About OrbitFx
A proprietary trading firm for crypto, built so that its own claims can be checked rather than taken on faith.
The problem this exists to solve
Proprietary trading firms sell evaluations: you pay a fee, trade a simulated account under a rule set, and if you pass you trade the firm's capital for a share of the profit. The model is sound. The execution, across the industry, has an unresolved trust problem.
The trader cannot see the price they were filled at being derived. They cannot check whether the slippage they paid was the slippage everyone paid. When an account dies one tick from the limit, they have the firm's word for it and nothing else. Every one of those gaps is somewhere a firm could put its thumb on the scale, and the trader has no way to tell the honest firms from the rest.
OrbitFx is an attempt to close those gaps structurally — not with a promise, but with an architecture in which the dishonest version of the platform would be harder to build than the honest one.
How that shows up in the build
The index is deterministic. The price computation has no clock reads, no randomness and no network calls. The same inputs produce the same output forever, which is what makes replaying a disputed account meaningful rather than approximate.
The execution delay cannot be varied per account. A deliberate delay between order and fill is the standard defence against latency arbitrage. Applied uniformly and published, it is honest execution modelling; applied selectively to profitable traders, it is fraud. The function that computes it takes no account, no balance and no P&L — it is not that we choose not to, it is that the code cannot.
The evidence log binds us too. Orders, fills and breaches are append-only, enforced by database triggers rather than by permissions. Permissions can be granted to yourself; the triggers stop the platform's own administrators the same way they stop everyone else.
Ambiguity resolves toward the trader. When venues disagree, or a feed goes stale, or the platform cannot be confident in a price, it halts trading rather than risk a wrong breach. A halt costs us volume. A wrong breach costs a trader their account and costs us the only thing we are selling.
What we deliberately do not do
We do not run a dealing desk, take the other side of trades, or hold client funds. Nothing a trader submits reaches an exchange — the accounts are simulated, which is the point of an evaluation.
We do not chart a price we are not filling against. The chart, the fill and the evidence pack all read from the same index series, which is why we do not embed a third-party price feed that would show a trader one number while filling them at another.
We do not publish performance claims we cannot substantiate. Pass rates, when we publish them, will come from the same records a disputing trader can audit.
We are in closed beta, and it matters
The rule parameters and the execution cost model are still being calibrated against historical market data. Until that finishes, every rule set carries a provisional version string — visible on the rules page and written into every account record, so it can never quietly become permanent.
Beta accounts are free. Payments, funded payouts and the wider instrument list open when the calibration and the legal work are done.
Read further
The trading rules are published in full before purchase, as is every parameter they contain. The FAQ covers what happens on a breach, how disputes work, and what automated trading is permitted.