Methodology
Nothing here is a forecast, nothing is advice, and every number is derived from data you can go and re-fetch yourself. Every cost figure you read here is a floor, because the rolling fee charged every 8 hours is levied at a rate Moon does not publish and is therefore left out of the arithmetic. The rest of the numbers come from a vendor price bar or from arithmetic on vendor price bars, and each assumption is labelled where it is used.
How the simulator prices your position
Your exposure is stake multiplied by leverage. A long makes
exposure x ((current - entry) / entry). A short makes
exposure x ((entry - current) / entry).
Which side of the book you are on matters more than that formula suggests. A long opens at the ask and closes at the bid, and a short does the reverse. Using the last traded price at both ends would quietly delete the spread and flatter every result. That convention is implemented and currently has no effect, because Moon publishes no spread and MarketMoves will not invent one, so the two sides of the quote meet at the mid. Third-party testers report a house spread inside Moon's settlement price, which would make your real results worse than the figures here.
Cost is modelled as Moon documents it and no further. Your opening fee is 1 per cent of the wager, which Moon's glossary states is "calculated solely on your wager amount" and "not calculated on your total leveraged exposure". So it does not scale with the multiplier and costs the same at 1000x as at 2x. An earlier version of MarketMoves charged a round trip on exposure instead, which overstated the cost of opening a 1000x wager about 160 times over, and every derived figure has since been rebuilt.
Two further costs are named and never multiplied out. A rolling fee is charged every 8 hours at a rate Moon calls dynamic and does not publish, and third-party testers add a performance fee on realised profit that Moon does not document. So treat every figure here as a floor: hold past 8 hours and you pay more than the page shows, by an amount Moon's own documentation will not let you compute.
Maintenance margin is 0 basis points, because Moon documents no buffer. Your wager closes once it has lost the wager, so your max loss price sits at exactly one part in the leverage: 1 per cent away at 100x, 0.1 per cent at 1000x. Your loss stops there. Moon states that an account balance "can never go below zero as a result of a bet".
Indicative example based on third-party market data. Moon's reference price, spread, fees, settlement and liquidation rules may produce a different result.
How the replay runs
The replay takes a fixed $100 stake at each multiplier of 2x, 5x, 10x, 20x, 50x, 100x, 200x, 500x, 1000x. Each one runs through the last 12 hours of one-minute bars for that instrument, long and short. A step dies the first minute the move against you from entry exceeds the distance its own leverage allows, after maintenance margin. That test uses the low for a long and the high for a short rather than the close, because a position does not survive a wick by being marked at the end of the minute. A dead step stays dead for the rest of the 12 hour window, with no re-entry.
Where minute coverage over the window is thin, you get the coverage figure instead of a curve built from gaps.
Volatility and range statistics
Realised volatility is the standard deviation of daily log returns over the stated session count, annualised by the square root of 252 for equities and 365 for crypto and metals. Average true range is the Wilder 14-session figure expressed as a per cent of price. Every figure carries the window it was measured over, so you can tell a 30-session reading from a 5-session one.
Derived daily bars
Where a daily bar is rebuilt from minute bars rather than supplied by the vendor, it is labelled as derived. A rebuilt bar and a vendor print are not the same object and the site does not present them as one.
The catalyst window
A move qualifies as material when it clears the larger of 1.4 times that instrument's own daily standard deviation, 1.5 per cent across the replay window, or 2.5 per cent on the session. The threshold is per instrument on purpose, because 2 per cent is a quiet day in one name and an event in another. Sources are searched back 36 hours. Opinion mills, price-target aggregators and press-release wires are excluded by name, and Form 4 and 13F filings are excluded as noise.
Moon availability
You only get an instrument-level Moon link where a person has confirmed the instrument is listed on Moon, and that confirmation expires after 30 days. Nothing checks it automatically, because no page here reads Moon's internal endpoints, so an unverified instrument gives you the reason there is no link instead.
Commercial assumptions, and their status
Fees, the multipliers on offer and the referral terms live in a dated configuration record, currently version 2026-08-28a. Nothing in it has been verified against a live Moon account, so every page using it labels it as modelling input, and the limits on the price feed are set out in data sources.