The forum wisdom is old and confident: blow yourself up before you taper, because that's how you supercompensate into the best form of your life. So he does — three brutal weeks, sessions he has no business finishing, sleep he isn't getting back — and shows up to the taper certain he's bought himself a peak. Two weeks later he's not fast. He's flat. Aubry and colleagues ran the actual experiment behind that forum wisdom, and the result contradicts the story riders tell themselves about it: the athletes who pushed hardest into genuine overreaching didn't taper into their best form. They tapered into a smaller rebound than teammates who trained just as hard and stopped short of it (Aubry 2014).

Section01 · What the data actually found

Thirty-three trained male triathletes went through the same protocol: one week of moderate training, then either three weeks of deliberately increased load (23 of them) or continued normal training as a control (10 of them), then a four-week taper — with cycling performance and VO2max measured at the end of the moderate week, at the end of the overload block, and then every week through the taper (Aubry 2014). Here's the part that matters: the researchers didn't decide in advance who "overreached." They classified it after the fact, based on what each athlete's body actually did during the block. Eleven showed a real performance drop during the overload weeks, combined with high perceived fatigue — that's the clinical definition of functional overreaching, F-OR. The other twelve got just as fatigued but never lost performance during the block itself — acutely fatigued, AF, not F-OR.

Same training block. Same three weeks. Two different physiological outcomes, and the taper treated them completely differently. The AF group's peak performance supercompensation after the taper was 2.6% ± 1.1% greater than the F-OR group's, and 2.6% ± 1.6% greater than the normal-training control group's (Aubry 2014). VO2max rose from baseline at peak performance in the control and AF groups. It did not rise in the F-OR group (Aubry 2014).

Read that last line again. The athletes who pushed hardest into genuine overreaching finished the taper with a smaller aerobic gain than the ones who just kept training normally the whole time. Deliberately overreaching didn't buy them anything. It cost them the gain the other two groups got for free.

Worth being precise about what the study can and can't tell you. Twenty-three subjects split after the fact into groups of eleven and twelve is a small comparison, and the classification depended on how each individual's body actually responded to an identical block — not a coin flip assigned in advance (Aubry 2014). That's a real limitation on how confidently you can predict, in advance, which specific week or which specific session tips a given rider over the line. It's not a limitation on the headline finding itself: within this cohort, more training stress stopped paying off exactly at the point where performance itself started dropping, and pushing past that point cost the taper's return rather than adding to it.

Section02 · Why crossing the line costs you, not rewards you

The taper's whole job is to let accumulated fatigue dissipate on top of a training stimulus you've already banked — cut the volume, hold the intensity, and the fitness you built shows up once the fatigue clears. That mechanism works when what's underneath the fatigue is a normal training stress. Functional overreaching isn't that. It's a distinct physiological state — the performance drop during the block, the immune-system cost that showed up as a 70% infection rate against 10% in the control group, the VO2max that wouldn't rise even after a full month of reduced load (Aubry 2014). A taper is built to dissipate fatigue, not to reverse a stress state your body has already tipped into. Cutting volume for four weeks gave the F-OR group's fatigue somewhere to go. It didn't give them back the adaptation the AF group got by not crossing the line in the first place.

This is also the part of the finding that quietly demolishes the "no pain, no gain" version of block-building. More load did help — the control group, who just kept training normally with no overload block at all, still came out behind the group that pushed hard and stayed short of F-OR. So the answer isn't to go easier across the board. It's that the relationship between training stress and taper payoff isn't a straight line you can extend indefinitely by suffering more. It has a peak, and past that peak, more suffering buys you less rebound, not more.

Section03 · The edge you can't feel from the inside

The uncomfortable part of this finding is that F-OR and AF felt the same from the inside — both groups reported high fatigue. The only thing that separated them was an objective performance test during the overload block itself. That's the actual signal, and it's not the one most riders are checking. "I feel wrecked" describes both groups in this study. "My power is already down, mid-block, compared to where it was a week ago" describes only the group whose taper failed to deliver.

That reframes the entire question a serious cyclist should be asking before a big block. It isn't "how hard can I make myself feel." Perceived fatigue was high in both the group whose taper worked and the group whose taper didn't. It's "has my output during the block itself started dropping" — because that's the only variable in Aubry's data that actually separated a taper that pays off from one that doesn't.

This is also why forum wisdom gets this backwards so reliably. "I'm completely wrecked" is a feeling you can manufacture with enough volume and not enough sleep, and it's genuinely hard to distinguish from the inside from the feeling that preceded a good taper in this study. The only thing that told the two states apart from the outside was a number on a power meter, not a subjective fatigue score. That's an uncomfortable amount of humility to ask of your own judgment mid-block, and it's exactly the judgment the forum-wisdom version of this advice assumes you don't need.

Section04 · Running the block without crossing the line

1. Push a real overload, don't half-measure it

   Higher training load did produce a bigger rebound than normal training — as long as it stayed short of F-OR. The goal isn't to avoid a hard block. It's to avoid the specific failure mode inside a hard block.
2. Test performance mid-block, don't just track how you feel

   The only thing that separated the group whose taper worked from the group whose taper didn't was an objective output drop during the overload weeks. Build in a short, standardised effort partway through your hardest block and compare it to where you started.
3. If output is already dropping, that's the stop signal

   Rising fatigue with maintained output is the group that tapered into its best form. Rising fatigue with dropping output, mid-block, is the group whose four-week taper still couldn't fix it.
4. Taper by cutting volume, not intensity or frequency

   The optimal taper strategy holds training intensity and frequency constant while cutting volume by 41–60% over roughly two weeks (Bosquet 2007). Slashing intensity instead detrains the exact quality you built the block to produce.
5. Recover like the infection number is real, because it is

   A 70% infection rate in the F-OR group against 10% in the control group (Aubry 2014) is a training-maladaptation signal, not a coincidence. If you've already crossed into F-OR, the priority is getting healthy, not shortening the taper further.

Section05 · Where riders get this wrong

Section06 · Applying it with THRIVE

Section 07Bottom line

The advice to deliberately overreach before a taper gets the mechanism backwards. Aubry's (2014) data says the taper rewards a hard block that stays short of true functional overreaching — not one that pushes past it. Cross that line and you get a smaller rebound, a VO2max that doesn't move, and seven times the infection rate of an athlete who never overloaded at all. Push the block. Just check your output mid-block instead of trusting how wrecked you feel — because that's the only signal in this data that actually told the difference between a taper that pays off and one that can't.