Level 1 Reading charts · Part 7/7
What Technical Analysis Can and Cannot Do
The first level ends with the uncomfortable question: does any of this work? The honest answer is nuanced – and it is the reason the next two levels are structured the way they are.
Three mechanisms that genuinely operate
Self-reference. When enough participants watch the same level, orders really do accumulate there. The effect is real but does not scale: the better known a level, the more attractive it becomes as a target for the other side.
Order mechanics. Stop orders cluster at visible prices. When they trigger, a short-term move follows that is not a change of opinion but forced execution. That explains false breakouts and fast post-break moves better than any pattern taxonomy.
Behavioural patterns. Loss aversion, anchoring and herding are well documented and produce recurring price structures. What they do not supply is timing.
Of all of this, the best-evidenced result is not the pattern catalogue but momentum: the observation that instruments which performed relatively strongly over the past six to twelve months tend to continue for a few more months. The effect has been studied across decades and markets – and it is nothing other than a trend statement in measurable form.
Four reasons patterns look better than they are
The base rate is missing. “In seven of ten cases a rally followed” is only a statement if you know how often a rally follows anyway. If equities rise in 54 per cent of all weeks, a pattern must clear 54 per cent by a wide margin to carry any information.
Selection bias in the examples. Textbooks show formations that worked. The ones that failed look identical until shortly before the end – and never make it into the book. That is why every graphic in this course is synthetic: a real example would always be a selected one.
Hindsight. In a finished chart every pattern is obvious. At the right-hand edge, where the decision is made, they are candidates – and most of them will not make it.
Degrees of freedom. Timeframe, scale, indicator parameters, the definition of a zone: with enough adjustable knobs, some combination always would have worked in the past. How quickly that search turns into an apparently excellent result is the subject of Level 3 .
What follows from this
Not that charting is worthless – but that its contribution sits elsewhere than usually claimed:
- It provides a shared language for price structure, and therefore comparable descriptions.
- It provides testable levels: a pattern with a clear refutation point is what makes risk management possible at all.
- It provides no forecast. If you want a probability statement, you have to measure it – on a large enough sample, with a clean separation of training and test periods.
That is exactly the path this course takes. Level 2 describes patterns and indicators precisely enough to make them testable at all. Level 3 shows what goes wrong during testing and how financial research deals with it.
To put an idea against historical data right away, the backtester and the return triangle are the place to start – with the caveat Level 3 spells out: one good backtest proves very little.
Next
End of Level 1. Continue with candlestick patterns in Level 2.