Skip to content
GetProfitable
Search

Hindsight bias, over-marking and ignoring context

Lesson 22 · about 9 min

Three errors account for most of the gap between how good chart reading looks in a course and how it performs in a live account. None of them is a knowledge problem; every trader who makes them knows better. They are habits of attention, and the fix for each is a procedure, not a fact.

1. Hindsight bias

Hindsight bias is seeing a chart with its outcome visible and believing you would have read it correctly without the outcome. It is the reason every chart in every trading book looks obvious and every live chart looks ambiguous.

The mechanism: once you know price went up from a level, your eye finds the hammer, the level, the trend, and ignores the two earlier candles at the same level that looked just as good and failed. The failed ones are still on the chart. You are not looking at them because they do not fit the story you already know.

How it shows up:

  • "I would have bought there." Only if you had a rule that also bought the three previous, similar places that lost.
  • Scrolling a chart from left to right with the whole thing visible and "spotting" setups. You are spotting outcomes.
  • Redrawing a level after a candle so that the candle is "at the level". The level is now defined by the outcome.
  • Choosing N for swing points after the fact so the trend label matches the move that happened.

The fix is the one from Lesson 1: cover the right side of the chart. Write the row before advancing. Choose N before starting and do not change it. Any analysis done with the outcome visible is practice at recognizing shapes; it is not evidence of an edge.

Key idea: A chart with its future visible is a different object from a chart without it. Only conclusions drawn from the second kind count.

2. Over-marking

Over-marking is a chart with so many levels, lines and zones that price is always "at a level". It is the natural result of Module 3 applied without the deletion rules, and it destroys the value of levels by making them everywhere.

  Over-marked                           Marked honestly

  ----------------------------          
  ///////////////////////////           ///////////////////////////   <- major swing high
  ----------------------------          
  ----------------------------          
  ///////////////////////////           
  ----------------------------          
  ----------------------------          - - - - - - - - - - - - - -   <- PDH
  ///////////////////////////           
  ----------------------------          
  ----------------------------          ///////////////////////////   <- last higher low
  ///////////////////////////           
  ----------------------------

Signs you are over-marking:

  • More than about eight lines on the trading-timeframe chart.
  • Levels from more than one year ago on a daily chart, with no recent reaction.
  • Lines you cannot attach a one-sentence trapped-group story to.
  • Two zones that overlap; that is one zone.
  • Fibonacci levels, pivot points, moving averages and trendlines all drawn at once. Each of these produces several lines; together they cover the chart.

The consequence is worse than clutter. If price is always near a level, every candle pattern passes the "at a level" test, and the confluence filter from Module 5 stops filtering. You are back to trading shapes with extra decoration.

The fix: the map from Module 3, Lesson 2. Major swings (N = 10), minor swings inside them (N = 3), the calendar levels for the session, and VWAP if intraday. Delete everything that fails the one-sentence rule, and delete intraday levels at the end of each session. If the chart still looks busy, raise N.

3. Ignoring context

Ignoring context is trading the candle or the level without the trend, or trading the trading-timeframe structure without the context timeframe. It is the error that Module 2's false-signal rates measure: most patterns are in the wrong place or against the trend, and a trader who does not check will take most of them.

How it shows up:

  • Shorting a shooting star at a minor swing high while the weekly is in a strong uptrend, because "it's a perfect shooting star".
  • Buying a range bottom on the 5-minute chart while the daily just printed a CHoCH down through that same price.
  • Taking a breakout in a market where ADX has been below 15 for a month and every previous breakout has failed.
  • Trading a pullback-to-level setup in the middle of a range, where there is no trend to pull back in.

Each of these has a level and a candle. Each is missing the third question. And each will, on a large enough sample, lose money, because the trapped group the candle describes is small and the group on the other side (the trend, the higher-timeframe structure) is large.

The fix: the structure check from Module 4, Lesson 4, run before looking for a candle, every time. Write the context and trading timeframe labels at the top of the chart or in the trade log. If they disagree, or either is "undefined", the answer is to wait, and waiting is a decision you can write down.

A fourth, briefly: outcome-based evaluation

Judging a trade by whether it won rather than whether it followed the rules. A trade that broke every rule and won teaches you to break rules. A trade that followed every rule and lost teaches you nothing you did not already know from the backtest, where 29 out of 50 lost.

Grade each trade on two axes: did it follow the rules (yes or no), and did it win (yes or no). The trades to study are the rule-followers that lost (was it noise, or is a rule wrong?) and the rule-breakers that won (what rule did you break, and would you have broken it if it had lost?).

Followed rules Won Lesson
Yes Yes Good trade; nothing to learn
Yes No Good trade; check whether it was noise
No Yes Bad trade; the win is the dangerous part
No No Bad trade; the loss is the tuition

Try it: Take your last ten trades (real or paper) and grade each on the two axes above. Then, for each, write which of the three errors in this lesson was present, if any. Most traders find at least one error in more than half of their trades, and that the errors cluster: one of the three is your habit. Name it.

Recap

  • Hindsight bias: a chart with its outcome visible makes every setup look obvious; only covered-chart analysis counts as evidence.
  • Over-marking: too many levels means price is always "at a level" and the confluence filter stops working; keep to a handful with a story each.
  • Ignoring context: a level and a candle without the trend and the context timeframe is the setup that the false-signal rates describe.
  • Grade trades on rule-following and outcome separately; the rule-breaking winner is the most dangerous trade you will take.
  • Each error is fixed by a procedure (cover the chart, delete lines, run the structure check), not by knowing about it.

See it drawn

Original diagrams for the ideas on this page. Illustrative, not real market data.

The parts of a candlestickAn up candle and a down candle with the same high and low, labelled with open, high, low, close, the real body and the wicks.UP CANDLEclose above openHigh 41.00Close 40.30Open 38.20Low 37.40upper wickreal bodyopen to closelower wickDOWN CANDLEclose below openHigh 41.00Open 40.30Close 38.20Low 37.40Same high and low; only the open and close swap places.
The parts of a candlestick. One candle sums up a slice of time: the thick real body runs from the opening price to the closing price, and the thin wicks reach out to the highest and lowest prices traded. Colour tells you which way the body ran.
One daily candle broken into four six-hour candlesA tall daily candle on the left and the four six-hour candles that make it up on the right, with dashed lines linking the day's open to the first candle and the day's close to the last.ONE DAILY CANDLEFOUR 6-HOUR CANDLEScloseopenhighlow=00:0006:0012:0018:00one dayThe same trading, summed up in one bar or spelled out in four.
How timeframes stack up. A daily candle is not different data, only coarser data: it opens where the first six-hour candle opened, closes where the last one closed, and its wicks reach the highest and lowest prices any of the four touched.
Trend structure: higher highs against lower lowsTwo zigzag price paths side by side; the left one steps upward with each peak and trough above the last, the right one steps downward with each peak and trough below the last.UPTRENDhigher highs, higher lowsHHHHHHHLHLHLDOWNTRENDlower highs, lower lowsLHLHLHLLLLLLHH higher high, HL higher low, LH lower high, LL lower low.
How a trend is built. A trend is just a sequence of turning points. While each peak and each dip sits above the one before it the market is trending up; once both start landing below the previous ones the structure has turned down.