Candlestick Pattern Settings for MT4: 7 of 59 Variants Beat the Spread

Once the patterns had been measured and compared, the practical question was left over: if you are going to trade them on MetaTrader 4 anyway, which settings should you use? This is the settings grid — every parameter I could vary, ranked by what it actually did to the numbers, on 16,926 EURUSD 5-minute bars from June 3 to August 26, 2026.

The short version is that most of the dials do nothing and the two that matter are not on the indicator panel.

What survived the spread?

Fifty-nine variants, one 1-pip cost charged on every trade, from the spread I measured in my spread and swap costs test. Seven finished positive.

Two of the seven are barely above zero and one of them rests on seventeen trades, which is not a result, it is an anecdote. Strip out the small samples and the honest survivors are the two bold rows. Everything else on this page is a list of things that did not work, which is worth more to you than the two that did.

Which timeframe should candlestick patterns be traded on?

This is the setting that mattered most, and it is not a pattern parameter. Same rule, same bars, different exit:

A three-hour bar out of the M5 chart is where the star lives. If you are reading a five-minute chart and closing trades inside an hour, no candlestick setting will save you — the top row is the only one that goes anywhere, and it goes there slowly.

The practical translation for MT4: find the patterns on M5 if you like, but manage the trade on H1. Press M5 to spot, H1 to hold. The first chart setup guide covers the timeframe toolbar if you have not pinned those two yet.

Does session filtering beat parameter tuning?

By a wide margin. Here is the same star rule at four hours, split by clock:

A ten-point spread in win rate across roughly equal samples, from a filter that costs nothing to apply. And it inverts for the engulfing bar — that one’s worst window is the US session at 36.7% and -2.18 pips, while its best is the quiet 21:00-24:00 block at 51.7%. Two patterns, opposite clocks.

I would not lean hard on any single cell here; these are sixty-trade samples and the split direction is the sort of thing that flips between years. The reliable part is that the hour of day moved the numbers more than any pattern parameter did, which is what my best hours to trade EURUSD measurements and my session against all-day test both found with completely different rules.

What about the “strong signal” filters?

Every candlestick guide tells you to demand a bigger body, a longer wick, a cleaner shape. All of it lost money here.

The wick ratio degrades monotonically — every step up in “quality” cost win rate, ending at one win in four. The ATR body filter has a mild sweet spot at 1x and then falls apart at 2x. Neither ever clears the spread.

The mechanism is the one my optimized against default settings test keeps demonstrating: tightening a filter reduces your sample faster than it improves your edge, and past a point you are optimising on twenty trades. A 19-signal pin bar rule is not a setting, it is a coincidence waiting to be discovered.

Does the EMA 200 trend filter help?

No, and it fails in the direction that should worry anyone who trades patterns “with the trend”.

Counter-trend beat with-trend on both patterns. That is the opposite of the standard advice and I would not build on it — a 200-period EMA on M5 is a 16-hour average, which is a slow reading of a fast chart and probably measuring something closer to yesterday than to now. What it does rule out is the claim that adding an EMA 200 bias improves candlestick entries. On this sample it made them worse.

What to actually put on the MT4 chart

Nothing exotic. MT4 draws Japanese candlesticks natively — Charts → Candlesticks, or Alt+2 — and every pattern on this page is read straight off them. There is no built-in pattern-recognition indicator in MT4, and the third-party ones in the Market do the same shape matching I did in code, with the same results waiting for them.

Two EMAs, candlesticks, an ATR for the stop. That is the whole toolkit, and the ATR stop is there because my fixed against ATR stop comparison found it the least bad structure, not because it rescues the entry. For lot sizes, the table in my position sizing guide does the arithmetic for a 10,000 account.

One warning about pattern indicators specifically. Anything that paints an arrow on a forming candle can un-paint it — the shape is only final when the bar closes. That is exactly the failure my arrow indicator repaint test measured at 88.2% in history against 46.7% live. Every number on this page assumes you entered at the close of a completed bar.

What I would actually do with this

• Set the hold before the pattern. Four hours was worth more than every parameter on this page combined.
• Do not tighten the shape. Wick 3x body took the pin bar to 26.3% wins. “Strong signal” filters remove trades, not losses.
• Use the clock. A ten-point win-rate spread across sessions, free to apply, larger than anything tuning delivered.
• If you want a confirmation, use the slow one. EMA 12/26 turned the engulfing rule positive; MACD nearly doubled its loss.
• Expect seven in fifty-nine. That ratio is the real finding. Anyone showing you one winning candlestick setting is showing you one cell of a table like this one.

Frequently asked questions

What are the best candlestick settings for MT4?

Candlesticks (Alt+2), patterns spotted on M5 or M15, trades held about four hours, and an EMA 12/26 pair as the only confirmation. Of 59 variants tested on 16,926 EURUSD 5-minute bars, that combination produced the two results that cleared a 1-pip spread on a full sample.

Is there a candlestick pattern indicator built into MT4?

No. MT4 draws candlesticks natively but has no pattern-recognition indicator in the standard install. Third-party ones apply the same shape rules tested here, so they inherit these results — and any that mark a forming bar can repaint before it closes.

Which timeframe is best for candlestick patterns?

The exit horizon mattered more than the chart. On M5 bars, the morning/evening star went from -0.17 pips at a 15-minute hold to +1.92 at four hours, then fell back to +0.58 at eight. The engulfing bar and pin bar never turned positive at any horizon.

Should I use an EMA 200 trend filter with candlestick patterns?

It did not help here. Engulfing bars taken with the EMA 200 trend won 42.4% at -0.85 pips; taken against it they won 45.0% at +0.09. Both pin-bar variants lost. Adding the filter made the entries worse, not better.

What stop loss should I use on a candlestick trade?

1x ATR(14) was the least bad structure tested, but be realistic about the target. A 2:1 target was reached only 29.4% of the time on engulfing bars and 32.8% on stars, against the 33.3% needed just to break even before costs. Details in my stop loss distance test.

Does a bigger candle body mean a better signal?

Only slightly, and only up to a point. Requiring the engulfing body to exceed 1x ATR improved the average from -0.64 to +0.16 pips. Pushing it to 2x ATR dropped it to -1.38 across 69 trades. Neither cleared the spread.

Where do these numbers come from?

16,926 EURUSD 5-minute bars, June 3 to August 26, 2026, entries at the close of the completed pattern bar. The full pattern-by-pattern breakdown is in my candlestick pattern test and the indicator comparison is in candlestick patterns against MACD and moving averages.