The false break — the one testable half of a liquidity sweep — beat buy & hold on 12 of 47 famous assets
Find a version of this that genuinely beats buy & hold and you have found something rare and worth the hunt — but first you have to split the part of a liquidity sweep you can actually test from the part you never can. The sweep is the well-worn idea that price lunges past a prior low to trip the stop orders resting there and then snaps back — a stop hunt. The catch is that those resting orders never appear on any price feed, and the word “sweep” is only awarded after price has already turned, so the core claim can never be shown wrong — and a claim no outcome can contradict is a claim no backtest can check. What closed bars do leave behind is a shape: a new multi-bar low that a later bar closes back above — the failed push lower, the false break. That one half is what we ship and test, the same way as everything here — the low and the reclaim are read only from bars that have closed, every fill is charged fees and slippage, and the result is always set against simply buying and holding. No kind window, no deleted losers, so a beat here is honestly won. Here is the whole glass: every famous asset, every winner and every loser.
So — does the false break beat simply holding?
On the honest evidence: only rarely. Of the 47 famous assets with enough reclaims and history to judge, the false break came out ahead of buy & hold on 12 — 26%. A handful genuinely came out ahead — but across dozens of assets a few beats are exactly what pure chance hands you, so the burden of proof stays on the idea, and one chart of a sweep that reversed on cue is not that proof. And the middle of the pack didn't just miss by a hair: the median false-break run landed at -40 pts against buy & hold over the window.
These are famous names that still trade today. Assets that delisted, went bankrupt or (for crypto) died or were rug-pulled aren't in the set (survivorship bias) — and that history isn't available from the free public data this tool runs on — so this rate isn't a representative base rate for every asset that has ever traded.
Where it caught the turn, where it got run over
The number that teaches the most here is not the average — it is the range. The one false-break rule, left completely unchanged, swings from a clean beat to a rout purely on which asset it is aimed at.
XRP — here the false break finished +343 pts ahead of buy & hold. Where the reclaim added anything, it was usually by catching one real bottom and holding the recovery that followed — as much a verdict on that asset's shape as on the rule.
Nvidia — the false break finished -863 pts behind buy & hold. On a strong, steady climb there are few breakdowns to reclaim, so the rule sits in cash through much of the rise, and the reclaims it does take get churned away by costs — being under-invested in a trend is the worst place a long-only rule can be.
That inconsistency is the real answer to “does the liquidity sweep work”: the outcome is not a property of the sweep at all, it is a property of the asset and the regime it ran in. A reclaim can mark a true bottom on something that ranges and get trampled on something that keeps falling — and the past cannot tell you which of those futures you are standing in.
Why a liquidity sweep looks like it always marks the bottom
- • You are shown the one sweep that snapped back and launched the rally. Choosing that chart after you already know price turned is hindsight, not evidence — the whole glass above is what is left once the chart can no longer be chosen for you.
- • The pushes that broke lower and simply kept going are never circled. A move only earns the name sweep once it has reversed, so the eye only ever meets the ones that turned — the failures were never labelled to begin with. Our rule counts every reclaim the geometry ever formed, the ones that held and the ones that broke straight back down.
- • The trading costs are quietly left out. Treating every reclaim as a fresh position means heavy turnover, and fees plus slippage on each round trip drag many textbook sweeps underwater — the friction was always real; most walkthroughs just never subtract it.
What we can test here, and the two things nobody can
We backtest the one half of a liquidity sweep that closed bars can actually show — the failed breakdown that reclaims — and decline the two halves that can only ever be read in after the outcome is known.
A liquidity sweep makes three claims in one breath. Two of them can never be shown wrong, which is precisely why they can never be tested; the third is plain price geometry, which is why it can. Here is where each part lands, graded only by whether any data could ever contradict it.
The premise is invisible. The whole story turns on a pool of stop orders sitting just past a prior low, waiting to be tripped — but a price feed records only the trades that printed, never the orders left resting and unfilled. The thing a sweep is said to hunt is absent from every chart and every backtest, and a premise nothing can measure is a premise nothing can refute.
The name is handed out at the end. The identical push past a prior low is called a sweep when price springs back and merely a breakdown when it keeps falling — and which one it was is settled only after the outcome is already on the screen. A label awarded by the ending can never be wrong about the ending, so there is nothing left for a test to disprove.
Set both untestable claims aside and what remains is a shape, not a story: a new low beyond the last N closed bars that a later bar closes back above — the failed push lower that reclaimed the level. That shape is on the chart whether or not anyone was swept, it is fixed the moment the bars close, and it is exactly what the False Break rule reads and backtests in the scoreboard above.
One caution, because this is where the honest version gets twisted. Observable means only that the rule can be backtested the same way every time. It does not mean the shape foretells anything, it does not mean it beats holding, and the whole-glass result above is the honest reason to expect that, pointed at your own asset, it most likely lags simply holding — the way most rules do. Isolating the testable half is what lets you find that out for yourself; it is not a hint about what you will find.
This split can move. If another piece of the idea turns out to have a closed-bar definition a machine can read the same way twice, it moves from the untestable column into the tested one — this is never a claim that the tested half carries an advantage.
Every false-break backtest is public — the beats and the washouts, nothing tucked away. Read the whole False Break scoreboard, then point it at the asset you actually care about and read, honestly, which side of the line it lands on.
Figures are computed live from simulated backtests over roughly the last five years and update as new bars close. False break here is the standard long-only rule — a later bar closing back above a support level that a new 20-bar low had just broken, with the reclaim arriving within about five bars, held until price closes beneath the lowest low of the last ten bars; other lookbacks behave differently, which is the whole point — any of them can be tested here. Hypothetical / simulated results — Past performance is not a reliable indicator of future results. This is an educational, analytical tool — not investment advice, and not a recommendation to buy or sell anything. You make all decisions and execute on your own broker.