Chart analysis can provide a context for experienced discretionary traders for coming up with subjective probabilities, which are then updated by proven, high expectation methods. Chart analysis alone is a negative expectation trading method because charts nowadays are available to everyone with access to the Internet.
System traders do not look at charts and they are not supposed to override their systems. The most successful traders of our time, known as HFT robots, do not even care about charts. Some discretionary technical traders usually look at charts because they provide a context for their trading method. Charts alone cannot provide a positive expectation trading method because everyone can see the same patterns everyone else sees and thus there is no edge. In the past, charts were available only to those traders who paid a fee to get them from those who prepared them manually, what was a very tedious job. The average trader did not have access to charts and thus some patterns on charts often provided some kind of an edge over those who did not have access to this information. With the advent of the personal computer and charting software in the 1980s, basic chart analysis started losing any edge it could offer and chartists, those who rely on classical chart analysis for their trading decisions, became slowly one of the most unprofitable group of traders. It is quite silly for someone to expect that money can be made from patterns seen by the whole world and the multitude of recent chart pattern failures in several key markets provides conclusive evidence that this is the case. I have explicitly described several of those failures in this blog. I will only mention a very recent one here that I think has caused tremendous pain to some retail traders who rely on charts and even huge losses to some professionals:
As it may be seen from the daily TBF (ProShares Short 20+ Year Treasury) chart above, the market made a bottom from November of 2011 thought March of this year and then broke above the neckline of the chart pattern. This pattern and other similar ones in TLT (inverted) and other bond related securities were heavily publicised as a market bottom (top). As it may be seen, this pattern went bust and bonds prices still continue to rally. This is what can happen to a naive chartist. On the contrary, all my algorithmic models were generating long signals for bonds during the time chartists were getting short to their neck. I warned them with this post.
As it turns out, some discretionary traders who rely solely on charts may be profitable for a year or two in a row. This is simply an artifact of the normal distribution. If you have a very large number of individuals who look at charts, some will have performance with mean return in the far right of the distribution by chance alone. However, to rule out luck, a very long performance record with positive results must be available. It is possible, as I have said before, that some have been able to find an edge from chart analysis alone; I do not rule out this possibility. However, as a group, the results are negative for chartists.
How to properly use chart analysis
Nowadays, classical chart patterns mean little and the high number of failures prove that. However, chart analysis can provide a context for establishing prior subjective probabilities that are subsequently updated by an objective high expectation algorithm. In simpler words, chart analysis can be used as a guide, a filter or provide a framework for money management. But chart analysis alone is not a trading method and it cannot be one because it is a highly subjective process that due to its proliferation cannot provide an edge.