Every strategy on SureFX goes through the same five steps, and nothing skips a step. This page says what each step is, what it rules out, and what you see on the strategy's card as a result.
A strategy is a fixed set of rules for one instrument in one direction, with the stop and the target decided at entry. Strategies that share a mechanism form a family; the family runs as one indicator and each strategy is a preset inside it, with every constant exposed as an input.
We keep the rules few and legible. A strategy you cannot explain in a paragraph is a strategy you cannot trust in a bad month.
Before anything runs live, the rules are replayed over the history they were designed on. The result is written down as a single number: the average result per trade, measured against that trade's own risk, which we call R. That number is the backtest figure on the card, and it never changes afterwards.
Backtests are where every strategy looks good. We publish them because they are the promise the live record is measured against, not because they are the evidence.
The same brackets are placed at random moments inside the same market conditions, many times over. If the strategy's entries do no better than random ones, the pattern is not doing any work, however good the backtest looked. Those strategies are retired before they reach the catalogue.
The viral strategies in the archive get exactly this treatment. Most fail it. Ours are not exempt.
Each strategy runs on a paper account at OANDA from the day its current rules went live. Entries fill at the live bid or ask at the moment the order is placed, so spreads and entry slippage are in the record; entries that have moved too far from the intended price are rejected. We state the fill model on every card because it is the difference between the record and a real account.
When a strategy's rules change, its record starts again from that date and the earlier window stays visible. A record that mixes two versions of the rules is not a record of either.
Each card carries a confidence range around the live result. With a handful of trades the range is wide; with hundreds it narrows. A strategy is called proven only when the whole range sits above zero, which is a claim about evidence, not a forecast. Until then the card says so.
The live record is large enough, and strong enough, that chance no longer explains it.
Running live, results published, not yet enough trades to tell the strategy from luck. Most of the catalogue at any given time. Rentable, with that said plainly on the card.
Stopped, with the reason on the card: failed the null, rules changed, or the live record said no. The record stays public.
You can rent any strategy in the catalogue, proven or still building a record. Some members want the backtest and an early look, others wait for the evidence. The card shows the backtest, the live trades so far and the confidence range at the moment you decide, and the choice is yours. We do not hide a strategy behind a threshold, and we do not dress a short record up as a long one.
Founding members get day-one access to every strategy, its live record and the archive, at a rate locked for as long as they stay.