What it costs to open, close and hold a perpetual futures position, coin by coin, on two anonymous kinds of venue. Everything is in basis points (1 bp = 0.01%).
Data window: funding for the 180 days to 31 Aug 2026; order-book costs measured 6–7 Oct 2026.
Information only. This page describes historical trading costs. It is not investment, legal or tax advice, not a recommendation to buy, sell or hold any asset, and not a recommendation to use any venue or product. Nothing here is personalised. Perpetual futures can lose more than people expect; decisions are yours.
On-chain perpetuals exchange. Funding is paid every hour. Base taker fee 4.5 bp per side.
Centralised exchange. Funding is paid every 8 hours. Base taker fee 6.0 bp per side. Funding figures come from a large peer exchange of the same type.
Click a coin to see its costs, its funding range and a calculator for your own case. Getting in and out: the round-trip execution cost for a $2,000 market order. Holding: funding over 30 days for the side you pick, Long or Short, as the typical value with the 10th to 90th percentile range of the 30-day windows of the last 180 days.
| Round trip, bp |
|---|
Keys: / search, ↑ ↓ move between coins, Esc close the panel. The star keeps a coin in your own list on this device.
More than 50 trading ideas have been backtested on historical data (to 11 October 2026). 40 of them did not meet the methodological thresholds. The rest are still under observation: none has met all thresholds yet, and they are not disclosed. Two of the 40 are shown in detail below. A full list of the 40, with data, results and re-check dates, is planned for the paid version.
One idea is one hypothesis in plain words, however many coins, horizons and settings were tried for it. The figures are historical results from this project’s own data and cost measurements. They are not a recommendation for or against any strategy.
Funding rates differ from one venue to another. One idea is to hold opposite positions on two venues and collect the gap. It was backtested on 209 coins from 2024 to the end of August 2026, rebalancing weekly and charging the measured trading costs.
Percent per year on the money in both positions together. 2026 covers 1 January to 31 August. Scale 0–5%.
| Year | Days | bp per day per position | Percent per year |
|---|---|---|---|
| 2024 | 337 | 2.18 | 4.0 |
| 2025 | 365 | 1.15 | 2.1 |
| 2026 (to 31 Aug) | 243 | 0.43 | 0.8 |
Data to 31 August 2026: settled funding on both venues, prices and measured costs. Annual figure = bp per day × 365 ÷ 2, because the money is split across two positions. It may be re-run on newer data.
The order book lists the buy and sell orders waiting at each price. One idea is that when far more buy orders than sell orders are waiting near the price, the price tends to move up next, and the other way round. It was backtested on 8 coins from January 2023 to September 2026, using 5-minute data from a large exchange’s public archive. There were 48 variants: four signals, three holding times (15, 30 and 60 minutes), two thresholds, and trading with or against the signal. Entry was five minutes after each signal, with trading costs charged.
Basis points per side (1 bp = 0.01%). Scale 0–7 bp. Fees only: spread and slippage, as on the map above, come on top.
| Item | bp per side |
|---|---|
| Best signal can pay (3.3 bp per round trip) | 1.6 |
| Maker fee, hypothetical | 3.0 |
| Venue A taker fee | 4.5 |
| Venue B taker fee | 6.0 |
Data: 8 coins, January 2023 to September 2026. The rules were written down before the data was examined. It may be re-run on newer data.
For spreadsheets and your own backtests. These are the project’s own measurements and calculations, with venues anonymous; they are not raw exchange data. Costs are per side in your model: use half the round trip for each open and each close.
The free page is what you see here. The paid column is a plan, not a product: none of it exists yet, and it will be built only if enough people say they would use it.
| What you get | Free (this page) | Paid (planned) |
|---|---|---|
| Venue names | Anonymous: Venue A and Venue B | The real names of the venues, shown on the site only (display, not an API feed), if the venues’ terms allow it (not yet checked) |
| How fresh | Delayed: funding to 31 Aug 2026, book costs from 6–7 Oct 2026 | Current: funding and costs refreshed through the day |
| Coins | 164 coins that trade on both venues | Same, plus new listings as they appear |
| Order size | $2,000 only | $500, $2,000 and $10,000, so you can see how cost grows with size |
| Cost to open and close | One snapshot period | History by day and by hour, and measured cost in fast markets |
| Cost to hold (funding) | Typical 30-day figure and range | Funding right now compared with that coin’s own history, and the next payment time |
| Alerts | None | Message when funding or cost for a coin you follow moves outside its usual range |
| Market events | None | Feed of new listings, delistings and status changes; summary of forced liquidations by coin |
| More venues | Two | Funding on more large venues; cost comparison on more venues later |
| Ideas backtested | Two worked examples of the 40 set aside: the funding gap between venues and reading the order book (above) | The full list of the 40 ideas set aside (it grows), with data, results and re-check dates. Candidates still under observation are not disclosed. |
| Your own tools | Copy as CSV or JSON | Download of the full history and an API for bots and spreadsheets: only the project’s own composite indicators, built from the data it collects. What the venues themselves publish is never streamed or resold, and venue-specific figures stay anonymous in downloads and the API. |
Planned items depend on what the data sources allow. Raw exchange data is not streamed or resold: every figure is the project’s own measurement or calculation, and an API would carry only its own indicators. Liquidation totals would be a lower estimate, because the source reports only part of the forced orders.
Built from independent order-book sampling and settled funding history. Measurements are historical and may not hold today. Venue names are intentionally left out. This is information about trading costs, not investment advice, and not a recommendation to trade any coin or use any venue.