Independent provider directory
The method Trust tests Playbooks FAQ The verified version
The method

How a crypto trading strategy actually works

A crypto trading strategy is a written bet about what a price will do next — and a discipline for proving, later, whether the bet was any good. The crypto-native version of that discipline is simple: commit the call in public before the outcome, so the record cannot be edited once you know the answer.

The mechanics, in order

From a written bet to a record you can check

The core engine: mean reversion

Most rules-based strategies rest on one observation: a price that stretches unusually far from a typical level tends, often enough to matter, to drift back toward it. A mean-reversion strategy takes the stretch as its signal — not a story about a project's future, just a measurable distance from a baseline — and positions for the snap back. The wager is not that you can outguess the market; it is that one familiar habit — overstretched prices snapping back — shows up often enough to profit from across hundreds of attempts, with every losing attempt kept in the tally. The diagram below shows the shape of it: the further the price strays from its own baseline, the larger the stretch the rule is built to trade.

How a mean-reversion setup places its entry, stop and targetSchematic of a mean-reversion trade. A price line falls a measured distance below its baseline (the dashed centre line), stretching further than usual. The rule places the entry at the stretched low, a stop a measured distance further down where a continued fall would prove the idea wrong, and a target back up near the baseline where the expected reversion is judged complete. The asset is unnamed; only the geometry is shown.baseline (mean) = targetstop — idea wrong below hereENTRYstretched low, on the rulethe stretch =the signalTARGETtime → (every level fixed before the position opens)expected reversion
Illustrative geometry, not a specific recommendation: the further a price strays from its baseline, the larger the stretch the rule trades — entry, stop and target are all set in advance, never after the candle prints.

The rule set, not the hunch

What turns the engine into a strategy is writing it down. A hunch says “this looks ready.” A strategy states, in advance and in writing, three things you can hand to a stranger:

  • Entry — the specific level the rules act on, not “somewhere around here.”
  • Stop — the level that admits the idea was wrong, decided before you are attached to the position.
  • Target — where the expected reversion is judged complete, so profit is taken on plan rather than on nerve.

The same logic, run end to end

An abstract rule is easier to trust once you have watched it run on a concrete example. Here is one, deliberately stripped of any instrument so only the reasoning is on show:

Worked setup · illustrative, not a recommendation

The figures below are invented round numbers for teaching the logic end to end. No instrument is named, and nothing here is a specific call. The reasoning is what transfers, not the levels.

  1. The stretch is the signal. A price has slid for several sessions and now trades at 2,000, while its own recent baseline — the middle of where it has lived — sits near 2,150. That is roughly a 7% stretch below the mean, wider than this market has typically strayed before turning. The rule reads the distance, not a story about why.
  2. Why this entry. The rule opens at 2,000 because that is the stretched level it was written to act on — not “somewhere cheap.” Entering on the named level is what makes the trade a test of the rule rather than of the urge to catch a bounce.
  3. Where the stop, and what invalidates it. The stop sits at 1,920, about 4% below the entry. Below there, the “stretch” is no longer a stretch but the start of a genuine new downtrend — the idea is admitted wrong and the trade is closed. The stop is the line that says the mean-reversion thesis failed.
  4. Where the target. The target is 2,130, just under the baseline the price reverted from. The expected snap-back is judged complete near the mean, so profit is taken on plan rather than on nerve. That is a reward of about 130 points against a risk of 80 — a little over 1.6 to 1.
  5. What a realistic outcome distribution looks like. A setup like this does not win every time, and an honest version says so. Across many such trades you might see a clear majority reach the target, a meaningful minority stopped out near the full 80-point loss, and a scatter closed early at the end of the window for a small gain or loss. The edge is not certainty on any one trade; it is a favourable average over a counted run with the losers left in.
The whole point: every level — entry, stop, target — was fixed before the position existed. On a verifiable method, those same levels are also committed on-chain before the outcome, so the closed trade is a clean test of a rule, not a memory coloured by how it felt.

Notice what the trader never does in that walk-through: decide anything after the position is open. Every level existed before the trade did. That single property — nothing improvised mid-trade — is what makes the resulting record meaningful, because each closed trade becomes a clean test of a rule rather than a memory of how the move felt.

Why crypto traders should demand verification

Crypto already runs on “don't trust, verify.” You would not accept a transaction because someone swore it cleared; you check the chain. A trading record deserves the same scepticism, because the incentives to embellish it are enormous. The cleanest proof that a strategy was a strategy and not a story is that each call was committed before its outcome was known — in a form a stranger can confirm without trusting the author.

The on-chain proof standard

This is where crypto's own tooling does the work. The verified method this desk uses anchors every call on-chain: the entry, target, stop and conviction grade are folded into a single SHA-256 fingerprint. Because the grade lives inside that fingerprint, the call cannot be quietly re-graded upward once it has already paid off, and because the timestamp is on Bitcoin, nobody — including the operator — can back-date it. The asset the model trades is beside the point; what matters is that the proof method is the one crypto traders already trust. The flow below is the whole mechanism, from a published call to a stranger re-checking it.

How a published call becomes a claim anyone can re-checkFlow diagram: a trading call is published with its entry, target, stop and conviction grade; those fields are folded into one SHA-256 fingerprint; the fingerprint is anchored to a Bitcoin block at publication; later anyone can re-hash the published call and confirm it matches the on-chain receipt, proving the call was fixed before its outcome was known.PUBLICATION TIME → (the receipt is dated before the trade can resolve)A match proves nothing in the call was touched after the outcome was known.1 PUBLISHentry / target /stop / grade+ signal time2 FINGERPRINTone SHA-256 ofthose fields —change one, all change3 ANCHORwritten to aBitcoin blockat publication4 RE-CHECKanyone re-hashesand matches thepublic receipt
Don't-trust-verify, drawn out: the call is frozen on a public ledger the instant it is published, so it cannot be quietly re-priced once the result is in.

That is why the example here is the #1-ranked provider's four mean-reversion models: same engine, written rules, every call verifiable before the outcome is known. The conviction on each call is graded, and the bar for the top grade is set against each model's own returns — which is why the same letter can mean a very different raw move on a fast clock than on a slow one:

Grade-A bar by model, set against each model's own measured returns. The clock column describes how long a position is carried — nothing here states what is traded.
ModelHolding clockGrade-A bar (per trade)
Day Tradesame session, a zero-to-sixty-minute window0.70% avg / trade
Multi Hourhalf a session out to two sessions4.50% avg / trade
Swing Tradethe flagship, carried roughly seven to twenty-eight days6.00% avg / trade
Investingcarried over a long horizonlong-horizon

An A is the top band of a model's own return distribution and a D is the lowest still published. The bar is set per clock, so an A on a same-session Day Trade call (around 0.70% a trade) and an A on a multi-week Swing call (around 6.00%) both read as “top-band for this horizon” rather than one absolute number stretched across very different holding times. There is no E grade — it was retired so the four-step scale keeps its meaning.

From here the natural next steps are the trust tests that decide whether any strategy clears the bar, and the playbooks for building and verifying one of your own.

Keep reading