Technical Analysis Is a Risk Framework, Not a Crystal Ball

Mention charts at most Australian dinner parties and someone will roll their eyes. Fair enough. A lot of what gets sold as technical analysis is just lines drawn after the fact, with the pattern chosen to match the story the speaker already believes.

That does not mean price and volume contain zero information. It means most people use them badly.

The useful version

The CMT Association treats technical analysis as a way to identify trends, support, resistance, volatility and potential turning points from market-generated data. The practical point is narrower. Before you take any position—whether the idea came from a DCF, a macro view, or a tip from a mate—you still need answers to four questions: where do I get in, where am I wrong, how much capital is at risk, and where do I take profit if it works?

A chart can force those answers into writing while the position is still hypothetical. Support and resistance levels are not magic. They are places where buying or selling pressure has shown up before. A moving average is not a forecast. It is a mechanical way to define whether the recent path is still intact. The value is the pre-commitment, not the prediction.

Lo, Mamaysky and Wang replaced the subjective “I see a head and shoulders” with algorithmic pattern recognition across thousands of US stocks from 1962 to 1996. Some patterns carried incremental information above random noise. Brock, Lakonishok and LeBaron found that simple moving-average and trading-range break rules on the Dow from 1897 to 1986 produced higher returns on the buy signals than the sell signals in that long sample. Neither study says you can retire on trend lines. Both say the data are not pure noise.

Where it breaks

Illiquid names are the obvious failure. A micro-cap with $40,000 of daily volume is not showing broad supply and demand. It is showing one or two players. Drawing a flag on that chart is closer to reading tea leaves than to risk management.

Gaps and event risk are the second. A chart assumes continuous prices. Earnings, regulatory announcements and geopolitical shocks do not respect your stop. The level you wrote down may not be the level you get filled at.

Subjectivity is the third and most common. The same chart can look like a breakout or a failed breakout depending on what the viewer already owns. Lo, Mamaysky and Wang flagged this explicitly: geometric shapes are in the eye of the beholder. Confirmation bias loves a canvas.

Overfitting sits on top of all of it. A setup that “worked every time” in a backtest can be the one the tester unconsciously kept because it confirmed the story. Transaction costs, slippage and changing market structure quietly kill most of them.

Who it actually fits

Technical analysis adds the most when the decision horizon is days to months and the investor has the bandwidth to monitor and review. For multi-year holdings the fundamentals dominate. The method rewards people who can be wrong often, keep losses small, and treat the chart as a set of rules rather than a story. It punishes people who need the market to confirm their existing view.

If your patience horizon is measured in years and your bandwidth is under an hour a week, this is a satellite tool at best. Use it to define the exit on a position that already has a fundamental thesis. Do not use it as the thesis.

The myth both sides sell

Promoters sell patterns as forecasts: this formation means the stock goes to $X. It does not. It means the historical tendency, after costs, has been higher probability for one outcome than another in similar conditions. Those conditions may not hold this time.

Sceptics sell the opposite: charts contain no information. The long samples above do not support that either. Price and volume reflect what people are actually doing with capital. The question is whether the person reading them has the discipline to treat the output as a risk boundary rather than a prophecy.

What to do with this

Pick one position you already own. Write the price or condition that would prove the thesis wrong, the maximum dollars you will lose if that level is hit, and the condition under which you will review rather than hold. Do this before the next market open. If you cannot write it, the chart was never doing the job.