Why past charts always look easy. Hindsight bias and lines that never miss

Open yesterday’s chart and everything looks clear. Price bounced off this line, broke that one, and reversed right at the zone you drew. But while the chart is actually moving, it never feels that clear. This gap has a name: Hindsight bias. It’s one of the reasons people trust a trading method more than it deserves.

What is hindsight bias?

Hindsight bias is the feeling, once you know the outcome, that it was predictable from the start, even though you weren’t sure at all before you knew.

On a chart this effect is especially strong, because you see the whole story’s ending in one picture. Your eyes go to the points where price reversed first, then you look for reasons to explain them. The points where price simply ran through a line rarely catch your attention.

More lines, more “accurate” lines

Try counting how many lines are on your chart. Say you use Fibonacci Retracement (a tool that draws pullback levels) with 5 levels: 23.6, 38.2, 50, 61.8 and 78.6.

• Drawn on 3 wave sizes (small, medium, large), that’s 5 × 3 = 15 lines
• Looking at two timeframes, such as M1 and H4, that’s 15 × 2 = 30 lines

With this many lines, price is near one of them almost all the time. So whenever price reverses, there’s almost always a line somewhere nearby. Looking back, that line seems like the reason for the reversal.

A past chart already holds the answer. With enough lines to choose from, you will always find one that matches it.

Real data: does price sit at the lines more than chance?

We measured this on gold, using every M1 candle between 06:00 and 09:00 Thai time (UTC+7), from July 2022 to August 2026, a total of 192,728 candles. We placed Fibonacci levels on 3 wave sizes on both M1 and H4, 6 frames in total. Each frame used only the data available at that candle, never looking ahead.

The first question: is price near a Fibonacci level (within 3 levels, e.g. 58.8–64.8 for the 61.8 line) more often than it should be? The comparison is how often price would land in the same zones if its distribution were smooth and ignored the lines completely.

Frame	Actually near a line	If lines were ignored	Ratio
M1 large wave	21.08%	21.24%	0.99
M1 medium wave	24.24%	24.37%	0.99
H4 medium wave	20.92%	21.91%	0.95

All 6 frames gave ratios of 0.95–1.00. That means price sat at Fibonacci levels about as often as chance would give, not more. And in the M1 small-wave frame, candles touched a line in about 60% of all candles. So “touching a line” in a small frame means almost nothing on its own.

The second question: does price at a line bounce more often? For each candle, we checked whether price bounced back the original way by half an ATR first, or kept pulling back another half ATR first (ATR is the average candle range). If the lines had no effect, you’d expect about 50%.

Frames where price sits at a line at once	Bounced first
0 frames	49.79%
1 frame	49.68%
2 frames	50.03%
3 or more frames	50.19%

Even when price sat on lines from several frames at once, which many people call a strong confluence, the result was still around 50%, the same as with no line at all.

This data is gold in one time window of the day, measured on the past. It does not predict the future, and it doesn’t prove Fibonacci fails in every market. But it’s enough to say that the “price bounced right off the line” picture you see on a past chart is not evidence on its own.

Why we only remember the hits

The brain keeps things that make a story. Price bouncing right at 61.8 is a good story. Price simply running through 61.8 is no story at all. Over time, your memory fills up with the hits, and your confidence keeps growing, even though the data hasn’t changed.

The result: you open bigger trades because you feel sure, or you blame yourself with “how did I miss something so obvious?”, when at the time it wasn’t really obvious.

How to protect yourself from past charts

1. Write your plan before price moves. Note the entry, Stop loss and target with the time, or save a screenshot. If you didn’t write it down beforehand, don’t count it as a correct signal.
2. Count every time, including the misses. A good trading journal includes the times price ran through a line, not just the times it bounced.
3. Have a comparison. Try placing made-up lines with no logic, such as evenly spaced lines, and count how often price bounces from those. If it’s about the same, there’s no evidence yet that your tool helps.
4. Use fewer lines. Keep the tools and wave sizes you actually use to a minimum, and set them in advance. Don’t redraw after you’ve seen price.
5. Review with the future hidden. When practising, move the chart forward one candle at a time and decide before you see the next one.

Common mistakes

• Looking at past charts and deciding a method is “very accurate” without ever writing down signals in advance
• Redrawing lines after price reverses, so they match the turning point exactly
• Adding tools one at a time so the chart can explain every point. The more you add, the more accurate it looks, but it isn’t really more accurate
• Being very hard on yourself after missing a point that looks “obvious” on a past chart