Backtest funding routes before they become positions
A scanner finds candidates. A backtester helps decide whether the recent funding behavior was persistent enough to deserve risk.
Why backtesting matters
Funding routes can appear attractive for a single snapshot and then disappear. Backtesting helps reveal whether the edge persisted across multiple funding periods.
SypherScore backtests route behavior using historical funding and price context so traders can compare signal quality before execution.
- Review positive funding periods.
- Compare estimated PnL and drawdown context.
- Check whether break-even time fits the intended hold.
How to interpret results
A strong backtest does not guarantee a future result. It simply reduces blind decision-making by showing whether the route had enough historical support to continue reviewing.
- Use recent windows for fast-changing DEX routes.
- Use longer windows for stability checks.
- Recheck live funding before entry.
What a backtest cannot tell you
A funding backtest replays what the rate did. It does not replay what the book did, and that gap is where most paper edges die. A route that looks profitable at 40% annualised can be untradeable because the leg you need is thin, because the venue moved its funding interval, or because everyone else saw the same number at the same time.
Treat the result as an upper bound on a good day, not as an expectation. The useful question is not «how much would this have made» but «how much room is there between this and the cost of running it».
- Depth is not in the data: a rate says nothing about whether you can get filled at size.
- Funding intervals change. A venue moving from 8-hour to 1-hour funding rewrites the whole series.
- Crowding is invisible in a backtest and very visible in a fill.
Choosing the window, and why it decides the answer
The single most common way to fool yourself is picking a window that contains one regime. Funding is regime-dependent: a trending market pays the short side for weeks, then reverses and pays it back. A backtest over that first stretch shows a strategy that cannot lose.
Run the same route over at least three separate windows that include a reversal, and look at the worst one rather than the average. If the worst window is survivable, the route is real. If the average looks good only because one stretch carried it, you have measured a market regime, not a strategy.
- Use at least three windows, and make sure one of them contains a reversal.
- Judge by the worst window, not the mean: the mean hides the drawdown that ends the position.
- A route whose result collapses when you shift the window by a week was never a route.
Questions
Does a positive backtest guarantee profit?
No. Market conditions, liquidity, funding schedules, and execution can change immediately after a backtest window.
When should I backtest?
Backtest after a route appears interesting in the scanner and before treating it as an actionable position.
How long should a funding backtest window be?
Long enough to contain at least one reversal of the funding regime, which in practice means weeks rather than days. Anything shorter measures the current trend rather than the strategy, and the current trend is exactly what will change.
Why does my live result differ from the backtest?
Almost always execution and depth. The backtest assumes you were filled at the observed rate on both legs; live, one leg fills at a worse price, or fills late, or does not fill at all. The second most common cause is that the funding interval or the fee tier changed and the historical series no longer describes the venue you are trading.
Can a backtest account for slippage?
Only crudely. You can subtract an assumed cost per round trip, which is better than nothing, but the real cost varies with size and with how busy the book is at the moment you trade. Use it as a sanity check, not as a number to plan with.