Candlestick Patterns in Futures: Use Them as Confirmation at a Level

A hammer candle in the middle of a quiet range, far from any level you care about, tells you almost nothing. The same candle shape at a price where you already had a plan is a different thing. Most candlestick mistakes in ES and NQ come from reading the shape and ignoring where it formed.

What a candle can and can't tell you

A candle summarizes one time period: open, high, low and close. The shape shows how price moved between buyers and sellers inside that period. It doesn't predict the next one. I haven't seen documented evidence that any single pattern gives a reliable edge by itself in ES or NQ, so I treat patterns as confirmation, not as setups. The setup is the level and the plan. The candle helps me decide whether the plan is being respected.

Levels can be the prior day's high and low, VWAP, round numbers, or fixed reference prices such as the KPL levels, which are anchored to the regular-session opening print and don't move during the day. They're reference prices, not predictions of where price will turn.

Five patterns worth knowing

Hammer

A small body near the top with a long lower wick, roughly two times the body or more. It shows price was pushed down and rejected within that candle. At a level you were already watching, it can support a long with a stop below the wick low. If price trades back under that low, the idea is wrong and you're out. Target the next reference level.

Engulfing

A bullish engulfing candle opens at or below the prior candle's close and closes above the prior candle's open, covering its body. It reads as a change in control. Context matters: after a push into a level it says more than in the middle of chop. A larger engulfing candle shows a stronger shift on that chart, though it also means a wider stop, which changes your risk.

Doji

Open and close nearly equal. It shows balance, not direction. A doji after a rally into a prior high is a prompt to pay attention, and the next candle decides what you do. A doji in the middle of a range means very little.

Pin bar

A small body with one long wick and almost none on the other side. The wick marks where price was rejected. A pin bar at VWAP or a reference level can give a defined stop just beyond the wick. If price trades through the wick, the rejection failed.

Inside bar

A candle fully inside the prior candle's high-low range. It marks compression. A common plan is to set alerts at the inside bar's high and low and trade the break in either direction, with the stop on the other side. Inside bars break both ways and fake out often, so wait for a close beyond the range and size the trade for a stop equal to the bar's full height.

Patterns I'd be careful with

Three-candle formations like morning star, evening star, three white soldiers and three black crows need three completed candles, so the entry comes after much of the move. Harami candles are small and ambiguous on fast charts. I'm not saying they never work. I'm saying I haven't tested them and wouldn't trade them on a hunch.

A three-part filter

My process asks for three things before acting: a pattern, a level, and a read on market internals such as NYSE TICK turning in the direction of the trade. Two of three is marginal, and I either skip it or cut size. This reduces how many trades you take. It does not make the remaining ones win more often, and I won't claim that without a documented test.

Cost math for a pin bar trade

Hypothetical example on ES: price rejects a level, the pin bar's wick low is 5,201, and you buy at 5,204 with a stop at 5,200.75. That's 3.25 points at risk, or $162.50 per ES contract at $50 per point, or $16.25 per MES at $5 per point. A target at 5,216 is 12 points, $600 per ES. Costs also include commissions and slippage on the entry and the stop, and a stop order can fill beyond your price in fast conditions.

How to test a pattern

1. Pick one pattern, one chart timeframe, one level type and one session window. Write the rule down, including entry, stop and target.
2. Mark every occurrence on historical charts without peeking at the outcome, or record each one live in a journal.
3. Log result in points and dollars net of commissions, plus an estimate of slippage.
4. Look at the sample size before the average. A few dozen trades can show whether you can follow the rule and where it breaks. They don't establish an edge.
5. Run the same rule on a different period before you trust it, and treat simulated fills as optimistic because they don't reproduce queue position or slippage.