HYPE research

Does High HYPE Funding Predict a Reversal?

A rolling 30-day event study using official hourly funding and completed HYPE candle closes. Inspect the sample before turning a popular market narrative into a trading rule.

Complete sample30 rolling daysSep 10, 2026 to Oct 10, 2026 UTCLatest funding +0.125 bps

Does high HYPE funding predict a reversal?

Not by itself. The all-hours 24-hour median HYPE return was -0.04% in this rolling 30-day window. Across 15 elevated funding episode starts, the 24-hour median was +0.12%, 0.15 percentage points above the baseline. Across 41 negative episode starts, it was -0.31%, 0.27 percentage points below the baseline. This descriptive sample does not support treating either funding extreme as an automatic reversal or continuation signal.

Evidence for this answer
Hyperliquid Docs: FundingHyperliquid Docs: Info endpoint

Source window

What is actually in the sample

Fetched Oct 10, 10:46 UTC
Funding rows
720
Official hourly records
Aligned closes
720
Completed 1h references
Elevated episodes
15
First hour above 0.125 bps
Negative episodes
41
First hour below 0 bps

Episode-level comparison

Forward HYPE returns after funding changes regime

Primary episode rows use the first hour after a threshold crossing. Each result is compared with all eligible hours in the same window.

All hourly observations

Every aligned hourly funding observation in the window.

Same-window hourly baseline

720 events
1h forward
+0.06% median54% positive · n=720Same-window baseline
4h forward
+0.04% median51% positive · n=717Same-window baseline
24h forward
-0.04% median49% positive · n=697Same-window baseline

Elevated funding episodes

First hour after funding moves above the 0.125 bps hourly base interest rate.

Descriptive sample, no significance test

15 events
1h forward
+0.12% median53% positive · n=15vs baseline +0.06 pp median · -0.42 pp positive
4h forward
+0.25% median53% positive · n=15vs baseline +0.21 pp median · +1.87 pp positive
24h forward
+0.12% median53% positive · n=15vs baseline +0.15 pp median · +4.12 pp positive

Negative funding episodes

First hour after funding moves below 0 bps, when shorts pay longs.

Descriptive sample, no significance test

41 events
1h forward
+0.23% median59% positive · n=41vs baseline +0.17 pp median · +4.79 pp positive
4h forward
+0.01% median51% positive · n=41vs baseline -0.04 pp median · -0.24 pp positive
24h forward
-0.31% median48% positive · n=40vs baseline -0.27 pp median · -1.71 pp positive

Share the finding

Text and image include the baseline, both episode samples, source time, and limits. No affiliate link or trading signal.

The rolling values are frozen at the displayed source time.

Inspect the observations

Recent funding episode starts

The latest six starts from each regime are shown directly. Newer rows can have pending forward windows.

Show 12 recent episode starts
negative funding
Oct 9, 13:00
-0.217 bps
1h-0.70%4h-0.54%24hPending
negative funding
Oct 7, 17:00
-0.083 bps
1h-0.54%4h+0.30%24h-4.72%
negative funding
Oct 7, 02:00
-0.010 bps
1h+0.47%4h+0.82%24h-2.52%
negative funding
Oct 6, 21:00
-0.113 bps
1h-0.66%4h-0.19%24h-4.45%
negative funding
Oct 6, 19:00
0.000 bps
1h+0.02%4h-0.21%24h-4.44%
negative funding
Oct 6, 17:00
-0.138 bps
1h+0.14%4h+0.55%24h-4.22%
elevated funding
Oct 5, 09:00
+0.215 bps
1h-0.68%4h-1.24%24h-0.09%
elevated funding
Sep 30, 22:00
+0.427 bps
1h+0.21%4h-1.81%24h-5.21%
elevated funding
Sep 23, 18:00
+0.161 bps
1h+0.37%4h+0.28%24h-0.04%
elevated funding
Sep 22, 11:00
+0.192 bps
1h+0.17%4h-0.87%24h+0.73%
elevated funding
Sep 21, 16:00
+0.240 bps
1h+0.23%4h+0.25%24h+1.91%
elevated funding
Sep 21, 09:00
+0.480 bps
1h-0.35%4h-0.02%24h-0.26%

Interpretation

What this study can tell you

It describes what HYPE did after comparable funding conditions in one recent rolling window and how those returns differed from the same-window hourly baseline. It can challenge a simplistic story, expose a small sample, and give funding context before a trade. It cannot establish that funding caused the move or that the next episode will behave the same way.

  • Episode windows can overlap and are not statistically independent.
  • All-hours differences are descriptive percentage-point gaps, not statistical-significance estimates.
  • The rolling sample can be dominated by one market regime and is not a causal test or trading backtest.
  • Returns exclude funding cash flow, fees, spread, slippage, liquidation, and any hedge on another venue.

Download the event rows

The CSV includes every funding timestamp, regime label, episode flag, reference close, and available 1h, 4h, and 24h return. Empty cells mean the forward candle had not completed or could not be aligned.

Download CSV

Reproducible method

How the comparison is built

  1. 01

    Collect

    Request two bounded 15-day funding windows and one 30-day set of 1-hour candles from Hyperliquid.

  2. 02

    Align

    Match each funding timestamp to the hourly candle close completed at that same UTC boundary.

  3. 03

    Classify

    Label funding above 0.125 bps as elevated, below 0 bps as negative, and collapse consecutive hours into episodes.

  4. 04

    Compare

    Calculate 1h, 4h, and 24h returns, then compare episode medians and positive shares with all eligible hours.

Put funding beside the rest of the HYPE trade

Live HYPE fundingCurrent payer, completed carry, and cost planner.OpenHYPE market briefPrice, OI, execution, flow, and risk context.OpenLiquidation mapActual monitored levels with cohort coverage.OpenMarket history CSVGrowing hourly HYPE price, OI, volume, and funding.Open
Risk notice
Crypto perpetuals and leveraged trading are high risk. You can lose money through liquidation, funding, slippage, oracle issues, protocol failures, and market volatility.

Affiliate disclosure: we may earn referral rewards if you use this link.

Finished checking the funding evidence?

If you have already decided to trade and are eligible, review the referral discount and its limits before opening Hyperliquid. This study is independent of that decision.

Open Hyperliquid
How to read this dataReviewed 2026-08-25 · Expand
Where this comes from
Hyperliquid public info endpoint
How fresh it is
The rolling study refreshes through a 15-minute server cache. Its source window ends at the latest completed UTC hour.
Last reviewed
2026-08-25
How to use it
This is descriptive market research, not a signal, forecast, causal test, recommendation, or complete trade return.
Sources and correctionsData dictionaryRisk disclaimer

Sources

4 references · Expand
  • Hyperliquid Docs: FundingAccessed 2026-09-04
    Supports: Hourly funding, funding formula, interest-rate component, premium component, funding payment formula, and the direction of payments when funding is positive or negative.
  • Hyperliquid Docs: Info endpointAccessed 2026-09-04
    Supports: Public info endpoint families used for market, book, candle, account, portfolio, native borrow/lend reserve, vaultDetails, userFees account-rate, and maxBuilderFee approval reads.
  • Hyperliquid Docs: Perpetuals info endpointAccessed 2026-09-04
    Supports: Core and HIP-3 perpetual DEX discovery, current market contexts including rolling dayNtlVlm, perpetual metadata, funding history, predicted funding, DEX and market limits, DEX status, configuration metadata, clearinghouse state, and open-interest cap fields.
  • Hyperliquid Docs: Rate limits and user limitsAccessed 2026-09-04
    Supports: Current REST request-weight, WebSocket, and user-limit constraints for public-data tooling.