Fees & Slippage
Transaction costs are applied inside the keyless DRY_RUN backtest that produces every
figure, so all published returns, curves and stats are net of these costs. The
assumptions are fixed and documented here.
The assumptions (as applied)
The backtest broker is configured with:
| Cost | Value | How it's applied |
|---|---|---|
| Commission | 0.0005 = 5 bps | charged on each fill (entry and exit), on notional. |
| Slippage | 2.0 ticks | every fill is moved 2 ticks adverse of the bar price (a tick = the instrument's minimum price increment). |
A round-trip therefore pays commission twice plus slippage on both fills. These are the same across all strategies and the whole window.
:::note Net of costs Because costs are inside the simulation, you do not need to deduct anything yourself — the CAGR, drawdown, Sharpe, returns and curves already reflect 5 bps/fill commission and 2-tick slippage. :::
What is NOT modelled
:::warning Honest cost caveats
- Perpetual funding is not modelled.
BTC-PERPETUALcharges/credits periodic funding; the backtest does not include it. For a strategy that holds positions across many funding intervals this is a real omission — net live results would differ by the accrued funding. - Market impact / size is not modelled. Slippage is a flat 2 ticks regardless of order size; large orders that would move the book are not penalised beyond that.
- Slippage is fixed, not volatility-dependent. In fast or thin markets, real fills can be worse than 2 ticks; here they never are.
- Commission is a flat 5 bps, not the venue's exact maker/taker fee schedule, rebates, or fee tiers.
- Fills occur at bar granularity (the backtest is bar-by-bar), not tick-by-tick. :::
These assumptions are deliberately simple and conservative-ish but not exhaustive. Treat the figures as net of an explicit, fixed cost model — accurate within that model, and to be adjusted by a reader who wants to add funding or impact for their own venue and size.