Another Piece in the Training Intensity Distribution Puzzle (Or, Maybe More Reason to Back Off the Tempo)

After reading this article

I hunted down the sources.

Paper from earlier in the year looking at Complex I proteins and endurance metabolism. The crossing point where Complex I gets bypassed is, apparently, lower than the average endurance athlete and coach might think – perhaps even lower than the aerobic threshold.

Certainly seems to back up the polarized model, and support some of the points that @brendanhousler raised in his “Tempo, the Silent Killer” blog post earlier this year.

For road racing and XC Mtb, you need some tempo training because chunks of races end up in that zone. Tempo can yield strong aerobic adaptations – if you don’t do too much and build up a weighty fatigue debt. But too much looks to compromise both Max aerobic power and lower-intensity aerobic efficiency.

As modelling gets better, I would hope that TrainerRoad, WKO, Xert, etc., would be able to develop predictors for training distribution based on the watt and HR data (the two big points they have to work from). Short of having a metabolic cart hooked up to the athlete regularly, it would be pretty hard to determine, but being able to more precisely individualize intensity distribution and dose seems like the next big step with how to use the data the gadgets give us.


High tempo / low sweet spot (In particular at low cadence) is recommended as a way to reduce VLaMax. This is likely needed more in athletes who skew towards fast twitch muscle fibers. So as usual, there’s not a one size fits all approach that works best.

I think there’s still a big missing piece on prescribing training which is understanding an athletes physiology - and response to training - and using that to tailor / adjust training over time.

Apparently TR is working on something like this. I’ll be very interested to see what they land on.


Larsen and his colleagues call this threshold “complex I max,” or CImax for short, and it marks the point where you body first starts to abandon its most efficient metabolic mode. In the trained cyclists in Flockhart and Larsen’s study, this occurred around 55 to 65 percent of maximum heart rate.

Funny how i look back at my easy workouts; lot of them have avg. heart rates of 130-137 (my max is 200). But also that CImax is for trained cyclist. So… 1) what are trained cyclists 2) how can i use this information (CImax) 3) how do you find it or is it same for everyone, because can it go lower than that??? Haha.
If CImax occurs at 55-65% of max HR, is this the best place to train for it? Or is aerobic threshold still a good limit?

I remember Stephen Seiler saying that polarized model is good even though you don’t have time for big volume and maybe this could be the reason if tempo(?) is not good for aerobic efficiency.

isn’t this to goal of INSCYD model? This is from The Science of Winning and doses are for swimming:

Some adaptations are individualized, and the INSCYD model supports a relatively straightforward field test (no lactate testing) to determine which way the adaptation is playing out for a specific athlete. The only issue is that the INSCYD software model is proprietary and not widely accessible. I’ve only looked at one of the nearby coaching options that use INSCYD, and they sell a 4 pack of assessments for $740 (or $185 per assessment). I don’t consider $740 for 4 months of individualized training to be cost effective however I’m simply trying to enjoy fast group rides.

Thanks for the article links.

I just read the article. One of the takeaways is that muscle glycogen is the preferred fuel for maximum ATP production, and is more efficient than glucose. The other is that this seems to re-enforce what Stephen Seiler proposes as an optimum training method. I also was unaware that lactate is produced by Type II muscle cells and metabolized by Type 1 muscle cells.


I don’t think this article is remotely saying what most people in this thread apparently believe it does.

1 Like

The Svendsen story is interesting. It sounds like his heart was in it. With his VO2max he probably needed to do those back to back to back long, slow 100 mile days.

The article was describing the output from a mathematical model that they have built to model metabolism.

As such, the output is only as good as the model, which the lay reader has really no way to judge.

The type of research that yields the most compelling insights on exercise are real world tests on athletes.

As such, when I read the article, I didn’t really take away too many Ah-ha conclusions.

It’s been known for a long time that VO2max and efficiency are inversely correlated. This article hypothesizes a reason for that. But I don’t see any new insights that might inform training.


:+1: Yes, a mathematical model to describe the why. Reading the Silicon Valley Expert Analytics’ blog post about it, the highlighted sentence in last paragraph could read like INSCYD if you replace “efficiency mode” with FTP, and “power mode” with VLamax. Interesting blog post here:

1 Like

Not surprisingly - given my stance on polarized training - I actually find this more to support touching all intensity zones than avoiding tempo. And this is also related to what Weber&Co are saying.

As Weber points out: “Polarised training seems to favour development of your VO2max in a simplified manner.”

With avoiding tempo one misses to make those fast twitch fibres more aerobic. In the polarized model these get only recruited with vo2max work (or when you ride long, really really long). Without this work and the focus on vo2max work vlamax rises/stays high. You’re more inefficient.

This is why cycling pros do so much tempo work. To lower vlamax. Grand Tours, epic one-day races are all about being efficient. The one who has left most in the tank in the end wins. Hence, they train all: base endurance, tempo/SST and high end for those critical moves in a long race. This makes coaching far more difficult (who knows, perhaps this is one reason why the pol model is so appealing to so many: easy to grasp and the promise of having found the holy grail).


Whenever I start thinking too hard about all this stuff I generally end up concluding that I should stop worrying about it and just ride my bike more!


Would be interesting to understand how (or if) Svendsen’s test results and training approach compared to other young pros. A lot of the basis for the article is what seems like an N=1 anecdote. I’m sure there’s a lot more science and data behind it, and Hutchinson is just using Svendsen’s story as a way of making it more relatable and digestible for the average reader who doesn’t have a degree in sports science. But without comparing him to other riders it’s not at all obvious to me that inefficiency was the main reason his pro career didn’t work out.

Was he doing significantly different training to other pros? Or was he doing the same training but his physiology meant he responded differently? Or were his training and response both pretty standard and his failure to make it as a pro was totally unrelated? He’s hardly the first promising young rider to fail to live up to expectation, if anything he’s in the majority. And he was on a team (Joker) who have produced riders like Kristoff and Boasson Hagen, so it seems unlikely they would have him on a plan thar differed much from what has worked with other riders.

1 Like

55-65% of max heart rate is even lower than MAF heart rate. If I kept my heart rate to those levels I’d barely be moving. I’m not sure I’d get fitter doing track stands!


This :point_up_2: :point_up_2: :point_up_2:

The INSCYD testing looks very interesting - and after hearing Sebastian Weber on Mikael Eriksson’s podcast describe the theory behind it, I think this kind of testing is practically the best way forward I.e. testing that can be done with a power meter instead of in a lab. Too bad it’s pretty expensive! maybe TR will come up with something similar??


I think LT1 is a better threshold to designate “lower effort” pacing. This is a well known approach and is applied today in training - at least by pro riders.

Remains to be seen of introducing “complex I” based training will catch on. I personally don’t think so, as I think an approach like INSCYD - a power based approach that gives a lot of very useful information - is the way forward.