FAILED

0.618 retracement (13% swing filter)

Price does not move in a straight line: after an advance it gives back a predictable fraction of the way before continuing. That fraction is not just any fraction — it is the ratio found in the nautilus shell, the pyramid of Giza and the arrangement of leaves — and buying exactly at it is buying at the point where the move starts again.

Measured in crypto — Binance spot · 540 pairs, delisted ones included · 0.2% per round trip

Net per trade

+0.31%

after fees

The fee is charged on both legs: every trade pays to open and pays to close.

Died at

benchmark

loses to buy and hold (same horizon)

Worst drawdown

−87%

1,056 days underwater

How far the account fell below its own best previous moment.

Timeframe

1 day

The time grid this was measured on. The same technique on a coarser or finer grid is a different measurement, and can earn a different verdict.

invariance
costs
placebo
benchmark
out of sample
multiple testing

Equity curve

Risking 1.0% per trade, marked to market every day — not only when a trade closes. The line at 1.0 is the capital it started with.

At its worst the account was worth 87% less than its own best previous moment, and it spent 1,056 days below that peak.

The technique against chance

Same number of trades, same holding time, same assets — only the dates were drawn at random.

The average trade and the typical one

The mean sits above the median: the lottery signature — the typical trade loses and a rare handful pays for everything.

What was measured

We coded the rule exactly as it is described and let it trade on its own from 2017-08-17 to 2026-07-28. It found trades in 529 coins, 32,225 in total. Each one enters and exits at the price that existed on that day — the program never sees what comes next, which is the most common mistake people make testing strategies at home. Wherever the description was ambiguous we took the reading least favourable to the technique; whatever was left open is stated in the hypothesis, at the foot of this page.

How many of those trades actually count

Only 29 of the 32,225. Cryptocurrencies rise and fall almost in unison, so the same rule fires on dozens of coins on the same day — and that is one bet repeated, not dozens of different bets. Counting them all separately is the trick that makes a bad test look impressive.

What it paid, before any deductions

+0.51% per trade. This is the number the technique claims, and the only one on this page that has not yet been through a single control.

Why it did not pass — it died here · against buying and holding

We compare it against simply buying and holding for the same stretch of time, with no rule at all. Sitting still paid more. The technique takes work, demands attention, and delivers less than doing nothing.

What would have happened to the money

Risking 1.0% of capital per trade — the most widely taught rule — across the whole period: The capital would have ended at 1.14× what it started with. But that number is not trustworthy: draw the order in which the signals arrive and it ranges from 0.88× to 1.36×. That is the draw talking, not the technique. At its worst the account was worth 87% less than its own best previous moment, and it spent 1,056 days below that peak. In none of the 12 paths tested did it fall to less than half of what it started with.

The account would not fit every signal

48% of the trades had to be turned down: by the time they appeared, the money was already tied up in other positions. That matters because the average return per trade the technique claims is computed over trades that nobody could have taken all of.

What this result does NOT say

For an edge to be assertable here it would have to reach 4.54% per trade — that is the size that survives this archive's multiple-testing correction, and it rises as the archive grows. So FAILED means “we found nothing above that size”, and never “it cannot possibly work”. The instrument is far more sensitive than that: on synthetic data, with a clean effect, it separates from 0.69% upwards. The distance between the two numbers is the price of a real market and the price of publishing many claims. The difference matters, and it is the rule of this house: the card confronts the claim, never the person who made it.

What this card does not measure

Every verdict holds for the conditions it was measured under. These are this card's — and outside them the result does not apply.

One market, one universe: Measured on 529 spot cryptocurrency pairs, delisted ones included. It says nothing about futures, equities or indices, nor about how the same technique behaves in another market.
One window of time, not every window: The measured period runs from 2017-08-17 to 2026-07-28. A market moves through regimes, and a technique can work in one and fail in another — the card measures the regimes that fit inside this window, not the ones still to come.
One cost structure: The cost charged is 0.10% per leg, in and out. Anyone paying more than that gets a worse result, and anyone paying less gets a better one — the verdict holds for this fee.
One exit rule: The trade was closed by: the technique itself (held until the opposite signal). The same entry measured with a different exit is a different strategy, and can earn a different verdict — it happens in this archive.
The number of trades is not the sample size: There are 32,225 trades, but only 29 independent market episodes: a single move fires the technique across dozens of assets at once, and counting those as separate observations inflates any result. It is the smaller number that governs the arithmetic. With 29 episodes, what the data supports is a range from -1.44% to +2.47% per trade — the published average is the centre of it, not the exact measurement.
Daily bars: Measured at the daily close. Nothing here measures what happens inside the day, and an intraday technique is not auditable with this data.

Numbers and reproducibility

The six controls

control	status	t	threshold	episodes
invariance	passed	—	—	—
costs	inconclusive	0.33	2.05	29
placebo	inconclusive	-0.02	2.05	29
benchmark	failed	—	—	—
out of sample	passed	—	—	—
multiple testing	failed	—	—	—

invariance32225 signals across 750934 barsconcentration: +34.35% of the profit sits in the top 5% of trades — with the dates shuffled, +55.26% (fails above +50.00%)
costsgross +0.513% · cost 0.200% · net +0.313% (t=0.33) · the range runs from -1.645% to +2.270%
placeboactual +0.313% · placebo +0.351% · excess -0.038% ± 1.826% (t=-0.02 against a threshold of 2.05, 29 real groups, 1,611,250 sham dates, draw error ±0.025%) — the status flips inside the placebo's own Monte Carlo error
benchmarktechnique +0.31% · buy and hold (same horizon) +0.37% · excess -0.06%
out of sampleasset half A: +0.485% (t=0.51, 29 episodes) · asset half B: +0.136% (t=0.14, 29 episodes) · liquid half (>= US$ 2,059,212/day): +0.058% (t=0.07, 29 episodes) · illiquid half: +0.627% (t=0.45, 28 episodes) · period 1/4 (2017-09-04 a 2022-01-03): +3.383% (t=2.28, 15 episodes) · period 2/4 (2022-01-05 a 2023-09-23): +0.740% (6 episodes — too small, does not count) · period 3/4 (2023-09-24 a 2025-03-18): +0.943% (6 episodes — too small, does not count) · period 4/4 (2025-03-19 a 2026-07-26): -3.808% (5 episodes — too small, does not count)
multiple testing6 variation(s) tested · t=0.33 across 29 episodes (equivalent to t=0.31) · p≈0.7540 · false positives expected by chance ≈ 4.52

Equity — outside the six controls, and here is why

The t of the trade series is invariant to bet size: 0.5%, 1% and 3% agree to the sixth decimal. Nothing here moves the verdict — it moves what the account would have lived through.

risking 1.0% per trade: ×1.14(×0.88–×1.36 depending on the drawn ordering)
worst drawdown from the peak: −87%
days below the previous peak: 1,056
signals refused for lack of capital: 48%
paths where the account halved (out of 12): 0

Reproducibility

period: 2017-08-17 to 2026-07-28
assets that traded: 529
variations tested before this one: 6
gross per trade: +0.51%
net per trade: +0.31%
exit rule: the technique itself (held until the opposite signal)
median duration bars: 6
mean duration bars: 12.7
max duration bars: 803
fee per leg: 0.001
episode days: 112
seed: 20260728
fibonacci level: 0.618
fibonacci filter: 0.13

Binance spot klines (delisted pairs included) · collected from 2026-07-27 23:31 to 2026-08-03 10:49 · 540 assets · 750,934 bars · 2017-08-17 to 2026-07-28

Hypothesis, filed before the result

After a confirmed up leg, buy when price retraces 61.8% of it — «a US$ 5 advance from start to peak should retrace US$ 3.09 before rising again». The source gives the LEVEL and does not give the exit; the target adopted is the one it predicts itself in the next sentence («a move that has retraced 0.618 should test the previous highs»), and the invalidation is a break of the low that originated the leg — without it the position would have no end. Both choices go declared on the card. It depends on the swing filter, which enters as a sweep. Family-level prediction, recorded before measuring: (1) NONE survives the corpus-wide Benjamini-Hochberg — part 74 measured that this requires ~3.94% per trade in crypto, and retracement techniques enter against the move, where the average effect is small by construction; (2) the gravedigger will be the PLACEBO, and not cost nor the invariant — this family's central claim is that specific levels (0.618, 0.382) matter, and the placebo is exactly the control that asks whether any level would have done the same; (3) the two retracements (0.618 and 0.382) will give results STATISTICALLY INDISTINGUISHABLE from each other — if the golden ratio had anything special about it, it would separate from its own complement; (4) automated Elliott will do BETTER than the retracements, because it is a trend system in disguise (the 5 against 35 oscillator is a moving average crossover) and not a bet on a level. Prediction (4) is the one I would least like to confirm, and that is why it is written down.

filed on 2026-08-04, before the number existed

The original, as it was filed

Depois de uma perna de alta confirmada, comprar quando o preço recua 61,8% dela — «um avanço de US$ 5 do início ao pico deve recuar US$ 3,09 antes de subir de novo». ⚠️ A fonte dá o NÍVEL e não dá a saída; o alvo adotado é o que ela mesma prevê na frase seguinte («um movimento que recuou 0,618 deve testar as máximas anteriores»), e a invalidação é o rompimento do fundo que originou a perna — sem isso a posição não teria fim. As duas escolhas vão declaradas no card. ⚠️ Depende do filtro de oscilação, que entra como varredura. Previsão da família, registrada antes de medir: (1) NENHUMA sobrevive ao Benjamini-Hochberg do corpus — a parte 74 mediu que isso exige ~3,94% por operação em cripto, e técnicas de retração entram contra o movimento, onde o efeito médio é pequeno por construção; (2) o coveiro será o PLACEBO, e não o custo nem o invariante — a alegação central desta família é que níveis específicos (0,618, 0,382) importam, e o placebo é exatamente o controle que pergunta se qualquer nível teria dado igual; (3) as duas retrações (0,618 e 0,382) darão resultados ESTATISTICAMENTE INDISTINGUÍVEIS entre si — se a razão áurea tivesse algo de especial, ela se separaria da sua própria complementar; (4) o Elliott automatizado irá MELHOR que as retrações, porque ele é um sistema de tendência disfarçado (o oscilador de 5 contra 35 é um cruzamento de médias) e não uma aposta num nível. ⚠️ A previsão (4) é a que eu menos gostaria de confirmar, e é por isso que ela está escrita.

Quotations from the source were translated from the Portuguese record and back into English — they are not the author's exact words.

Pre-registration exists to keep prediction apart from rationalisation: written after the number, every hypothesis is right.

This claim has been audited once — there is no history to compare against.