for these efforts. I calculated the rate of fall in W’bal = 1.85 KJ/min. Perhaps the HIIT days in my training can focus on fast start type events that expend half of this rate to start out in an interval blocks of 3-4 minutes and then prolong the interval as the fitness increases. This workout might have several minutes of active rest breaks. I can even try manipulating the active recovery break durations by crunching them down.
Those were my thoughts on how I would apply this to my specific situation.
For long distance triathlon, I’m not sure focusing on the W’ aspect will help but sprint triathlons/duathlons often have fast starts.
DISCLAIMER
It is one thing to say I had nothing left because my W’ bal reduced to 0 and another thing to actually verify if that is indeed the case from a physiological point of view. Current understanding is that we cannot be certain W’ = 0J in the modeling should necessarily be an anchoring point for physical exhaustion. But if you can do a VO2max test and verify that you have reached VO2max at that point, then you can be sure W’ is empty. And that would also coincide nicely with exhaustion.
One has to be careful about the differences in what mathematical modeling says and what reality is.
5 Likes
Ahhh no wonder my chart is empty. Thanks.
Any implementation of the intervals.icu w’bal calculation for garming gps units? A real time display would be just awesome.
I’m using this one:
But there are others:
I’ve using Dr Skiba and tried to match the settings with the ones in intervals.icu but they differ quite a lot. I can put the Dr.Skiba negative with several short burst with rest in-between and just today i went negative in the intervals (with a 3min effort) but far from it in the dr.skiba.
Have a read here
and you will better understand why the W’ model has a lot of shortcomings.
It’s best to use a CIQ field that let’s you change between the 2 different calculation methods. Just try ou both and see which one is better adapted tou your specific profile.
To be honest, I don’t know which one @david is using in intervals.
1 Like