Candlestick Confirmation Strategy: Make Patterns Earn the Entry
A candlestick confirmation strategy treats the pattern as the last check, not the first. A hammer or a bullish engulfing bar appears somewhere on most charts most days. On its own it is noise. The same bar at a tested low, in the direction of the larger trend, after the candle has closed, is a different proposition. This guide lays out one testable version of that idea. Every setting is a starting point for research, not a recommendation. No backtested performance is claimed.
Our Candlestick Pattern Detection complete guide covers how the block recognises each formation. This post is about the rules that decide whether a detected pattern is worth acting on.
Why Does a Pattern Need Confirmation?
Pattern detection answers one question. Does this bar have the shape of a hammer, a shooting star or an engulfing bar? It says nothing about where the bar formed or what the market was doing before it.
That gap is where most candlestick systems fail. Investopedia’s hammer definition is explicit that the pattern needs a preceding decline to mean anything. A hammer in the middle of a range after a sideways week has the right shape and the wrong context.
The detection block in Arrow Algo outputs a true or false value for the chosen pattern on every candle. A confirmation strategy wraps that output in two or three extra conditions. The pattern becomes the trigger. The conditions become the filter.
The Three Gates
The version below is long-only and uses bullish patterns. Mirror every rule for a short version.
Gate 1: Location
The pattern must form near a recent low. Use the lowest low of the last 20 candles as the reference. Require the pattern candle’s low to be within 1.5 ATR of that reference.
A 20-candle low is a plain, observable level. It is the lower band of a Donchian channel. Our Donchian Channels guide covers the same level from the breakout side.
Gate 2: Trend
The close must be above the 200-period EMA. This keeps the strategy buying pullbacks inside an uptrend rather than catching falling knives in a downtrend.
This is the gate most traders want to remove. A hammer at the bottom of a long decline looks like the best trade on the chart. In testing, those are also the patterns most likely to be followed by another leg down. Keep the gate for the first round of tests. Remove it later and compare.
Gate 3: A Closed Candle
The pattern must be evaluated on a completed candle. A hammer with ten minutes left in the hour can become a bearish bar by the close. Detection on a live, unfinished candle is the most common source of signals that vanish. Our candle close confirmation guide covers why waiting for the close matters for every indicator, not just patterns.
In Arrow Algo this means feeding the detection block from the completed-candle outputs of the Data Watcher. The block then only sees bars that have closed.
Entries, Stops and Exits
Enter long when all three gates are true on the same candle and the pattern output is true. Allow one position at a time.
Place the stop below the pattern candle’s low. The low of a hammer is the point the market rejected. If price trades back through it, the pattern has failed and the reason for the trade is gone. Add a small buffer of 0.25 ATR so a single tick through the low does not stop the trade.
For the exit, two options are worth testing separately:
- Fixed reward: a target at two times the distance from entry to stop. Simple, and it gives a clean risk-reward reading per trade.
- Trailing stop: an ATR-based trail that starts at the pattern low and ratchets up. Our ATR trailing stop strategy covers the mechanics. This lets the occasional pattern that catches a real trend run.
Do not test both exits in the same run and keep the better one. Test each, read the trade list, and decide which distribution of outcomes you can live with.
Settings Worth Testing
| Setting | Starting point | Range to test |
|---|---|---|
| Pattern | Hammer | Hammer, bullish engulfing, both |
| Location lookback | 20 candles | 10 to 50 |
| Distance from low | 1.5 ATR | 1 to 3 ATR |
| Trend EMA | 200 | 100 to 300 |
| Stop buffer | 0.25 ATR | 0 to 0.5 ATR |
| Timeframe | 4-hour | 1-hour to daily |
Change one setting at a time. If the result only looks good at one exact combination, the pattern is not the edge. The settings are.
Expect the trade count to be low. Three gates plus a pattern is a strict filter. On a 4-hour BTC chart it may fire a handful of times a month. That is the point, but it also means a two-year test may contain too few trades to trust. Our backtest statistical significance guide explains how to judge whether the count is enough.
What Goes Wrong
Loosening the location gate until the trade count looks healthy. Widening the distance from 1.5 ATR to 4 ATR will produce more trades. It will also produce hammers in the middle of nowhere, which is what the gate exists to exclude.
Counting the pattern twice. A bullish engulfing bar that is also within range of the low will pass the detection block and the location gate. That is one signal, not two reasons. Do not size up because several conditions agree. They were designed to agree.
Entering on the pattern candle’s close at the backtest price. In a live run, the signal fires when the candle closes and the order fills on the next candle’s open. Check the backtest fills on the next open, not the signal candle’s close.
Stopping at a fixed percentage. A 2% stop ignores the pattern. The stop belongs below the low the pattern created. Some trades will have a 0.8% stop and some 3%. Size each position by the stop distance so risk per trade stays constant.
Building It in Arrow Algo
The visual builder has every block the strategy needs. No code is involved.
- Data Watcher on your pair and timeframe. Use the completed-candle outputs so every downstream block sees closed bars only.
- Candlestick block with the pattern property set to Hammer. Connect open, high, low and close. Its true output is the trigger.
- Min block with a 20-candle period on the low. This is the location reference.
- ATR block with a 14-candle period. An operation block multiplies it by 1.5. A condition block checks that the current low minus the 20-candle low is below that product.
- EMA block with a 200 period. A condition block checks that close is above it.
- Condition block with the AND operator and three inputs. Feed it the pattern output, the location check and the trend check. Its output connects to the buy action block.
- Stop: a trailing stop block or a condition that sells when close falls below the pattern candle’s low minus a 0.25 ATR buffer. The lag block can hold the entry candle’s low for the comparison.
Run a backtest. Open the trade list before the summary. Look at where each entry sits on the chart. If the entries are not visibly near lows, a gate is misconfigured.
The strategy stays in your workspace until you choose to deploy it. If you do, your rules send signals to your own exchange account. The exchange decides whether to accept and execute them.
Key Takeaways
- A detected pattern is a trigger. Location, trend and a closed candle are the filters that make it worth acting on.
- Put the stop below the pattern’s low, not at a fixed percentage. Size the position by that distance.
- Expect few trades. Check the count is large enough before trusting the result.
- Test one exit style at a time and read the trade list, not just the summary.
- Every block in this guide is available in Arrow Algo’s drag-and-drop builder.
Disclaimer: This content is for educational purposes only and does not constitute financial advice. Trading involves significant risk and you should only trade with capital you can afford to lose. Past performance is not indicative of future results. Always conduct your own research before making any trading decisions.
Ready to build your own automated trading strategies without writing a single line of code? Start for free at Arrow Algo and join thousands of traders who've made the switch to systematic trading.