Pressure Break is the more direct of BookPilot's two signals. When one side of the order book is much heavier than the other and stays that way, BookPilot trades in the direction of the weight. The useful detail is in how "heavier" and "stays that way" are defined.
Step 1: add up the book
The scanner sums the resting size on the first Pressure Break Depth levels of each side. The default is 2, so it adds the best bid to the next bid down, and the best ask to the next ask up. A depth of 1 looks only at the touch.
Going deeper makes the measure steadier but slower, and more of the size it counts can be pulled before price gets there. Staying shallow makes it fast and sensitive, and noisier.
Step 2: start a candidate
When the bid total is at least Pressure Break Ratio times the ask total, a long candidate starts. When the ask side is the heavy one, a short candidate starts. The default ratio is 4.0.
A simple example: the two bid levels hold 130 and 95 contracts, a total of 225. The two ask levels hold 30 and 25, a total of 55. The ratio is about 4.1, so a long candidate begins. Below a ratio of roughly 2.5, the guide warns, the scanner fires on ordinary book noise.
Step 3: make it hold
A candidate is not a signal yet. The ratio has to stay at or above the Pressure Break Hold Ratio (2.5) for the Confirm time (250 ms). If it slips under 2.5 inside that window, the candidate is dropped. That short hold filters out size that flashes into the book and disappears.
After it fires, the scanner will not fire again until the ratio falls back below the hold ratio, so one long-lasting imbalance produces one signal rather than a stream of them. Each scanner also rests for 5 seconds after firing.
Step 4: the filters
A fired signal still has to pass several checks before an order goes out:
- Spoof veto. While the book's spoof score is 0.6 or higher, Pressure Break is vetoed, because the pressure is probably fake.
- Conflict check. If Pulled Wall fires the other way at the same moment, there is no trade.
- Forecast Confirm. With it on, the micro-forecast (a logistic model of the next 2-tick move) has to give the signal's direction at least the Confirm Threshold of 0.55.
- Rules of Engagement. The strategy must be armed and flat, inside its trading hours, allowed by the With Bias, Long only or Short only mode, and under the risk preset's hourly trade cap.
Step 5: the order
BookPilot places a limit order 2 ticks behind the best bid for a long, or behind the best ask for a short. It is cancelled if it has not filled within 3 seconds, or if price runs 4 ticks away in the trade's direction. Once filled, the bracket defaults to a $150 target and a $100 stop per contract: 30 and 20 ticks on NQ, 12 and 8 ticks on ES.
| Setting | Default | Suggested range |
|---|---|---|
| Pressure Break Depth | 2 | 1 - 5 |
| Pressure Break Ratio | 4.0 | 3.0 - 6.0 |
| Pressure Break Hold Ratio | 2.5 | 2.0 - 3.5 |
| Confirm (ms) | 250 | 100 - 1000 |
| Pressure Break Limit Offset (ticks) | 2 | 0 - 3 |
| Pressure Break Cancel (ticks) | 4 | 2 - 6 |
The Invert switch
Invert flips the direction, so a heavy bid becomes a short signal. It is off by default. The guide recommends turning it on only if your own testing shows that the pressure is being faded on the market you trade, not as a quick fix after a bad session.
Tuning without guessing
The guide's starting profiles give a sense of scale. Conservative uses a ratio of 5.0 with a 500 ms confirm. Active uses 3.0 with 150 ms. Balanced is the shipped default. Change one value at a time and replay the same Market Replay days after each change, because the Strategy Analyzer has no book history to test against.
The panel's Blocked line tells you which filter stopped each signal. If most blocks come from the forecast, the filter is doing a lot of the work. If many trades lose straight away, tighten the scanner with a higher ratio or a longer confirm. The full settings sheet is in the BookPilot user guide, and the BookPilot product page lists what the plan includes.
Book pressure can vanish in an instant, and no filter catches every fake. Treat Pressure Break as a rule to test in Market Replay and SIM first, well before any evaluation, funded or live account.