Does this prove over/unders don't work any better than riding at tempo?

This is not an earth shattering concept. One that is followed my many different coaching trees. You’re basically trolling and clearly don’t care to provide anything productive.

I’m out.

1 Like

Z2 does a lot more than that, in fact it’s optimal for certain adaptions.

I want to emphasize the issue of the underpowered-ness of these studies. I think everyone knows that this means (informally) the study is less likely to find statistically significant differences between groups. I think it is less well understood that underpowered studies can inflate effect sizes, so when statistically significant differences are found, they may not reflect reality.

In other words, we should be very skeptical of the results of studies with small sample sizes. This is a fundamental and inescapable issue.
When the sample size if small, if a study fails to find a significant effect, that’s (possibly) because there is too much noise in the data. If a study does find a significant effect, that’s also (possibly) because there is too much noise in the data.

How large should the studies be? It depends on the study design (e.g. cross-over or parallel, which has its own implication) and also how big of an effect we expect.
How large should effects be? This is really tough! The effect of increasing training or increasing intensity on their own should be large (depending on how long the intervention lasts), but I think we’re mostly interested in issues like “Hard Start vs Steady State” or “POL vs PYR”, which are not likely to have large effect sizes. Table 3 in the study under discussion lists the effects they analyze as mostly small or moderate.

In general, for any question where the answer is not already obvious (a moderate or smaller effect size), general calculations suggest a sample size over ~130 comparing two groups. A cross-over study can get away with fewer, maybe ~35, but they have other issues. And that’s for a moderate effect size. Small effects need around 3.5 times the number.

A slightly related gripe:
The analysis in the Setpto-Hawley paper is crazy. At least they didn’t try to fit a quartic polynomial. I would be skeptical of the analytic abilities of anyone who cites that study as their primary evidence of the efficacy of those intervals.

Finally, remember that different adaptations take different amounts of time and the fastest adaptations may plateau the quickest. A study may find different results at 3 weeks vs at 3 months (and most studies are closer to the former).


I had a check and found some studies looking at the relationship. The mean 75% of max HR was 69% of VO2 max. Not hugely out, but yes a little under 70% of VO2 max. Lower for untrained.

1 Like

Some of my earlier explanations were a little off…

But i found this good paper that explains most of the concepts.

I know @empiricalcycling has spoken out against this model, but i don’t see as much disagreement. In general, the model explains what we see with power duration curves. Those with very high peak power (anaerobic) don’t tend to have as well developed lt2… or i should say will be further away from MAP than a gc contender ( within their own power duration curve). Interested to hear his critique. I do know that one propnent of the model does seem to overvalue the ability of the model to predict one’s power profile, but there does seem to be a good correlation.

He said changing fractional utilization takes a long time, and i agree, but i have noticed a rapid reduction in fractional utilization after training for anaerobic capacity. ( supporting the model concepts).

There’s not much disagreement because most sports scientists aren’t as acquainted with the biochemistry. It shows when I talk to people who study actual metabolism, who say “oh that theory is ridiculous, it doesn’t work like that,” and don’t care to publicly make their thoughts known because their entire peer group feels the same, as opposed to the cycling training peer group.

1 Like

We all have friends that fit this description and we can also see the sprinters in the pro peloton that seem like a different species, depending on the stage.

This is a good observation to start with, perhaps @empiricalcycling could take it from here.

Do you have a specific episode you had already recorded that provides a critique of the model?

I wasn’t talking about disagreement on how the model is calculated per se, but there doesn’t seem to be much disagreement with the observations we have with individual power profiles and how they tend to fall in to a pattern/buckets.

Then we have certain training interventions which aim to try and change or amplify those profiles.

In addition to @Bioteknik excellent citation I would add the Coggan classic ‘Determinants of Endurance’ which reaches pretty much the same conclusion but has the excellent advantage of being free to read & consider FOR FREE in its entirety.

1 Like

Interesting that someone with their LT at approach 56% of VO2 max, can last 15 mins, at 32% above that. Do you know which LT that refers to? If it is LT2 seems amazing someone could work that far above it, for so long.

“(LT) was determined…by graphing an individuals’ venous blood latate concentration measured after 10 min of cycling at five intensities ranging between 50 and 90% of VO2max. The bouts were performed during two testing sessions during the week before the performance evaluation. During cycle testing, a Quinton cycle ergometer (model 845) equipped with toe clips was used. This ergometer provides a constant power output independent of pedal frequency. However, the subjects were required to maintain a pedal frequency of 75-85 rpm during all testing. Venous blood samples were obtained from a catheter in an antecubital vein before and immediately after exercise and assayed for lactate concentration.”

“At fatigue, the blood lactate concentration averaged 7.4 +/- 0.7 mM in group H, whereas it was approximately twice that (i.e., 14.7 +/- 1.0mM;…) in group L at fatigue.”

As a special challenge to the forum…we should all attempt to attain a blood lactate of 15 mmol some time this week. Report whatever sensations you experienced here. :smiley:

1 Like

Let me explain figure 1 there in more colloquial language…
Participants’ LT = MLSS, and the range was 59-86% relative to VO2max. What they did was have people ride at 88% of VO2max (over threshold for everyone) and looked at how long they could hold it. Go figure participant 14 was the 59% guy, and 1, 3, and 4 were >85%. For control, everyone had almost exactly the same weight and vo2max. So all we see here is the concepts behind critical power and W’ at work.

The study is a good one because it was fairly thorough and has a pretty open data set too that many have used. One of the things they looked at is fiber type, which also correlated to a high MLSS/VO2max ratio, what we’ve been (technically incorrectly) calling fractional utilization.

I think what’s missing here is received wisdom from anecdata vs looking at actual physiologic limiters. The relationship between sprint power and aerobic power only really has crossover with fiber type, which will partly change efficiency and fractional utilization here (look at the same study, table 5) even though it only partly explains that relationship. Turns out type II fibers can be highly aerobic and fat burning with high capillary density too, and we’d probably need to look at more architectural differences between fiber types here since the histochemical staining doesn’t differentiate hybrid fiber types of IIa and IIx MHC types.

The thing about big sprint power is that the muscles create the demand for ATP first based on how hard and fast/frequently they contract, and that creates the demand. This is the “pull” that gets other aspects of metabolism running. Me with a 2000w sprint can have every aspect of muscle identical to someone with 1500w and I’ll make more lactate simply because I can create more demand. This doesn’t tell us anything about the factors that determine vo2max (mostly stroke volume) and threshold (seems to be mostly vo2max and capillary density, some fiber type). So while it seems having a big sprint may reduce fractional utilization, I estimate the delta is only about 5%. And none of this determines actual endurance performance. One of my pro sprinters (1900w max, 410 ftp) was being used by his team for breakaway control duties his endurance ability is so good compared to the rest of his team, which by colloquial logic should make them better at such things, since their sprints were not nearly as good.


Thanks for the 10 minute tips #27, listened at work today. Did I hear correctly that you put 30/30s in the ‘extremely over / unders’ programming bucket? :joy: Last week I did some 30:180s and you and Kyle are right - more power on the repeats just makes it hurt more! Next time I’ll be wanting to kill myself quicker on the 5 minutes 30/30 jobbers you guys talked about.

Then, what explains that Ewan, Cav, Groenewegen, Jakobsen…consistently struggle against the time limit to finish a a hard climbing stage?

For me, 14-16 mmol/l gives me the shakes. I did a series of sprint workouts last summer to see how high I could get my lactate. I could have done one more rep in each, but I was at a point where I was wondering whether another rep was worth it or not.

When i did a sprint interval study, peak lactate for me was around 17-18 mmol after 6x30s maximal efforts with 5 min rests. This had me dry heaving in the trashcan.

It was three or 4 sessions of this as we were doing a blinded study to see if a beta alanine product had any effect on sprint power. The intervention sessions didn’t differ for me… as the product was blinded so i don’t know what one was the one in question vs placebo, but one definitely had lower rpe. But if W’ is unchanged, beta alanine won’t be increasing avg power over 6x30s efforts. (My take on the results)

That’s deliberate. Bike racing at that level is as much about energy management as anything else, and with sprinting it matters a lot. If a sprinter isn’t coming in first, they want to be last because otherwise you’ve spent energy you don’t need to spend. Why be in the third group on the road when you can be in the fifth and still make it to Paris? If you’re 100th instead of 180th, you still finish the race.

1 Like

Just finished both of the @empiricalcycling podcasts and enjoyed them both. My takeaways were: Focusing on some super specific physiological process is probably almost always a mistake. Over/unders properly done can be beneficial to riders (like me) for whom punch and acceleration are a big limiter. Naturally punchy riders can probably do well with steady state intervals.

1 Like

Very commendable! You know you are in a rare cohort when ‘getting my blood lactate as high as I can’ sounds like a good idea. It’s very hard to take your own blood lactate when it’s in the upper teens! :rofl:

Very well done. That sounds horrible. :grin: