Many traders use charts of past prices to trade stocks. They feel that, if the chart makes certain patterns, it can predict what the stock is likely to do in the near future.
One of the problems with chart reading is that pattern recognition is fairly subjective (the human brain likes to organize random data into pictures). One trader may see a pattern where another one doesn't. This makes charting-based trading systems hard to test.
However, studies that have been done using objective definitions of patterns consistently detect no trading advantage. They show that trading on chart patterns are equivalent to buying randomly.
I think that another flaw of chart based systems, which nobody really talks about, is scaling. Chart scaling, in my opinion, is the reason that patterns looks so seductive and accurate in hindsight.
Traders develop patterns by studying charts of big moves that have already occurred. They frequently "see" consolidations and patterns before the big moves.
For example, one pattern is called sideways consolidation - where the market trades almost horizontally in a narrow range. Then, prices eventually "breakout" and make a big up or down move.
The problem is that, after a market's range has expanded, the scaling of the chart changes as well. This scale change smooths out the movements that occurred prior to the big move, and creates the sideways consolidation. Prior to the big move, traders would not have seen the sideways consolidation, because the chart scale would have been different.
Let's assume that the market moved in a range between 10 and 20 for 6 months. The chart would be scaled from 10 to 20, and the chart will show lots of peaks and valleys. After the market moves to 85, the chart now reflects a range from 10 to 85. This compresses the peaks and valleys between 10 and 20. The result is that the 6 months before the move looks like a sideways consolidation pattern.
Unfortunately, you can't trade on patterns that occur after the move.
This is why my trading system does not use charts - only prices. Numbers are objective, and not open to interpretation. I only use charts to find prices for stocks I want to back test using my spreadsheet.
