By Keel Research Team · Updated May 12, 2026
A Bollinger squeeze flags volatility compression on a chart. Combined with above-average relative volume, it surfaces Hyperliquid perpetuals coiling for a directional move. The Squeeze Setup screen ranks HL pairs by squeeze tightness and volume confirmation, updated hourly.
The Squeeze Setup screen looks for Hyperliquid perpetuals where price has compressed into a narrow Bollinger Band range while volume has started rising. X-axis: vol-squeeze percentile — a measure of current Bollinger Band width relative to its recent average, where the bottom quartile (≤25th) is the tightest compression. Y-axis: relative volume percentile — top half (50th+) ensures the squeeze is being scouted by real flow, not dead tape. Z-axis: ROC at the 50th+ percentile, a directional bias so we’re looking at squeezes resolving up rather than down. The combination filters out the most common squeeze failure mode — quiet ranges that stay quiet — by requiring volume participation. Squeezes on liquid HL perps tend to resolve within 1-3 days on the 4h timeframe; the screen refreshes hourly and the Constructor lets you switch timeframes live.
The screen below is pre-loaded with the Squeeze Setup preset. Adjust signals, thresholds, and timeframe inline — your changes update the cohort in real time. Share or backtest the resulting state directly from the toolbar.
Keel is a Strategy OS for AI-assisted systematic trading on Hyperliquid. Backtest a strategy that uses the Squeeze Setup signals as entry filters, optimize parameters across thousands of variants, then deploy live with funding-aware execution and full risk controls.
Free to start — connect a Hyperliquid wallet when you’re ready to go live.
A Bollinger squeeze is a chart pattern where the upper and lower Bollinger Bands contract to an unusually narrow range — volatility has compressed. Traders watch for it because compressed volatility historically expands in one direction, providing a breakout opportunity.
A squeeze without volume is just a range. Volume rising into a squeeze means real participants are positioning — when the breakout happens it has flow behind it. We filter for top-half relative volume to avoid signaling pairs whose tape has gone dead.
The default screen uses 4h, which catches squeezes that resolve within 1-3 days. 1h is more sensitive (more signals, more false breakouts); 1D is slower and pairs better with a multi-day swing approach. Switch the timeframe live in the Constructor.
Market data refreshes hourly. The screen recomputes percentile rankings, thresholds, and qualifying cohorts on each refresh. Funding rates update on the same hourly cadence as Hyperliquid’s native funding settlement.
Two common failures: false breakouts (the move breaks one direction, fakes, and reverses) and dud squeezes (volatility stays compressed indefinitely). The relative-volume gate addresses dud squeezes; managing the false-breakout case requires entry rules (e.g., wait for a candle close beyond the band) rather than the screener itself.
Yes. Open the screen, click "Backtest in Keel," and the current state — signals, thresholds, universe, timeframe — passes into a Keel workspace. From there, you can run a full backtest with realistic fees, slippage, and funding modeled, then optimize and deploy live.
The Hyperliquid momentum screener ranks perpetuals by rate-of-change, MACD slope, and relative volume across the top-100 HL universe. Nam…
The funding rate screener finds Hyperliquid pairs paying the most carry — high funding combined with low realized volatility. Short the p…
The oversold screener finds Hyperliquid perpetuals trading deep in the bottom of their recent range with MACD slope starting to turn up. …