How MoonMath calculates past entry backtest

The formula

The multiple is the exit close divided by the entry close, and the return is that minus one. A stated amount is carried through the same ratio: it buys amount ÷ entry close tokens, valued at the exit close. A requested date resolves to the nearest observation only if it lies within half the series' own spacing; otherwise the calculation is refused.

What it reads

• A series of price observations with timestamps in seconds
• The entry date to analyse
• The amount to invest, so the result is stated in money rather than as a multiple
• Optionally an exit date, defaulting to the most recent observation

What it assumes

• A candle's closing price is the price attributed to that moment; nothing within the interval is modelled
• The whole amount executes at that day's closing price: no fee, no slippage, no partial fill, and no effect on the price it is paying
• The tolerance comes from the series' own spacing, so a denser series answers more precisely and a sparser one refuses more often
• At least two usable observations must exist

Run it

Pick a past date and an amount, and see what that entry would have produced by the most recent observation in the series — as money, not as a multiple, with the price chart it was read from. The result is stated against the two observations it was computed from, not against the dates that were typed, and a date the series does not cover is refused rather than answered with a neighbouring candle.

How the market itself works

This page describes MoonMath's arithmetic. The mechanics section describes the market: how an order book forms a price, why order-book depth is not the same thing as pool liquidity, and what a public book does and does not let you observe.