How long you hold decides what leverage you get away with
Hold 500× for one minute and you will probably still be there. Hold it for one week and you will not: across 46 markets the highest multiplier that came through one week intact was 5×. That is 3 leverage levels down the leverage range for the same trade, and the only thing that changed was the clock. Over one session, the length most positions actually last, 26 of 46 markets top out at 20×.
A leverage level counts as survived when its distance to liquidation stayed clear of the 95th-percentile adverse move over rolling windows of that length, taking the worse of the long and the short side.
the same trade loses 3 leverage levels as the clock runs
the widest session moves, drawn
The 8 widest one-session moves among the 50 tracked markets, each tile carrying the last 60 daily closes on its own vertical scale so the shape is comparable while the level is not. Strategy (MicroStrategy) leads the strip at +11.98% on the session dated 2026-08-27, against +3.01% at the far end of the ranking.
Session change is the vendor's last completed daily bar against the one before it. Ranking is on absolute change, so falls are ranked alongside rises rather than dropped, and 0 of the 8 tiles shown are negative.
where the tape actually moved
Strategy (MicroStrategy) leads it up at +11.56% on the session printed 2026-08-27. At 100x the stop sits 1.00% out, and MSTR covered that much on 81% of the last 90 sessions.
| Market | 30d trend | Session | Last | Vol 30d | Session leverage level | 100× hit | Window |
|---|---|---|---|---|---|---|---|
| Strategy (MicroStrategy) MSTR | +11.56% | $137.39 | 74% | 10× | 81% | 2026-08-27 | |
| NVIDIA Corporation NVDA | +8.62% | $228.05 | 43% | 20× | 67% | 2026-08-27 | |
| Coinbase Global COIN | +4.96% | $190.75 | 76% | 10× | 79% | 2026-08-27 | |
| Palantir Technologies PLTR | +4.75% | $185.90 | 92% | 20× | 68% | 2026-08-27 | |
| Broadcom Inc. AVGO | +4.41% | $371.56 | 42% | 20× | 74% | 2026-08-27 | |
| Intel Corporation INTC | +4.32% | $92.07 | 72% | 20× | 86% | 2026-08-27 |
Session change is the vendor's last completed daily bar against the one before it, dated in the final column. The trend cell plots the last 30 daily closes on each market's own scale, so the shape is comparable and the level is not. Session leverage level is the highest Moon multiplier whose distance to liquidation exceeded the 95th-percentile adverse excursion over rolling one-session windows, taken from survival.json at the 95th percentile. 100× hit counts the share of the last 90 sessions that travelled the 1.00% a 100x position has to spare.
305 of 828 replayed positions liquidated
Every leverage level on Moon's leverage range was replayed long and short on one-minute data for 46 markets over a median replay window of 5.8 hours, and 305 of the 828 resulting positions were liquidated, 36.8% of the sample. The upper bar in each pair is the long side, the lower bar the short side, and the pair reaches its full width when every market killed that leverage level. At 1000× the rates were 83% long and 98% short across 46 markets. The lowest leverage level with any liquidation at all is 10×, at 0% of longs and 2% of shorts.
Replay applies Moon's opening fee of 1% of the wager and its liquidation distance of 1.00% at 100x, and it excludes the rolling fee charged past 8 hours because Moon does not publish that rate, which makes every mortality figure here a floor. One window per market is one sample, not a distribution, so the 36.8% headline describes the replayed window rather than an expectation.
Bitcoin: what the hold costs in room
Adverse excursion on Bitcoin rises with the holding period, from a 511-window session median of 0.78% to a 95th percentile of 1.55% and a 99th of 1.68%. The weekly figure is 11.4 times the hourly one, which is the mechanism behind a survivable leverage level that falls as the hold lengthens. Set against Moon's leverage range at the session horizon, the highest leverage level whose distance to liquidation clears the 1.55% figure is 50×, out of 9 leverage levels offered on this market.
Adverse excursion by holding period
Room to liquidation, one session
Excursion is the worst move against the position inside a rolling window of the stated length, taking the worse of the long and short side, measured on 511 session windows of 390 1min bars. Room to liquidation is Moon's own arithmetic at a maintenance rate of 0.00% of notional.
a flat fee crosses a notional fee at 23.1×
Opening a $100 wager costs $1.00 on Moon against $0.09 on Hyperliquid at 2×, and $1.00 against $1.73 at 40×, which is the highest leverage level Hyperliquid offers on this market. Moon charges on the wager, so its line is flat at every leverage level, while Hyperliquid charges on notional, so its line rises in proportion to leverage and the two cross at 23.1×, or 69.4× on maker fees. Hyperliquid is the cheaper open at 2× and Moon is at 40×.
Every figure here is computed by venues.py from the registry dated 2026-08-27b on a $100 wager, taker fees, referral discounts included where a venue publishes one. The crossover solves the flat fee against the notional fee, so the wager size cancels and 23.1× holds at any size. Moon's opening fee costs 1.00% of the distance to liquidation at every leverage level, because leverage scales the numerator and the denominator alike. Neither curve carries cost of carry: Hyperliquid publishes its funding rate and Moon does not publish its rate for a wager held past 8 hours, which makes both lines floors rather than totals.
why the movers moved
- Strategy (MicroStrategy)▲ 6.59%8-K filed 24 Aug, cause unconfirmedunconfirmed
- Zcash▲ 2.06%Grayscale listed the first Zcash ETF on NYSE Arca, and the spot token repriced with it.confirmed
Only 18 of 50 tracked markets sit on more than one venue
Each square is one tracked market, shaded darker the more of the 5 venues list it and left pale where only one does. A pale square leaves you one quote with nothing to check it against, and that is 32 of 50 markets, 64% of the universe. Coverage never reaches all 5; the deepest anywhere is 2. Moon lists 50 of the 50 markets, the widest of the 5, and Avantis lists 0. Every one of the 50 squares is listed somewhere.
Coverage comes from venues.py against the registry dated 2026-08-27b: a market counts as listed when the venue covers its asset class and the venue's own instrument record carries a maximum leverage for it. Listing is not the same as a checked instrument on Moon's book, which is why no square here links to a venue, and 50 of the 50 tracked markets carry a dated human check against Moon's own book.
the ones that bite
Ranked by 30-session realised volatility, highest first. Volatility is not opportunity wearing a different hat. It is the reason the liquidation column reads the way it does on these particular names.
the ones that sit still
These take more multiplier before the stop lands inside an ordinary day's range. That is arithmetic, not permission. Quiet markets have loud days too, and none of them send a warning first.
the shares people come here to size
The equity and ETF side, ordered by 30-session realised volatility. Ranked on measured movement rather than on click history, because there is no click history worth ranking on yet and a made-up popularity score is still made up.
pick one and see what it does to a position
All 45 instruments are recorded as listed on Moon on the operator's attestation for the whole set, which is a weaker record than an instrument-by-instrument check and is dated so you can see its age.
Availability states across the universe: 45 verified. Only a state of verified opens a link, and a check expires after 30 days. How that check is done is set out in the methodology.
what is already on the calendar
Scheduled reports for names in this universe. A print outside its own consensus is where an overnight gap comes from, and a gap does not stop at a liquidation price on the way through it.
| Session | Market | When | Consensus |
|---|---|---|---|
| Wed 02 Sep | Broadcom Inc. AVGO | after the close | $2.83 expected, $1.26 a year ago |
| Tue 08 Sep | Oracle Corporation ORCL | time not confirmed | $1.40 expected, $1.20 a year ago |
Dates and consensus from the Nasdaq earnings calendar, read 2026-08-27 19:30 UTC. Companies move their own dates and a consensus is an average of estimates, not a forecast this site is making.
read these before you touch the slider
what this actually is
A simulator and a data reference. Not a trading venue. No money is staked here, no order is routed, no position exists anywhere. Prices are indicative third-party data from Twelve Data. Moon prices its own book, so its reference price, spread, fees, funding and liquidation rules will produce a different result from anything on this page.
What it is good for is arithmetic, which is the part nobody checks before sizing. Leverage gets sold as a multiplier on the outcome. It is a divider on how much you can absorb. At 100x a 1.00% move against you closes the position, and most markets here cover more ground than that on an ordinary Tuesday, which is why every market page counts the sessions it actually happened on over that market's own last 30 rather than reasoning from a market-wide average.
what moon charges, and what it does not tell you
These are the modelling assumptions this site runs on, not confirmed venue terms. The simulator uses them because it needs a number, not because anyone has checked them against Moon's book.
The opening fee is charged on the wager and not on the exposure, so it reads 1% at every leverage level on the leverage range. At 1000× the max loss price sits 0.10% from entry, and the fee eats 1% of that distance, the same share it eats at 2×. So climbing the leverage range gains you nothing on cost. What it takes away is room: 1000× leaves you 0.10% before you are closed, where 2× leaves 50.00%.
Held past 8 hours a wager also pays a rolling fee. Moon describes the rate as dynamic and does not publish it, so no figure on this site models it and anything carried overnight costs more than shown.
MarketMoves earns a commission when a reader opens an account through a link on this page. It costs the reader nothing and does not change the figures shown. Nothing on this page is an order, a balance or a position. See the affiliate disclosure and the methodology.