A fugue takes a single idea and restates it in voice after voice, each entering against the last. That is what this is: five trading strategies built from two underlying signals, run side by side on your own Binance account so you can watch how the same idea behaves at different risk, on different instruments, at different speeds.
Free · no card · your keys never grant withdrawal access
The voices
Two signals underpin all five. Everything below — including the parts that failed their own tests — is stated on the strategy's own page inside the platform, permanently, not just here.
Ten liquid futures majors. Lower turnover, 3× leverage cap. Meant to be left running.
The most promising result here, and still not proven: it came in below the standard set for it before testing began. The learning layer on top of it cannot yet be told apart from chance.
The Steady signal on a wider market list, 7% of capital per trade, up to 10× leverage.
Leverage multiplies whatever an unproven signal does next, in both directions. This is a louder version of Steady, not a better one.
A directional model on ten futures majors. No extra leverage, no added turnover.
This signal did not hold up when tested on data it had never seen. It has no demonstrated edge. It is here so you can watch it run, not because it is expected to make money.
Buys Binance options — calls and puts on the majors — using the Direction signal.
A different shape of risk: the most you can lose on a trade is the premium paid, with no leverage and no liquidation. But it inherits Direction's lack of edge, so a run of options expiring worthless is a realistic outcome, not a freak one.
Tokenised US stocks and ETFs on Binance. Long only, no leverage.
In testing, its results could not be told apart from random chance. There is no real-order routing for this strategy at all — it cannot trade real money even if you ask it to.
Scope
Split deliberately. Everything in the left column is running right now; everything on the right is intended and not built. Nothing on this page is described as finished when it isn't.
How it works
Paper mode needs no exchange key at all. You only connect Binance when you want to see your real balances, or trade for real.
Email and password. Paper trading is open immediately — no card, no exchange key, nothing to lose.
Choose paper or real, and set the capital yourself. Real money asks you to type the amount back before anything starts.
Live equity, open positions and every signal that fired. One button closes everything and shuts it all down.
Pricing
You trade on your own connected account and we never hold your funds — so charging a share of your P&L would be both the wrong incentive and the wrong relationship. These are introductory prices while this is small, and none of these strategies has a proven net-of-cost edge yet — which is exactly why pricing is for access, not results.
Every strategy, paper mode. No card, no exchange key. Where everyone should start.
Real-money trading on your own connected Binance account.
For running more than one book at the same time.
Fugue Capital is a research project open to a small group of users. Nothing here is investment advice, and no number shown anywhere on this platform — backtested, paper, or live — is a promise of future performance.
Direction has no demonstrated edge: out-of-sample testing found no reliable signal, and it is offered for observation only. Options inherits that same signal. Steady's base signal showed a modest, repeated edge in backtest that has still not cleared the standard set for it before testing; the learning layer above it has not been shown to beat chance. Amplified runs that same unproven signal at substantially higher risk. Equities could not be distinguished from random in testing.
Trading with real money carries real risk of loss, up to and including everything you allocate. You are responsible for your own decisions and for the laws of your own jurisdiction. This platform never takes custody of your funds: you connect your own exchange account with an API key, and a key that can withdraw is refused outright — at connection, and again every time a strategy starts.