Triangular FX Arbitrage strategy: rules and backtest
Triangular FX Arbitrage extracts profit from a mispricing of three currencies along a closed chain of conversions.
| Parameter | Value |
|---|---|
| Type | Mean-reversion |
| Timeframe | W1+ |
| Complexity | Pro |
| Instrument | USD/JPY |
How the signal works
The strategy tracks three currency pairs forming a triangle (for example, EUR/USD, USD/JPY, EUR/JPY) and looks for a moment when the product of the rates around the loop deviates from unity. Such a deviation lets you walk the chain of conversions at a profit without market risk.
Pure arbitrage of this kind on liquid markets has all but vanished: mispricings are spotted and removed within fractions of a second by automated systems. The available window is measured in milliseconds, and the competition is at the level of infrastructure speed, not the idea.
Why this strategy cannot be honestly tested on our data
Arbitrage by definition requires at least three synchronised instruments, and the available data is a single pair, EUR/USD. Building an arbitrage chain from it is impossible in principle, not for lack of history.
We deliberately show no backtest here: presenting attractive figures computed on unsuitable data would mislead the reader.
Pros and cons
- The profit is almost independent of market direction.
- Theoretically market-neutral.
- Based on a measurable mispricing, not a forecast.
- Pure arbitrage on liquid markets has practically vanished.
- The window of opportunity is milliseconds, a race for speed.
- Inaccessible without high-frequency-trading infrastructure.
Pitfalls
For a retail trader the main mistake is treating triangular arbitrage as an accessible opportunity: mispricings are removed faster than a human can react, by those very high-frequency systems. Trying to play on their field without the corresponding infrastructure is doomed to lose on speed.
Who it suits
A topic for understanding how the currency market works, not for use by a retail trader. Triangular arbitrage is the domain of high-frequency firms with infrastructure competing for milliseconds.
Frequently asked questions
Why can the strategy not be tested on a single pair?
Triangular arbitrage by construction works with three linked pairs. Profit is extracted from a mispricing along a closed chain, and on a single pair such a chain cannot be formed.
Does this arbitrage still exist?
On liquid markets it has practically vanished: mispricings are removed within fractions of a second by automated systems. The window of opportunity is measured in milliseconds.
Can a retail trader do this?
In practice, no: it requires high-frequency-trading infrastructure to compete on speed. It is the domain of specialised firms, not the individual trader.