# Liquidation heatmap (estimated)

Source: https://tactiq.pro/docs/liquidation-heatmap · Reading the market · updated 2026-10-31

The liquidation heatmap is an estimate from a model of where leveraged positions may be liquidated. It is labelled as an estimate everywhere it appears.

The liquidation heatmap estimates the prices where leveraged perpetual positions would be force-closed. It answers: *where would a move trigger forced selling or buying?* Turn it on with **Indicators → Positioning (estimated)** or press `j`. It is available on 5-minute to daily candles.

> This layer is a **model**, not data. Nobody publishes where positions sit or at what leverage. Tactiq labels every number from it as *estimated*.

![A 1-hour BTC chart from 26 October to 1 November covered by the estimated liquidation heatmap: bright yellow bands above price near 83,000 and 85,500 and below near 75,000 and 74,000, a dark purple channel hugging the candles, and the legend at the bottom saying estimated.](https://tactiq.pro/docs/chart/liquidation-heatmap.webp)

*BTC perp, 1h, 26 Oct–1 Nov 2026: estimated liquidation levels; bright = more positions would be forced out there.*

- **What you see:** Bands of colour above and below the candles: bright yellow is where the model puts many liquidations. The dark channel around the candles has few: levels there were used up (price traded through them) or sit closer than even 100x leverage allows. In the picture, the bright band near 79,000 that sat under price for days disappeared in the fall of 29 to 30 October: those long positions were (in the model) liquidated. The legend at the bottom says *estimated* and how well the model checks out so far.
- **How to read it:**
  - **Bright bands below price** are estimated long liquidations (forced selling). **Above price**, estimated short liquidations (forced buying).
  - **A dense band near price** is where a move can feed on itself.
  - **A band that vanishes** after price crosses it has been used up.
- **What it tells you:** Where a move might trigger forced orders, if the model's guesses about leverage are right. It does **not** show real positions, and price is not "pulled" to a band.
- **How to use it:** Use it for risk, not entries. A bright band just past your stop means a move there could speed up and slip your stop. A bright band ahead of a trade in your favour can be a sensible place to take profit. If price runs into a band, check the big-trade squares (real liquidations) to see whether forced flow is really happening.
- **Common mistakes:**
  - Treating the bands as real orders. They are a model built from open interest.
  - Expecting price to "hunt" every bright band. Tactiq has not measured that.
  - Ignoring the check line in the legend: early on there is too little data to test the model.

### How the model works

1. **New positions.** When a bar's open interest rises, the model assumes new positions, opened at the bar's typical price. They are split between longs and shorts by the bar's taker-buy share.
2. **Leverage mix.** Each opening is spread over 10x, 25x, 50x and 100x (30 %, 30 %, 25 % and 15 %). A long at leverage L is liquidated near price × (1 − 1/L + maintenance margin). A short is liquidated near price × (1 + 1/L − maintenance margin).
3. **Closing.** When open interest falls, every open estimate shrinks in proportion.
4. **Used up.** A level that price trades through is gone: those positions were liquidated.

Tactiq checks the model against the real forced-order stream. That stream is a sample (Binance sends at most one liquidation per second), so the check is a sanity test, not a proof.