Funding rates are small periodic payments exchanged between traders holding long and short perpetual futures positions. Individually tiny, they repeat every few hours and compound into a real cost — or income. A bot that holds positions through funding periods without counting them can win on price and still lose money overall.
What is a funding rate in one sentence?
A funding rate is a small payment made every few hours between the two sides of a perpetual futures market — usually longs paying shorts when the rate is positive, shorts paying longs when it's negative. A perpetual future (perp) is a contract for betting on a price without owning the coin, and unlike a normal futures contract it never expires. With no expiry date to pull its price back to reality, exchanges use funding as the leash: when the perp trades above the real coin's price, longs pay shorts, nudging traders toward the side that corrects the gap. The rate is typically quoted per 8-hour period, and you only pay or receive it if you're holding a position at the moment funding is exchanged.
What does funding actually cost in numbers?
Small per period, meaningful per month — here's the arithmetic on a $10,000 position. Say you hold a $10,000 long perp position and the funding rate is 0.01% per 8-hour period (a common, unremarkable level). Each period costs $10,000 × 0.0001 = $1. Three periods a day is $3; thirty days is about $90, or 0.9% of the position per month — roughly 11% a year if it persisted. Note what the fee is charged on: the full $10,000 position, not your margin (the money backing the trade). If you posted $1,000 of margin at 10x leverage, that $90 monthly funding bill is 9% of your actual capital. In heated markets funding can run many times higher than 0.01%, and it can also flip negative — in which case holding the position pays you.
How can a strategy win on price and still lose on funding?
Because price profit and funding cost accrue on different clocks — one depends on being right, the other only on holding. Take a bot that buys the $10,000 perp and exits 30 days later 0.5% higher: $50 of price profit. If it paid 0.01% funding per 8 hours the whole time, it also paid about $90 in funding. Net result: a $40 loss on a correct call. The trap is invisible in any system that simulates prices but ignores carrying costs — there, this trade looks like a winner. Slow strategies that hold for days or weeks are the most exposed; a bot that's in and out within a single 8-hour window may never pay funding at all. Duration, not direction, decides how much funding matters.
What should a bot do about funding rates?
Count them as a cost of holding — in testing, in live trading, and ideally in the decision to hold at all. Concretely: any simulation of a perp strategy should subtract funding period by period, the same way it subtracts trading fees; a strategy's expected edge per trade should be compared against the funding it expects to pay over its typical holding time; and position reports should show profit net of funding, not just price movement. Some strategies invert the problem and funding deliberately — though that income competes against price risk and can vanish when the rate flips. Liquid Edge — non-custodial automated algorithmic crypto trading and portfolio management — is built for exactly this kind of accounting: strategies defined as complete rule sets in , evaluated with their real costs attached, running on funds that stay in your own custody. The honest limit: future funding rates are unknowable, so even careful accounting is an estimate — one more reason to prefer strategies that survive a range of funding assumptions, not just the friendly one.
