Breakout and retest: confirmation costs something

Research status: untested hypothesis. No backtest, paper or live results exist. Every threshold, period, R value and risk fraction is an untested candidate parameter, not a recommendation.

The idea

Chasing a breakout candle means buying at the most extended price with a distant stop. Waiting for a retest gives a tighter, clearer stop, but some breakouts never come back, and those missed trades count too. Confirmation isn’t free. This playbook defines the retest precisely so that the trade-off can be measured.

IF–THEN rules

At the breakout bar, H and D are the high and low of the previous 20 completed bars, and W = H − D (must be positive).

1. Breakout: a completed close above H + 0.05W, and that bar’s volume is more than 1.5× the average of the previous 20 bars. Don’t trade the breakout candle.
2. Watch window: the next 5 bars.
3. Retest: the first completed candle in the window whose low is within H ± 0.05W, that closes above H and closes above its open.
4. Cancel the watch if any watched candle closes below H − 0.05W, or the first candidate fails shared checks. Don’t keep searching until something works.
5. THEN enter at the next bar’s first executable price.

• Stop: retest candle low − max(2 ticks, 0.02W). Frozen.
• Target: H + W (the old range’s height projected upward). A projection, not a prediction. Net reward must be at least 2R after costs.
• Failure exit: a completed close below H − 0.05W exits at the next open (stop has priority).
• Time stop: 12 bars.
• One watch per UTC day. No concurrent watches, positions or pending orders.

When NOT to trade

• No retest within 5 bars: that’s a missed trade, not permission to loosen the definition or chase.
• Spread or volume fails the shared liquidity checks, or the next open removes the room to 2R.
• Scheduled event window or data feed issue: see the shared no-trade filters.
• A mirrored “breakdown short” isn’t covered or authorized here.

Hypothetical example (made-up prices)

Item	Value
20-bar range	H 1,000, D 960, W 40
Breakout close	1,006 (> 1,002) on 1.8× average volume ✓
Retest (bar 3)	low 999, close 1,004, close > open ✓
Entry (next open)	1,004.50
Stop: 999 − 0.80	998.20
Risk per unit incl. 0.60 costs	6.90
Target H + W	1,040 → net reward 34.90, about 5.1R

If it works: price reaches 1,040 within 12 bars: about +5.1R. A large R multiple is a feature of the geometry, not a sign the trade is likely. If it fails: price closes at 997, below H − 0.05W (998), but the stop at 998.20 already triggered intrabar and filled at 998.00: about −1.03R. Missed trade: price never retests and runs to 1,040 without you. That’s a recorded miss, not a mistake.

Ratigan Capital rules in this playbook

• Early Is Wrong: buying the breakout candle is a separate, untested strategy.
• Patience as a Position: a missed breakout is the cost of the rule.
• Never Average Down: a failed retest is an exit, not a cheaper entry.
• Price Is the Only Truth: volume and the old ceiling describe price behavior; the stop overrides them.

Why it can fail

High-volume breakouts fail too, and a retest can be a pause before a deeper fall back into the range. One unusual venue print can dominate the volume filter. Later entries can raise confidence while worsening price. The test must track missed continuations and false retests, and model realistic fills.

Pre-trade checklist

1. Breakout close and volume filter both true?
2. Retest within 5 bars, first candidate, not cancelled?
3. Shared filters clear; one watch per day?
4. Stop, target, time stop and size written; net ≥ 2R?

Sources

• StockCharts ChartSchool: Support and resistance
• StockCharts ChartSchool: On Balance Volume (volume concepts)
• Sullivan, Timmermann & White (1999), Journal of Finance (RePEc listing)