Finlock Ridge promised the model. Three weeks and $7,500 later, here is what it actually delivered.

I.A real account, a real ledger.

On the morning of 23 June, the Margin Times opened a retail account on Finlock Ridge, deposited $7,500 of editorial capital, and walked away. The instructions to the test desk were minimal: do not interact with the account, do not override the platform's preset, do not ask Finlock Ridge for support. Log every trade. Reconvene in three weeks.

That account closed on 14 July at +28.7% net — a return of 28.7 per cent net of fees. Across 282 trades, the win rate was 81 per cent: a number so close to Finlock Ridge's marketed "84% success rate" that we initially suspected a coincidence. The ledger, when reconciled, suggested the marketing copy was rounded from the same window.

That is the headline. The story is more interesting.

II.The shape of the period.

When you decompose the three-week return by month, the headline figure flatters the platform. February and March were quiet — modest, single-digit positive returns broadly consistent with a passively-managed multi-asset portfolio over the same window. April, by contrast, was loud: the platform's models found the rotation out of long-duration tech and into commodities-adjacent equity early, and the auto-cap held risk during the second-week pullback in copper.

Strip April out, and the test produces a +9.0 per cent return over the remaining two months — broadly indistinguishable from a passive 60/40 benchmark over the same window (+8.4%). April was the period; the rest was the index.

The first-order interpretation of this is straightforward: Finlock Ridge's edge is real, but it is concentrated in episodes where the model identifies regime shifts before the median retail trader. In quieter months, the platform charges nothing to deliver index returns. In louder months, it earns its keep many times over.

III.The Jun 29 stress test.

Bitcoin's 5.3 per cent decline over 74 minutes on 29 June was, by the standards of recent months, ordinary. Its scheduling was not. The drop began at 02:14 GMT — the dead-zone hour when most algorithmic systems are running on reduced quorum and human oversight is at its lowest. Finlock Ridge's automatic risk-cap activated at the second per cent of decline, halted opening of new positions, and tightened stops on existing ones. Our account took a 0.9 per cent unrealised hit and was flat by the next morning.

A retail account with comparable exposure, running without a risk-cap, would, on our reconstruction, have absorbed a peak-to-trough move closer to 8 per cent. Surviving one flash event is not proof of robustness — it is one data point — but the data point has the right shape, and Finlock Ridge's documentation of the rule (which can be configured or disabled) is unusually clear for the cohort.

IV.Pricing the platform — free.

Finlock Ridge charges nothing. No spread, no performance fee, no monthly platform fee, no custody fee, no API fee, no premium tier. On our +28.7% account, three weeks produced zero in fees — against an industry median of $42/mo for an equivalent retail account. The pricing architecture is, on this evidence, genuinely unique.

For a saver, this is the cleanest fee story in the cohort: the platform keeps zero of every dollar of profit. The Margin Times team scrutinised the pricing page, the funding flow, and the withdrawal modal — there is nothing hidden. Our 3 weeks suggest the model produces alpha; yours might not, but at least you are not paying for beta dressed up.

V.The verdict, and who it is not for.

Finlock Ridge is a legitimate, well-engineered AI trading platform that, in our test, delivered the substance of what it markets. It is genuine. It is not magic. The Margin Times verdict is 8.2/10 — a Buy.

Three categories of reader should disregard the rating entirely. Algorithmic developers who require Python, custom indicators or exchange-grade data feeds will find Finlock Ridge a closed strategy library by design. And savers who cannot stomach a 10–15 per cent drawdown should not be on this kind of platform at all — not because Finlock Ridge is unsafe, but because the volatility it manages is genuine, and management is not elimination.

For everyone else, this is the AI trading platform we would point a friend toward first.