Does Ichimoku Repaint? 0 of 48,320 Values Changed, but Reading the Chikou Off the Chart Lifted Wins From 37% to 58%

“Does Ichimoku repaint?” gets two answers online: no, it is a set of moving midpoints, and yes, because the Chikou span moves. Both are half right. I measured each half on 5,642 fifteen-minute EURUSD bars and 4,577 gold bars, and on the 5-minute charts they were built from, the same bars behind my Ichimoku cloud breakout test. The comparison that matters is not repaint against no repaint. It is what the history chart shows against what you could have known at the time.

Does the Ichimoku cloud change on closed bars?

No. I took 40 random cut-off points on each market, recalculated Ichimoku 9/26/52 using only the bars up to that point, and compared the last 151 bars with the values from the full history.

EURUSD and gold M15, 6,040 bar positions on each, June to August 2026.

That is what you would expect: every one of those lines uses only bars that have already closed. The cloud is drawn 26 bars into the future, but it is built from the past, so the part you see ahead of price is already fixed. The current, unfinished bar moves Tenkan and Kijun while it forms, like any indicator, and stops when it closes.

So what is wrong with the Chikou span?

Nothing, on a live chart. The Chikou span is today’s close, drawn 26 bars to the left. On the right edge of your chart the last 26 candles have no Chikou point above them yet. They get one only after 26 more bars.

The trouble starts when you scroll back. Above every past candle there is now a Chikou point, and it is the close from 26 bars later. A rule like “buy when Chikou is above price” looks easy to check by eye: look at a candle, look at the line above it. But that line tells you where price went next. A trader on that candle could not see it.

The honest way to read Chikou at a candle is to compare that candle’s close with the close 26 bars before it. That is what the line meant at the time.

How much better does Ichimoku look on the chart than live?

I took the Ichimoku “strong signal” that many courses teach: a Tenkan/Kijun cross, price on the right side of the cloud, a cloud of the same colour ahead, and Chikou above price for a buy or below it for a sell. I ran it twice on the same bars. Once with Chikou read live, once with Chikou read where it is drawn. Exit: stop 1.5 ATR, target 3 ATR, closed after 48 bars, spread included.

After a 1-pip cost on EURUSD and $0.30 an ounce on gold. “Chart” reads Chikou at the candle it is drawn over, which is the close 26 bars later.

On EURUSD M5 the live rule lost 1.40 pips a trade. Checked on the chart it would have looked like a system winning more than half its trades. On EURUSD M15 the gap was 4.28 pips a trade, more than four times the spread. And the chart version fired fewer times, 41 against 67, because the Chikou condition on the chart removes the trades that went the wrong way.

With a shorter exit the gap grows. Closing after 12 bars, three hours on M15, sits fully inside the 26 bars the chart line already knows:

Seven in ten winners, on every market. That is the number a trader scrolling back through a chart will believe, and it is why Ichimoku screenshots look so convincing.

Would splitting the test into two halves catch it?

No. On EURUSD M15 the chart version made +4.46 pips a trade in the first half and +2.61 in the second. On gold M15, +$21.90 and +$12.35. Both halves positive, both comfortable. A split-half check, the one I use in every test such as gold vs EURUSD with one rule, catches a rule that got lucky in one stretch. It cannot catch a rule that reads the answer, because the answer is there in both halves.

This is the same trap as a repainting arrow indicator, without the repaint. With arrows the past changes. With Chikou the past stays put, but what is drawn over it came later. The fix is the same: count a signal only from the moment it could be seen.

How do I check any indicator for this problem?

1. Cut the history. Note an indicator value on a closed bar, then look at the same bar after 50 more bars. If it changed, the indicator repaints. Ichimoku passes this test, as above.
2. Ask where each line is drawn. Any line shifted to the left (Chikou, centred moving averages, a ZigZag point) shows over a candle something that happened after it. Never judge a rule by eye against such a line.
3. Look at the right edge. If the last bars of the chart have no value for a line, the line on older bars is using bars that came later.
4. Run it forward. Watch the rule on a demo chart for a few weeks and write down signals as they appear. My backtest vs demo test shows how far the two can drift even without a look-ahead, and tick data vs spread shows a second trap: the order of prices inside a bar.

In MT4 and MT5 you can do step 1 in the Strategy Tester in visual mode: pause, note a value, run on, compare. My guide to backtesting on MT5 covers the tester.

Does this mean Ichimoku is useless?

No. Read live, the same strong signal made +$3.87 an ounce a trade on gold M15 and passed every check in my Ichimoku 15-minute chart test. It lost on EURUSD, where the live win rate was 37.3%. The point is that the honest version wins about four trades in ten, not seven. If a course shows you seven in ten from a chart, it is probably reading Chikou where it is drawn. My Ichimoku settings guide for MT4 shows how to set the lines up.

Frequently asked questions

Does the Ichimoku cloud repaint?

No. Recalculated on 80 cut-off histories, 0 of 48,320 values on closed bars changed. Tenkan, Kijun and both cloud spans use only closed bars.

Does the Chikou span repaint?

Not in the usual sense: once drawn, a Chikou point stays. But it is drawn 26 bars back, so on a history chart the point over a candle is the close 26 bars later. Reading a rule off it uses information nobody had at that candle.

How much does the Chikou look-ahead change a backtest?

On EURUSD M15 the strong signal won 37.3% of trades read live and 58.5% read off the chart, and went from -0.68 pips a trade to +3.60. With a 12-bar exit the chart version won 68% to 77% of trades on every market tested.

How should I use the Chikou span live?

Compare the current close with the close 26 bars ago. If it is higher, Chikou is above price. That is the only version you can trade.

Is Ichimoku safe to backtest in MT4?

The built-in Ichimoku in MT4 and MT5 does not repaint. An expert advisor that reads the Chikou buffer at a past bar index would be reading a later close, so check how the code reads it.