Heikin Ashi and Renko Charts: When They Show the Trend and When They Mislead

A normal candlestick chart shows the real open, high, low and close of each period. It’s complete, but it’s noisy: even a clean uptrend is full of red candles. Heikin Ashi and Renko are two alternative charts that try to strip out the noise so the trend is easier to see. The price you pay is that they no longer show market prices directly. This article explains how each is calculated, when it helps and when it misleads.

What is Heikin Ashi

Heikin Ashi (Japanese for “average bar”; MT4/MT5 spell it Heiken Ashi) is a candle built from averages instead of raw prices. Each candle still has four values, calculated like this:

Value	Formula	Meaning
HA Close	(Open + High + Low + Close) ÷ 4	Average price of this real bar
HA Open	(previous HA Open + previous HA Close) ÷ 2	Midpoint of the previous HA body
HA High	Highest of real High, HA Open, HA Close
HA Low	Lowest of real Low, HA Open, HA Close

The very first candle has no previous HA candle, so most software sets the first HA Open = (Open + Close) ÷ 2 of the real bar. The effect of this starting point fades within a few bars.

Notice that HA Open uses nothing from the current bar’s real prices. It comes entirely from the previous HA candle. So Heikin Ashi behaves like a moving average with memory: each candle leans on the one before it, which leans on the one before that.

Worked example: 3 bars

Three real bars (Open / High / Low / Close):

• Bar 1: 100 / 104 / 99 / 102
• Bar 2: 102 / 105 / 101 / 104.5
• Bar 3: 104.5 / 105 / 101.5 / 102 (a red bar, closing 2.5 below its open)

Bar 1

• HA Close = (100 + 104 + 99 + 102) ÷ 4 = 405 ÷ 4 = 101.25
• HA Open = (100 + 102) ÷ 2 = 101 (first bar, so real Open and Close)
• HA High = highest of 104, 101, 101.25 = 104; HA Low = lowest of 99, 101, 101.25 = 99
• Close > Open, so it’s green

Bar 2

• HA Close = (102 + 105 + 101 + 104.5) ÷ 4 = 412.5 ÷ 4 = 103.125
• HA Open = (101 + 101.25) ÷ 2 = 101.125
• HA High = 105; HA Low = lowest of 101, 101.125, 103.125 = 101
• Green

Bar 3

• HA Close = (104.5 + 105 + 101.5 + 102) ÷ 4 = 413 ÷ 4 = 103.25
• HA Open = (101.125 + 103.125) ÷ 2 = 102.125
• HA High = 105; HA Low = lowest of 101.5, 102.125, 103.25 = 101.5
• HA Close 103.25 is above HA Open 102.125, so it is still green, even though the real bar is red and the real close is 102

This one bar shows both sides. The good: a single red bar in an uptrend doesn’t flip the HA colour. The bad: if this really is the start of a reversal, HA tells you later, and the HA “close” of 103.25 sits 1.25 above where the market actually closed.

How to read Heikin Ashi

• Green candle with no lower wick: steady buying pressure, because the real Low never went below HA Open
• Red candle with no upper wick: steady selling pressure
• Small body with wicks on both sides: indecision; the trend may be weakening or turning sideways
• A colour change means the average has turned, which always happens after real price has already turned

In the figure above, the first uptrend has several real red candles mixed in, but HA stays green the whole way. In the sideways stretch (yellow band) HA still flips back and forth. On the last bar the real close is 100.65 while the HA close is 101.61, almost a full price unit apart. If you read the price scale on an HA chart and think the market is at 101.61, you are wrong straight away. Some software shows the HA close on the last-price label rather than the market price, so check your platform’s settings.

Why you can’t trade Heikin Ashi prices

HA Close is an average of four prices. It does lie within the real bar’s High-Low range, but you only know an HA signal once the bar has closed, and at that moment the market price is the real close, not the HA Close. HA Open and the candle’s wicks are calculated numbers too.

So:

1. Entries use real market prices only
2. Stop loss and Take profit are placed from real prices, for example below the real Low of the bar or based on ATR, not below the HA candle’s Low
3. Position size comes from the real Stop loss distance, as in position size from Stop loss

Use HA candles for direction, not for prices. Every number that involves money, from entry to Stop loss to lot size, must come from real prices.

How optimistic are backtests on HA?

We tested this on synthetic, purely random prices with no hidden trend, so no system should be able to make money. We built 20 series of 5,000 bars each and applied a simple, popular rule: Buy when HA turns green, Sell when it turns red, always reversing. That gave 24,229 trades. Then we scored them two ways:

Filled at	Average profit per trade	As share of average bar range	Win rate	Series in profit
HA Close (price on the HA candle)	+0.608	+44.5%	50.3%	20 of 20
Real close of the same bar	+0.016	+1.2%	35.7%	14 of 20

The average High-Low range of a bar was 1.366. Same rule, same signals; only the fill price changed. Nearly half a bar’s range per trade turned into roughly zero, before subtracting spread, which would push it negative. The reason: when HA has just turned green, HA Close is usually below the real close (80.1% of the time in this test, by 0.297 on average), so the backtest “buys cheap”, and when it turns red it “sells high” compared with any price actually available at that moment. This is an experiment on random prices, not a result from a real market, but it shows that this kind of profit comes from how prices are calculated, not from skill.

If a backtest on an HA or Renko chart looks unusually good, re-run it with fills at real market prices before you believe it.

Some charting platforms warn about this or offer an option to fill strategy orders on real OHLC prices while you view an HA chart. If you test yourself, check whether it’s on. For more on the gap between tests and live trading, see why backtest and live results differ.

What is Renko

Renko (from the Japanese renga, “brick”) ignores time completely. The chart is made of equal-sized bricks, and a new brick appears only after price moves a full box.

• Box size is either a fixed price amount, such as 1.00 or 10 pips, or based on ATR so the box tracks volatility
• A brick in the same direction needs a close one full box beyond the edge of the last brick
• A reversal brick, in traditional Renko, needs price to move two boxes against the trend from the top of the last brick (one box beyond its bottom)
• If price doesn’t move a full box, nothing is drawn, however many bars go by
• One strong bar can create several bricks at once

Example on the same price series

Real closes, box size 1.00, starting at the 101 level:

• The first up brick, 101–102, forms on bar 5 at a close of 102.09, the first close to reach 102
• Bar 13 closes at 109.82, starting from the last brick top at 106, so this single bar creates 3 bricks: 106–107, 107–108 and 108–109
• The highest close is 110.17 on bar 18, which creates the 109–110 brick. The first down brick (108–109) then needs a close below 108, which happens on bar 28 at 107.28: 10 bars after the peak and 2.89 below the highest close
• 40 bars become 17 bricks. With a 2.00 box they would become just 7

Renko pitfalls

• Time disappears. The 9-bar sideways stretch above leaves no trace. The chart looks calm, but in reality price was swinging back and forth while spread and swap costs kept running
• The latest brick isn’t final. A forming brick isn’t real until price closes a full box. If your software builds bricks from prices inside an unfinished bar, a brick can appear and then vanish before the bar closes
• ATR-based boxes can redraw. With an ATR box size, some software recalculates the box as new data arrives, so past bricks can change when you reload the chart. Check how yours does it
• Brick edges are not fill prices. In the example, bar 13 created 3 bricks from a single close at 109.82. By the time you know there is a new brick, the bar has already closed. A backtest that enters at the first brick edge of 107 is using a price that has already gone, and after a gap the brick edge may never have traded at all. It’s the same problem as with HA
• Box size changes the picture. A 1.00 box shows the pullback; a 2.00 box may not show it at all. No size is correct. Pick one and keep it, rather than tweaking it until the past looks pretty, which is a form of overfitting

MT5 ships Heiken Ashi as an example indicator but has no built-in Renko chart; you need an add-on tool. Charting platforms such as TradingView offer Heikin Ashi, Renko and other types as chart styles.

Related chart types

• Line break (usually 3-line): a new line is added when the close goes beyond the high or low of the last line, and it reverses only when the close goes beyond the extreme of the last 3 lines. Like Renko it removes time, but with no fixed box size
• Range bars: every bar has the same High-Low range, and a new bar starts when price exceeds it. Quiet markets produce few bars, volatile markets produce many

Both share the same caveats: time is removed, and the prices on the chart may not match moments when you could actually trade.

When they help and when they mislead

Situation	Heikin Ashi	Renko
Clear trend	Long runs of one colour help you avoid exiting too early	Highs and lows are clear; easy to draw trendlines
Sideways	Still flips colour; can whipsaw	The range vanishes; the market looks calm
Fast reversal	Changes colour later than real price	Needs a two-box move against the trend
Big news / gaps	The average hides part of the move	Several bricks at once, at prices that never traded
Placing entry and Stop loss	Don’t; use real prices	Don’t; use real prices

A common and safer approach is to use HA or Renko as a “view” chart on a higher timeframe and place orders on a normal candlestick chart, similar to multi-timeframe analysis. And if your approach profits from following trends, accept that it will lose more often when the market goes sideways; see ride the trend or fade it.

Common mistakes

• Putting the Stop loss below the HA candle’s Low or below a Renko brick edge when the real price may be somewhere else
• Trusting a backtest on an HA or Renko chart without checking which prices it fills at
• Thinking HA predicts reversals, when it is an average of prices that have already happened and always changes colour after price does
• Watching Renko and forgetting that time passes, holding trades over several days without counting swap
• Tweaking the box size or settings until the historical chart looks best, then expecting the future to match. This trap is related to hindsight bias on charts and trading biases
• Reading candlestick patterns such as Engulfing or Doji on HA candles, which are not the candles those patterns were defined on