method
How it works, and where it breaks
Most strategy pages explain only the upside. This one leads with the conditions under which the approach is expected to lose — because that is the information a reader needs before committing capital.
The strategy in one paragraph
A regime-adaptive trend follower. The system reads the broad market using a long moving average on Bitcoin: above it, the book is positioned long; below it, short. Positions are selected by relative momentum across a basket of liquid pairs and sized so every position risks a similar amount. Trades are held until momentum decays or the stop is reached.
Position sizing
Every position is sized so that being stopped out costs the same fraction of the book regardless of which pair it is. Volatile pairs therefore take smaller positions. This keeps one bad trade from dominating the record.
- Risk per signal is fixed as a percentage of the book, scaled by that pair's recent volatility.
- Stops sit a multiple of average true range away, not at a fixed percentage.
- No take-profit — winners run, and close when momentum turns.
- Concurrent positions are capped so exposure cannot quietly concentrate.
Why the win rate looks low
A trend-following system is expected to lose the majority of its individual trades. Most positions never become trends, and those close small; a minority become large winners. Profitability comes from the size of the winners relative to the losers, not from being right often. A high win rate with a weak payoff ratio is usually a worse system, not a better one.
Where this is expected to fail
These are known limitations, measured on the strategy's own history — not hypothetical risks.
Long stretches of losing trades are normal
In the tested history the longest run of consecutive losing signals was long enough that any reasonable person would question the system. This is inherent to the style, not a malfunction.
Performance concentrates in trending markets
When the market moves sideways both directions tend to lose. No filter we tested fixed this, so we report it rather than tune it away.
Win rate and stability trade off against each other
Tuning for a higher hit rate consistently reduced total return, because it shrinks winners before they can pay for the losers.
A short record is not evidence
Below roughly 100 closed signals, results are dominated by chance. The live count is shown on every page so you can judge it against that threshold.
Costs and slippage are real
Exchange fees are charged on every entry and exit. Live results include them; a backtest that ignores them will always look better.
What the hash chain does and does not prove
The chain proves the record has not been edited after the fact: each signal was hashed before its outcome was known, and altering any entry breaks every hash that follows. It cannot prove the strategy will stay profitable. Those are different claims, and only the first is made here. A verifiable past is necessary for trust, not sufficient for it.
What would change our mind
The record is the test. If live results diverge materially from what this method predicts — particularly if the payoff ratio collapses — that is evidence the approach has stopped working, and it will be stated here rather than quietly re-tuned. Changes to the method are recorded as new ledger entries carrying the parameters that were in force.