Technical Analysis vs Quantitative Analysis
Technical analysis reads charts using patterns, levels and indicators that are interpreted by a person. Quantitative analysis instead states a rule precisely enough to apply mechanically to historical data and then measures the result, including on periods the rule was never built from.
Both of these work from the same price history. One interprets it with a trained eye and the other insists on a rule that can be run over the data and measured. The difference is not sophistication — it is whether the claim can be shown to be false.
What each one is
Technical analysis reads charts through patterns, levels, indicators and trend, interpreted by a person. Technical analysis covers the tradition.
Quantitative analysis states a rule precisely, applies it mechanically to history, and measures what happened — including on periods the rule was not developed on. Quantitative analysis covers the method, and price action covers the most judgement-heavy end of charting.
One produces readings and the other produces claims. Whereas a chart reading describes what a person sees, a quantitative claim exists as a specification that anybody can run and disagree with.
Where they differ
Whether the claim can be checked. A chart reading is difficult to test because it depends on which level, which pattern and how strictly. A specification can be run by anyone and produces the same answer every time, which is what makes disagreement productive.
How each handles randomness. This is the substantive point. People find convincing patterns in sequences that contain none — the ability is reliable and it does not switch off when the data is noise. A test is the only instrument that distinguishes the two cases.
What each can adapt to. A person can weigh an unusual situation — a company event, a market condition, something that has never happened before. A rule cannot, and will apply itself confidently in circumstances nobody anticipated.
How each is usually taught. Technical material is overwhelmingly retrospective: charts marked up after the outcome. Quantitative work insists on out-of-sample results precisely because fitting the past is easy and means nothing.
Where they agree
Both work from price history, and neither has access to anything the market has not already printed.
Both can be fitted to the past. A reader can mark up a chart to fit the outcome; a modeller can tune parameters until the backtest looks good. It is the same error in two languages.
Both fail in the same conditions. Direction runs on this site’s shared series average 2.01 bars with a longest of 11, which is where false patterns and false signals both come from.
And both cost a round trip when acted on — 0.0098 here, about 2% of the median bar range of 0.493.
Which one to use
Use a test when you want to know whether something works. That is the only question quantitative analysis exists to answer, and no amount of chart experience substitutes for it.
Use judgement when the situation is genuinely novel. Rules cannot handle events they were never specified for, and a person who recognises that something is unusual is doing work no backtest can do.
Use both by specifying your chart reading. Writing down exactly which level, which condition and what would disprove it converts a reading into something testable — which is available to anybody willing to be that precise.
And when a pattern looks obvious, test it before trusting it. The obviousness is the warning: the eye produces that feeling for real patterns and imaginary ones identically.
Why finding patterns in noise is the central problem
Because the faculty that reads charts well is the same one that over-reads them. Pattern recognition does not come with a confidence signal that distinguishes signal from noise, so the experience of seeing something real and seeing something spurious is identical from the inside.
And because a small sample produces the most convincing shapes. A handful of bars can form a textbook pattern by chance, and the fewer observations there are the more likely that becomes.
The original data
Of the 24,971 videos in the search corpus, no title compares these two directly. Technical analysis appears in 451 videos at a median of 8,006 views across 327 channels. Quantitative analysis appears in 1 video, at 1,456 views.
Four hundred and fifty-one videos against one. The interpretive tradition is among the largest subjects measured on this site and the discipline of testing has a single video in the entire corpus — which is worth holding in mind when judging either by how much material exists about it.
On the chart above the honest answer for most readers is no, and the pattern will feel exactly as convincing either way.
When it fails
The characteristic failure in technical analysis is confusing familiarity with evidence. Somebody who has looked at charts for years has seen a particular pattern resolve well many times, and remembers those occasions far better than the many where it did nothing — so a strong sense of reliability builds from a sample selected by memory rather than by counting. The feeling is genuine, the experience is real, and neither is evidence. Counting is the only remedy, and almost nobody does it because the impression is already so convincing.
A second failure in quantitative work is over-fitting, which produces a specification that describes the past precisely and predicts nothing.
A third is testing only on the data the rule was developed from, which is the same error stated more politely.
A fourth is applying a rule in conditions it was never specified for, where it will act confidently and without basis.
And a fifth is treating either as a complete method, since neither supplies position sizing or a risk rule.
Related
Technical analysis covers the interpretive tradition. Quantitative analysis covers stating and testing a rule. And price action covers the most judgement-heavy end of chart reading.
The uncomfortable part is that the human eye is excellent at finding structure and cannot tell the difference between structure that is there and structure that is not. Testing does not replace judgement — it tells you which judgements were worth having.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.