What is your W' bal(/anaerobic capacity) in Kj? Did you work to change it?

I’ve read in several places that there is an “optimum” W’ for a male road cyclist in the 14-18Kj range

Very well-trained male cyclists do well with W’ values in the low to mid 20s. For females, optimal W’ values are around 11 to 15kJ for moderately trained and 14 to 18kJ for well-trained cyclists.
Critical Power for cyclists explained | Better than FTP? - BikeRadar

In my own training data (intervals.icu and Xert) I get a W’ of 33-36Kj, which seems like a lot/too much. My strength has alwas been 20sec-2min efforts so that correlates well with this.
I have limited time to train and focus on aerobic base (6h/wk) so I’m wondering if:
a. I should try to bring up my aerobic base at the expense of W’
b. It is even possible on 6h/wk

Any opinions and personal anecdotes appreciated

As I understand it, as your ‘FTP’ or aerobic ability comes up, your FRC/W’ comes down. I don’t see a lower FRC as a problem, more of a indication I am expressing more aerobic ability at that time. I think it would be unusual to have a high FTP and FRC (relative to the athlete). Interested to follow this thread as I am ~4w/kg @ 80KG and my W’ usually hovers around 17-21Kj

1 Like

We can all benefit from a bigger aerobic base, but I suppose one of the key questions is what are your target events (if any) and how does your current W’ profile limit them (if at all)?

1 Like

I target 2-2,5h road races. No climbs of more than 3min duration, so even at my level it is pretty much full gas and repeatability + aerobic base seems to be my limiter.
Have previously been @ ~4,7w/kg ftp, but this lately have been sitting around 4,0. 81kg.

My current profile limits me in that a 2h race is almost a constant downward slope of FRC. W’ in training software might look like it restores, but my legs sure don’t feel recovered at the end… So more aerobic capacity would be welcome, if I could achieve that on 6h/wk that would be great :sweat_smile:
Also, I don’t train these short efforts specifically.

FRC decreases because you are working above FTP, ergo a higher FTP would mean you would not need to rely on anaerobic/FRC (as heavily)… It sounds like you’ve probably got a fair bit of training behind you, and if <2min efforts are your strength, I would focus on training your ‘weakness’. Aerobic work is a challenge on 6 hours. I can get in 6-10hrs/wk and I had good results with progressive sweetspot (just set @ 90% FTP with progressive TiZ), the benefit is you can do 3/wk and not cook yourself, but get some good adaptations. By the end of the block I could tap out 80min @ 90% and keep HR ~80% max.

1 Like

Think both FRC and FTP come from the same fitting equation, so maybe if your FT is set too low, your FRC might come put too high?

Think 33kJ would mean a pretty decent sprinter.

My (current) FTP is definitely not set too low :sweat_smile:
I have a decent long sprint, but the legs need to be somewhat fresh
Alltime powercurve for reference

I actually had pretty decent success with SS last season… didn’t seem to blunt the W’ though, but am willing to trade some for a higher FTP! Again, on ~6hrs a week :hugs:

Just beware that different CP models will give different W’ results, with the original 2 parameter model giving the lowest estimate, Morton’s 3 parameter the same but the Peronnet-Thibault model is quite a bit higher (which is the model used in WKO, and is similar to golden cheetah’s extended CP, I’m not sure about Xert). For me there is about a 6 kJ difference.

1 Like

I will preface this with saying I am a relative noob to cycling with only a few years experience. I have wandered a fair ways down the rabbit hole of this data though. I feel comfortable speaking to the math and the software but I am in no way qualified to speak to the training or what you should do with it.

That does seem kinda high, but plausible. I suspect, as others have said, your CP is low. For some folks CP and FTP drift from one another by a non trivial amount where CP is generally higher. My CP is at least 10 watts higher than my ftp on TR. If I set my CP 15 watts to low in intervals it takes my W’ from the 18-20kj I think it really is to 30+. I would download GC and look at your last 30-90 days for their CP Tab, under the trends tab. Click the down arrow on the CP model, go to chart settings, go to ‘model’ and change it from 2 parameter to 3 parameter. That should give you a CP and W’ of some sort. I then feed that one into garmin and intervals.

[ScreenShot of CP settings in golden cheetah]

I find using this method it matches up really well with the best ‘real world’ W’ test I have found, which is zwift wtrl racing on sprint days. I race with my IRL club so there is the mental part to push really hard. There is lots of riding just under and at threshold between efforts so if CP is off the curve wont recover correctly. There are 4-6 all out efforts of various lengths between 10 and 90 seconds. When I look at the W’ garmin plugin or intervals.icu after the race using data fed from Golden cheetah to set my CP and W’ I find I am within 1 or 2 KJ of 0 nearly every sprint i go to 100% on.

I would also just look at your last 30/90/120 days. W’ trains up and down on the quicker end of the spectrum. Looking at lifetime PRs relative to each other then the 1m and 15m may be from different years will skew data for this metric.


W’ is the area under the curve(work) when going over FTP/CP as you draw down max power, so as others have mentioned, increasing FTP will mean that your W’ will decrease because the Kj available get is less now. One way to increase W’ is to increase work at max sprint power

Intervals lists my W’ as 9000, I think that was pulled out of Golden Cheetah about a year or more ago which probably says how often I trac that metric.

According to wko5 mine moves between 10-13kj.
U23 females can outsprint me so might be accurate.

Different tools use different models. And they need to be fed some maximal efforts, and as Tom Bell explained in the BikeRadar article one advantage of CP is you only need a couple of (shorter) max efforts.

My primary analysis tool is WKO, but I do some comparison against Xert and Intervals. In practice that means I try and do short/medium/long max efforts as defined by WKO. And when WKO has enough max efforts, the models are fairly accurate.

A few weeks ago I knew the WKO model didn’t have good short power data, it listed a 27-sec or 32-sec duration (I forgot which) as the short target needing a strong effort. WIth that info I sessioned a ~30-second climb and did 4 very strong efforts (max or close to max), separated by about 20 minutes required to loop around and back to the climb.

After those 4 strong efforts, the only portion of my power curve that changed was from 7 to 31 seconds. Before and after shown here:

The dotted line between 200 and 400W is my FTP (270W), so if you calculate the area between the FTP line and power curve (or modeled curve), you have a pretty good estimate for W’ (assuming you have enough (near) max efforts).

With that one change to my 90-day power curve, this is how modeled anaerobic capacity changed in 3 different models:

Metric Before After
WKO FRC 13.7kJ 16.0kJ
Intervals.icu W’ 13.5kJ 13.5kJ
Xert HIE 28.9kJ 29.3kJ

WKO usually has me between 16 to 18kJ, at least when it has enough max and near max efforts at short/med/long durations. I think that is closer to reality than the other models. I believe all 3 use a modeled power curve to estimate anaerobic capacity, and that’s one reason they are different (in addition to each having different definitions).

As @redlude97 mentioned, increasing your short power will drive the curve up and increase W’. And the flip side: if your short power stays the same, raising your FTP - that dotted line above - you can see that W’ - the green portion between FTP and your power curve - will decrease. For example in January I did a 32-min field test and (modeled) FTP increased from 250 to 270, while (modeled) anaerobic capacity decreased from 14.4 to 13.7kJ (FRC in WKO).

1 Like

What’s your Xert TP?

numbers from here: WKO5 is here and it looks... different - #537 by WindWarrior and you can see Xert maps me to a modeled curve at higher 1-sec power and lower threshold power. Lopsided in favor of a higher W’

I think you can force TP to be closer to 270 in Xert and then it will recaulate hie. My intervals and Xert w’/hie are pretty similar but I don’t have any experience with WKO.

thanks, and well that was interesting. Edited Xert to 1125W / 16kJ / 270W and it resulted in 230W lower threshold :rofl:

Now go ride at LTP/LT1 for 3 hours and report back :thinking:

1 Like

16kJ with a peak power of 1125W is possible but unlikely. TP of 250W, HIE of 24kJ with an LTP of 190W is likely a better representation, assuming 1125W is real. One check you can do is to make this calculation:

(HIE*1000)/(PP - TP). A value of about 30 is most common. There was a post on the FB Xert Users group on this and most althletes fell very close to this number.

For those looking to geek out a bit more on this stuff. CP>TP>FTP. HIE*1000>W’ (Don’t know how FRC is calculated). Biggest difference between Xert’s HIE and W’ is that W’ can be fully depleted whereas HIE can’t. There is always a residual capacity of HIE left at failure in Xert’s model. CP/W’ leaves no capacity left at failure (which is why W’bal will always drop below 0 after most all out efforts to failure. You can often go above CP when this happens, even for as long as 10-20s). CP is overestimated as a marker of a point of equilibrium by as much as 10-15W as a result, W’ is underestimated and W’bal calculations won’t ever work reliably. :disappointed:


Thanks. FWIW I seem to have high PP and ok long TP, while relatively weak from 10-sec to 5-min. Lately I’ve been doing a lot of 2+ hour endurance rides at 190 so that LTP seems reasonable. In particular 30-60 second max/hard efforts I find quite challenging, and if its more mental than physical I can see an argument in support for 20-24kJ. Or maybe that “strong PP / weak anaerobic / ok diesel” simply makes me an outlier and difficult to map onto a modeled curve.

In response to your check here a few datapoints, and some calcs:


  • I’m not sending a lot of TP signals to any model
  • did a 276W / 32-min effort recently, it wasn’t a max effort, and unfortunately ended early so not the best signal. It was paced based on a 291W / 11-min “pretend you are doing a 20-min test” effort from a couple weeks prior


  • past 90 days the top 5 are 1100 to 1127W and I haven’t been working on this, those were ad-hoc ‘work on coordination’ efforts at the end of a multi-hour endurance ride
  • top 15 over past year are 1100W to 1236W (from a 20-sec sprinterval workout last summer)
  • takes some focus and good cadence/incline-or-wind to hit 1200+ peak but it has been repeatable and data looks clean coming from my current Quarq and previous Stages gen3 (I have removed spikes from what appears to be 4 outliers on Stages at 1400-1600W)

Using the 276W and a week before 1078 PP on a ride ending hard push… and some random sensitivity analysis (not knowing the distribution of 30):

HIE = 34 * (1078-276) / 1000 = 27kJ
HIE = 30 * (1078-276) / 1000 = 24kJ
HIE = 26 * (1078-276) / 1000 = 21kJ

So lets use 26 and plug in some slightly higher values based on last 90 days:
HIE = 26 * (1127-280) / 1000 = 22kJ

Looking at “HIE = 30 (a variable) x (PP-TP)/1000” formula it reminded me its possible to use simple middle school geometry to cross-check, and as a ball-park first order approximation:

W’ is more than 20kJ, and again its a simple rough ball park number. I picked 8-min because a) the contributions above and below the red diagonal are roughly equal so I can use half the area of the square, and b) the small contribution of area below/beyond 8-min so this rough W’ calc is a pretty decent lower bound for a reality check on other anaerobic capacity numbers (where each has different assumptions).

@baronbiosys your thoughts?