Why 0.01% on Hyperliquid isn't the same as 0.01% on Bybit

Perpetual futures keep their price close to spot via the funding rate — a periodic payment between the two sides of the contract. The exchange sets the rate, but an equally important parameter is cadence: how often that payment happens. Most traders ignore cadence when comparing rates and end up catching either phantom profit or real loss.

The basic math

A funding rate is a percentage of position notional transferred from one side to the other at each funding event. The interval between events is the cadence, and it varies sharply across exchanges:

This means: a rate of 0.01% on Hyperliquid and a rate of 0.01% on Bybit are not the same magnitude. The first pays out 24 times per day, the second 3 times. Over a day:

• Hyperliquid: 24 × 0.01% = 0.24% daily funding
• Bybit: 3 × 0.01% = 0.03% daily funding

Eight-fold difference, and it's entirely from cadence — the on-screen rates are identical.

Normalising to an 8-hour window

To compare cleanly, rates get reduced to a common window, usually 8 hours:

rate_8h = rate_per_cycle × (8 / cycle_hours)

Apply it:

• Hyperliquid 0.01% / 1h → 0.01% × 8 = 0.08% per 8h
• EdgeX 0.01% / 4h → 0.01% × 2 = 0.02% per 8h
• Bybit 0.01% / 8h → 0.01% × 1 = 0.01% per 8h

Now they're in the same coordinate system. And now you can see: the same 0.01% rate on Hyperliquid is 8× larger than on Bybit once normalised.

Where this creates the edge

Funding-rate arbitrage works like this: long on the exchange with the positive rate (you collect funding) + short on the exchange with a low/negative rate (you pay less or even get paid) simultaneously. Equal-sized positions = delta-neutral. Profit is the funding-rate spread.

Suppose:

• Hyperliquid BTC-perp: +0.005% / 1h → +0.04% per 8h
• Bybit BTC-perp: +0.01% / 8h → +0.01% per 8h

Divergence: 0.03% per 8h. Open long Hyperliquid + short Bybit at $10k per side. Over 7 days (21 × 8h cycles from Bybit's perspective, or 168 × 1h cycles from Hyperliquid's):

• Long Hyperliquid: 168 × 0.005% × $10k = +$84 funding income
• Short Bybit: 21 × 0.01% × $10k = −$21 funding paid
• Net funding: +$63 over 7 days

Minus taker fees on open + close for both sides (~$20 round-trip):

≈ +$43 net over 7 days on a $10k position ≈ ~22% APR.

At stable rates. If you'd ignored cadence and tried to compare "0.005% vs 0.01%" head-to-head, you'd have seen "negative divergence" and skipped the trade.

Where this creates the trap

Reverse scenario: a 0.05% rate on Hyperliquid looks "small" next to 0.05% on Bybit. In reality, it's 0.4% per 8h on Hyperliquid against 0.05% per 8h on Bybit — a 0.35% divergence, a fantastic setup. Without normalising, you don't see it.

The flip side: short trades over 1–2 Bybit cycles (8–16 hours) frequently don't recover round-trip fees. Taker 0.04% × 2 sides × 2 (open + close) = 0.16% round-trip. To cover that out of 0.01% / 8h funding, you'd hold at least 16 cycles (~5 days). On Hyperliquid the same distance is 16 hours.

Break-even by cadence

A simple formula for the minimum hold time just to recover fees:

hold_cycles = round_trip_fee% / divergence_per_cycle%

Apply it for a typical setup (taker ~0.04% × 2 sides = 0.16% round-trip, divergence 0.04% per 8h):

• On an 8h-cadence pair: 0.16% / 0.04% = 4 cycles = 32 hours
• On a 1h-cadence pair: 0.16% / (0.04%/8) = 32 cycles = 32 hours

Same number of hours — but 32 cycles on Hyperliquid means 32 funding payments accruing income, vs only 4 on Bybit. In practice:

• On 8h-cadence pairs the lag between "I just opened" and "I received my first funding credit" can be up to 8 hours
• On 1h-cadence — at most 1 hour

For exit discipline this matters: on Hyperliquid you see funding movement every hour and can react quickly to a reversal. On Bybit — only 3 data points per day.

Where the watchdog matters most

The main risk in funding arbitrage is rate reversal. The side that was paying you can stop and start charging. On 8h-cadence pairs this happens slowly (and there are notably fewer funding data points in the history), but on a 1h-cadence pair a flip can complete inside 4–6 hours.

That's why Finder's funding scanner carries a dedicated "Next flip" column estimating time-to-possible-reversal from the last 24h of rate history. On stable rates the flip is far; in volatile conditions the predictor flags risk before it materialises.

The takeaway

A funding rate without cadence is half the number. Comparing rates across Hyperliquid and Bybit is only meaningful after normalising to a common window. Two practical consequences:

1. Divergences that look invisible "head-to-head" turn out to be fat. 0.01% vs 0.005% looks like nothing; in 8h-normalised form it's 0.08% vs 0.01% — a workable setup.
2. Trades over 1–2 cycles on 8h exchanges almost always lose money. Round-trip fees of 0.12–0.16%, divergence of 0.04% per 8h — you need at least 4 cycles, i.e. 32 hours of hold.

If you're planning to trade funding-rate arbitrage, keep this formula handy. If you want Finder to do it for you — the funding channel is free during beta and normalises every rate automatically.