What are perpetual futures?
A perpetual future is a contract that lets you speculate on an asset's price without ever owning it and without a settlement date. A traditional dated future settles on a fixed day; a perp just keeps running. That single design change is the whole point.
Instead of an expiry to force the contract price back to spot, perps use a periodic cash payment called funding. When the perp trades above the underlying spot price, longs pay shorts. When it trades below, shorts pay longs. This nudge keeps the two prices close without anyone ever taking delivery.
The instrument was described academically by economist Robert Shiller in 1992 and made practical by BitMEX in 2016. Today perps dominate crypto volume, usually trading many times the size of the spot market for the same coin.
How do perps differ from spot and dated futures?
Spot means you buy the actual coin and hold it. A dated future is a contract with a fixed expiry that settles to spot on that day. A perp sits between them: a contract like a future, but open-ended like spot.
| Feature | Spot | Dated future | Perpetual |
|---|---|---|---|
| Expiry | None | Fixed date | None |
| Leverage | Usually none | Yes | Yes |
| Easy to short | No | Yes | Yes |
| Tether to spot | Is spot | Convergence at expiry | Funding payments |
| Ongoing cost | None | Basis | Funding + fees |
For a bot, the practical differences are leverage and symmetry. On a perp you can go short as easily as long, so a strategy can trade both directions of a market with identical code. That symmetry is why so many automated strategies are built on perps rather than spot.
What is mark price, and why does your bot care?
Every perp venue tracks three prices, and confusing them is a common way to get burned. The last price is whatever the most recent trade printed on that exchange's order book. The index price is an average of spot prices across several major exchanges. The mark price is a smoothed, manipulation-resistant reference, usually the index price plus a decaying moving average of the perp's basis.
Your unrealized profit and, critically, your liquidation are calculated from the mark price, not the last price. This matters because a single exchange's order book can spike or wick far away from the real market for a few seconds. If liquidations used last price, a thin book could be pushed to trigger a cascade of forced closes. Mark price stops most of that.
For bot logic, always read mark price for risk checks and last price for execution. A bot that sizes stops off last price will fire on noise that mark price ignores.
How does funding keep a perp tethered to spot?
Funding is the mechanism that replaces expiry. At each funding interval, one side pays the other a percentage of position notional based on how far the perp has drifted from the index. It is a payment between traders, not a fee the exchange keeps.
- Positive funding: perp trades above spot, longs pay shorts. This is the normal state in a bull market.
- Negative funding: perp trades below spot, shorts pay longs. Common in sharp sell-offs.
- Interval: commonly every 8 hours on venues like Binance and Bybit, though some venues and volatile conditions use 4-hour or 1-hour cycles. Hyperliquid, for example, funds hourly.
Rates are small per interval, often a few basis points, but they compound. A persistent 0.01% every 8 hours is roughly 11% a year paid by longs, which quietly eats a leveraged position. Some bots trade funding directly by holding a delta-neutral spot-versus-perp position to harvest it. We cover the mechanics in Funding Rates Explained, and it is worth reading before you assume funding is negligible.
How does liquidation work on a perp?
Liquidation is what happens when your margin can no longer cover your losses. Each position has a maintenance margin, the minimum equity the exchange requires. When mark price moves against you far enough that equity drops to that floor, the position is force-closed at the liquidation price and you lose the margin backing it.
The higher your leverage, the tighter the buffer. Rough math on an isolated position, before fees and the maintenance margin cushion:
- 10x leverage liquidates on roughly a 10% adverse move.
- 20x liquidates on roughly a 5% move.
- 50x liquidates on roughly a 2% move, which crypto can print in minutes.
Isolated margin caps the loss to that one position's collateral. Cross margin shares your whole balance, which lowers liquidation risk on any single trade but can drain the entire account if a position blows up. Most systematic bots run isolated margin per position so one bad fill cannot cascade. How much leverage is sane is its own topic, covered in Leverage and Trading Bots and Position Sizing for Bots.
Why are perps the default instrument for crypto trading bots?
Several properties line up in the bot's favor:
- Liquidity. The BTC and ETH perps are among the deepest markets in crypto, so slippage on reasonable size is small.
- Symmetry. Long and short are the same operation with a sign flip, so one codebase trades both sides of a trend.
- Capital efficiency. Leverage means a strategy can run more notional per dollar of collateral, though that cuts both ways.
- One API surface. Placing, sizing, and closing orders happens through a single perp endpoint rather than juggling spot inventory and a separate borrow for shorts. See Trading Bot APIs for how that connection works.
- Funding as an edge. The funding stream itself is a tradable signal, not just a cost.
This holds on both centralized venues and on-chain perp DEXs. Copy-trading systems increasingly run on perp DEXs precisely because positions are transparent on-chain; see Copy Trading on Perp DEXs and the deeper look at Hyperliquid Trading Bots. The trade-offs between the two worlds are laid out in On-Chain vs CEX Trading Bots.
What can go wrong with a bot on perps?
The same features that make perps attractive make them unforgiving. Be honest with yourself about the costs before you go live.
- Fees compound with leverage. Taker fees are often around 4 to 6 bps per side on major venues, so a round trip is roughly 8 to 12 bps of notional. At 10x leverage that is 80 to 120 bps of your margin per trade. A bot that trades often can bleed to death on fees alone.
- Funding drift. A position held through many funding intervals in an expensive market can lose more to funding than to price.
- Liquidation cascades. In fast markets, forced liquidations feed each other and mark price can move faster than your bot reacts.
- Leverage magnifies mistakes. An overfit or mis-sized strategy that would slowly lose on spot can be wiped out in one candle on 20x.
None of this is a reason to avoid perps. It is a reason to model costs realistically. Most retail bots lose money after fees and funding, and leverage is usually why the losing ones die fast rather than slow. Trade small, size positions by risk rather than by leverage, and assume your real edge is thinner than your backtest suggests.
Frequently asked questions
What are perpetual futures in simple terms?
They are contracts that let you bet on a crypto's price, up or down, with leverage and no expiry date. Instead of settling on a fixed day, they use recurring funding payments between traders to keep the contract price close to the real spot price. You can hold the position as long as your margin covers it.
How is the perp price kept close to spot?
Through funding. Every few hours, if the perp trades above spot, long holders pay short holders; if it trades below, shorts pay longs. This creates a financial incentive to push the perp back toward the index price of the underlying asset, so the two rarely drift far apart for long.
What triggers a liquidation on a perp?
Liquidation happens when the mark price moves against your position far enough that your equity falls to the maintenance margin, the minimum the exchange requires. The exchange then force-closes the position at the liquidation price and you lose the collateral backing it. Higher leverage means a smaller adverse move triggers it.
Why do trading bots prefer perps over spot?
Perps are deeply liquid, let a bot short as easily as it goes long, and run through a single API instead of managing spot inventory plus a separate borrow. Leverage adds capital efficiency, and funding rates give bots an extra signal to trade. The catch is that fees and leverage also amplify losses.