Cyclic Failure: How Crashes and Melt-Ups Actually Form

A cyclic failure is a cycle low that breaks shortly after it forms — price drops back down through a trough that was supposed to hold, and the next larger cycle takes over to the downside. Crashes are what that looks like at scale. A melt-up is the same phenomenon in reverse: a peak that was supposed to cap price gets broken back to the upside, and buyers carry the market higher instead.

Most declines respect the cycle map. Price bottoms in its expected windows, peaks in its expected zones, and the analysis stays boring. Failures and melt-ups are the exceptions — the moments when a cycle does not do what the model drew for it — and they are worth studying precisely because they produce the biggest moves on the chart. This article defines both, then walks through four dated historical examples on the S&P 500 so you can see how each one actually developed.

What exactly is a cyclic failure?

In cycle analysis, every trough has a job: it is supposed to hold. When the 80-day cycle makes its low, that low should be the floor for the next 80-day span. A cyclic failure is what you call it when that floor gives way soon after it forms. Price rolls back down, takes out the fresh trough, and the tidy picture the nominal model drew is gone — the next cycle up the ladder is now running the show, with its larger downside projection.

The same logic runs in reverse at a top. Peaks have a job too: to cap price until the cycle rolls over. When price drops off a probable peak, then breaks back up through it and keeps going, the peak has been melted up. Massive buying creates it. Both events distort the standard cycle picture, and both are readable once you know what you are looking at.

Here is the anatomy of a failure, step by step, drawn from the examples below:

1. A cycle peak forms, and price begins easing off it — normal so far.
2. An FLD interaction that should hold, fails. In the historical cases below it was a B or C interaction: price gets above the 20-day FLD and cannot stay there. (The full interaction sequence is covered in the eight FLD interactions.)
3. Price crosses back down through the FLD.
4. The recent cycle trough breaks — the low that was supposed to hold, doesn't. This is the failure itself, and it is hard to recognize in the moment.
5. A valid trend line break confirms that the next larger cycle peaked in the past. (How that confirmation works: what is a valid trend line.)
6. Price runs until the larger cycle's projection is met — often not stopping until it reaches support at a much larger FLD.

What did the August 2022 failure look like?

In August of 2022 the 80-day cycle on the S&P 500 was broken after a peak formed, pushing price down into the 20-week cycle low. Two distinct events marked the way down.

The first was the cyclic failure itself — although, and I want to say this plainly, it is hard to see that at the time. Price had crossed down through the 20-day FLD on 22 August 2022 and kept going. The C interaction then told the story: price got above the 20-day FLD and failed, setting up a cross back below it the next week, on 12 September. From there price did not look back until it met the 80-day cycle projection.

The second event was the break of the probable 80-day valid trend line. The VTL rule is specific: when the 80-day VTL is crossed to the downside, it means the next higher cycle — the 20-week — made its peak in the past, roughly half the VTL's span back, about 40 days. That break turned a suspicion into a confirmation.

The whole process was quick. The 20-week cycle ended at 17 weeks — early, but an average of 17 weeks is fine; cycles drift around their nominal lengths. And the resolution carried its own information: the ending point of that 20-week trough came in almost equal to the trough from 17 weeks before. Sellers had pushed hard and gained little ground, which meant buyers were probably still willing. When price crossed up through the 20-day FLD on 18 October 2022, the cycle logic said that was likely an A interaction — a fresh cycle starting — because the 80-day projection had been met, the 17-week count fit, and the 20-week FLD was providing support underneath. Even a trader confused by the failure had a coherent map again within weeks.

How did a cyclic failure start the 2022 bear market?

The January 2022 top is the bigger example — a failure at the 80-week degree, which is what turns a correction into a bear market.

Into the end of 2021 everything was going well, though the phasing had gotten sloppy — early sellers in December made the interaction count hard to read. The tell came when the B interaction failed and price went down through the 20-day FLD on 5 January 2022. Then the defining event: price broke below the trough of the 80-day cycle. Getting below that trough is a failure — especially right after a major peak. The 80-day VTL break then confirmed the peak almost eerily, right on schedule.

What followed shows why failures matter. The market did not stop falling until the 80-week FLD was touched from above on 23 February 2022 — and that touch produced a strong bull candle, which also happened to mark the 20-week cycle trough, early again at 17 weeks. That is the power of FLD support: even in a failing market, the larger line caught the decline where the model said it could.

One more thing worth knowing: this same failure pattern appeared just before the pandemic volatility of February 2020, in almost identical form. Price broke several troughs in succession and did not stop until the 54-month and 80-week cycle projections were met, far below the February highs. Different cause, same structure. The news changes; the mechanics of a failure do not.

What is a melt-up?

A melt-up is the cyclic failure of a peak. July 2020 is the clean example: the bull trend off the pandemic low simply melted up the peak of the 20-week cycle and kept going.

The sequence looks like this. You monitor the average duration of peaks, and when price drops below the 20-day FLD in coordination with a probable peak, you treat the top as likely in. Then price does something the map did not draw: it breaks back up through that peak and holds. That break is the melt-up. From that point the message of the chart flipped — the structure said to stay dedicated to the long side and to keep working with price whenever it came back toward, or found support at, the 20-day FLD, which is exactly what it did through August 2020.

One honest operational note from that period: cycle software will keep trying to find the peak. It is built to. If price keeps running for days or weeks past the probable top, the peak is likely melted up, and the analyst has to overrule the peak-hunt and read the failure for what it is. Peaks in these regimes still follow the nominal model, just with somewhat longer averages than usual.

How did melt-ups flag the 2023 bull market early?

The late-2022 chart shows a variant worth studying, because it demonstrates melt-ups as an early-warning system rather than a single event.

Through the protracted 2022 selloff, the troughs of the large cycles started moving down less — the October 2022 low sat only modestly below the June low. Then, in December 2022, something genuinely distorted happened: traders pushed the 80-day cycle trough up above the previous 20-week trough from October. A shorter cycle bottoming above the prior larger-cycle low is a distortion in the cycles. Buyers were there, ready.

Then the peaks started melting up. A 40-week peak was taken out, and bulls pushed price above each successive prior peak into early 2023. By the time the 80-week cycle trough formed in March 2023, the stage was set, and the bull market of 2023 ran from there. A trader watching those peaks fail upward could have shifted bias early in the year — not from a headline, but from the chart. When there are fundamental reasons for a trend to reverse, the melt-ups are how the cycle map registers it.

How do failures and melt-ups compare?

Both resolve at levels the FLD framework projects in advance — the projection mechanics are covered in FLD crossings.

Can you see a failure coming in real time?

Mostly, no — and I would rather say that out loud than pretend otherwise. The August 2022 failure was hard to see as a failure while it was happening. The January 2022 phasing was sloppy for a month before the break. Failures and melt-ups are, to a real degree, known after the fact; what you can do in real time is define what would prove your read wrong, and respect it when it triggers. Here is what that looks like from my own ledger — a miss, on the record:

Logged 2026-07-09, scored 2026-07-18. On the 4PM show the desk's read was the S&P rising into a 40-day peak, with the invalidation stated on air: a close below that day's session low would mean the phasing was wrong. Within six sessions the index closed below that level. Invalidation triggered — scored a miss in the calls ledger. That is what a failing read looks like live: the map said one thing, price broke the level that was supposed to hold, and the invalidation did its one job. The grade stays in the record next to the hits.

That is the honest posture toward failures: you will not catch them at the moment they begin, but a stated invalidation converts "the cycle failed" from a disaster into information. The larger structure — which cycle now controls, where its projection sits — then gives you the map for what follows, the way the October 2022 A interaction did after the August failure.

Why bottoms and peaks behave so differently in the first place — why the lows stack and the tops smear — is its own subject, covered in why bottoms form faster than tops. And the full rules for how I phase charts, log calls before outcomes, and score them afterward live in the methodology.

Questions traders ask

Is every market crash a cyclic failure?

The major declines examined here — January 2022, August 2022, February 2020 — each began as one: a trough broke shortly after forming and a larger cycle took over. Most ordinary declines are not failures at all; they are cycles bottoming on schedule. What makes a crash different in this framework is precisely that a low which should have held, didn't, which unlocks the next larger cycle's downside projection. That said, this is a lens on how crashes develop, not a claim to explain every decline in history.

How do you tell a normal decline from a cyclic failure?

A normal decline holds its recent cycle trough; a failure breaks it. The supporting evidence arrives in a sequence: an FLD interaction that should hold fails, price crosses back below the FLD, the fresh trough gives way, and a valid trend line break confirms the next larger cycle peaked in the past. Any one of those alone is a caution. The stack of them is the failure signature.

Can a melt-up be predicted in advance?

No. A melt-up is defined by price breaking back up through a peak that had already probably formed — you only know it happened once the break occurs and holds. What you can do is recognize it quickly instead of fighting it: if price keeps running for days or weeks past a probable peak while your software hunts for a top that never confirms, the peak is likely melted up, and the late-2022 example shows how a series of them can flag a regime change early.

Where does a failure usually stop?

At the projection of the cycle that took over. In August 2022 price ran until the 80-day cycle projection was met. In early 2022 the decline did not stop until price touched the 80-week FLD from above — which also marked a 20-week trough. In February 2020 price broke several troughs and stopped in the area of the 54-month and 80-week projections. The larger cycle's FLD framework supplies the destination; the failure supplies the trip.

Keep reading