We built a trading strategy. We froze the rules for judging it before seeing the results.
The verdict arrives at 40 trades — around September 2026. We don't know the answer yet, and we've made it impossible for ourselves to pretend otherwise.
Updated manually as trades resolve. The strategy trades rarely by design — silence is selectivity, not failure.
The three tests it must pass — frozen on 2026-06-11, before the data
The full numeric thresholds live in our research record and are version-controlled; they were reviewed once and cannot be changed to fit the results. In plain language, at 40 trades the strategy must show all three:
1 · It makes money after real costs
Average profit of at least +0.10R per trade after subtracting the bid/ask spread we actually recorded at every entry and exit — not an assumed cost, the measured one — with basic statistical confidence (t ≥ 1.3). Trades we can't cost honestly are flagged, never silently filled in.
2 · It wins for the reason we claimed
The strategy is supposed to work because gold pays breakout traders only in trending regimes. So the trades, bucketed by how trendy the market actually was, must show the trending bucket carrying at least 60% of the net profit and beating both other buckets — with at least 8 trades per bucket. A profitable record that wins for the wrong reason fails this test.
3 · Its filter picks the right days
The strategy uses a filter that decides, each morning, whether today looks like a trending day. We score that filter on every day — including days with zero trades: the days it switched ON must turn out trendy at least 10 percentage points more often than average, and the days it switched OFF must not. A filter that's secretly a coin flip fails, even if the account made money.
If all three pass, the strategy earns a discussion about trading tiny, bounded, real size. If any fails, it stays in the lab — and you'll read exactly why, with the same numbers we read. Either way, the verdict publishes. Our research record is full of negative results; we consider them the product. See the Graveyard.
Be there when the verdict drops
One email when the verdict publishes, and a handful of research notes between now and then. No signals. No spam.
The form above is our newsletter provider embedded securely — you stay on zenflare.digital the whole time. After you subscribe, check your inbox for a one-click confirmation email; once you confirm, you're on the list and we'll email you the day the verdict drops. No signals, no spam.
What ZenFlare is
ZenFlare is a measurement institution for retail trading claims. We record millions of market ticks ourselves, test strategies against them with an honest simulator (one that models the ways stops really fail), run survivors in shadow against live markets without risking a cent, and publish what we find — especially when the finding is "this doesn't work, and here is exactly why."
Nothing here trades live. The strategy on trial runs in shadow and on paper behind a discipline layer with hard limits and a kill switch — and every attempt to break those limits is written to a permanent deviation log. Honesty isn't a slogan here; it's the architecture. How the machine works →
Nothing is for sale here today. No signals are provided. If signals are ever offered, it will only be after three gates have opened in public: the Trial passes its frozen bar, the strategy survives live testing on our own account, and the proper regulatory licensing is in place. Until all three — nothing.