Crypto Technical Analysis

“TA is for people who don't understand statistical significance or their own cognitive biases, and want to spend their life drawing lines on charts, guessing what TradingView indicators the other suckers are using, and very likely getting financially rekt in the process (but telling themselves otherwise). It represents the worst aspects of crypto and of humanity generally.”

-Dan McArdle, Messari co-founder

I literally laughed out loud when Dan sent me this. While I don’t fully agree with him, there are a few important truths in his sentiment towards technical analysis in crypto.

As a former quantitative trader, I know for a fact that TA can be profitable. But TA people who make money in the real world do not remotely resemble TA people on crypto Twitter/Telegram/Reddit.

The former understand three things that the latter don’t:

First, almost every successful trading strategy has a relatively small edge. TA is no exception. The best traders in the world generally have a win rate of less than 55%. Yet TA people in crypto seem to have 100% conviction in the predictions they make with their drawings. Overconfidence is the greatest recipe for financial ruin.

Second, every successful trading strategy is based on thorough analysis of history. Before the computerization of finance, successful traders spent years refining their strategies through a combination of live trading and post-trade analysis. With computerization, successful traders backtest their strategy over thousands or even millions of data points.

TA people in crypto don’t get that. They think spending a dozen hours studying classic price patterns or writing a blog post on a new metric make them trading experts.

Finally, price forecast is only half of any trading strategy. The other half is putting your money where your mouth is. How do you size your order? When do your close your position? Once again, the loudest TA people in crypto generally don’t do that. They make a price forecast, and never put real skin in the game.

Ironically, making real trades is the only way to learn the two hard lessons described above - overconfidence and statistical significance.

With those caveats, I do want to emphasize that TA can work in crypto, i.e., past prices could have predictive power on future prices. That’s because price actions (particularly in a retail / amateur driven-markets like crypto) can capture a lot of predictable mass psychology. Namely, market prices tend to trend over long time horizons.

If price recently rose over a horizon of a few months, then it is more likely than not to rise even more over the next few months. It’s a pattern we have observed for decades across nearly all asset classes Warren Buffett explained the phenomenon by observing what happens when “people see neighbors dumber than they are getting rich.” They FOMO, panic-buy, and push prices even higher.

On the other hand, when a long position falls in value, we are reluctant to cut losses. Selling at a loss is an admission of failure, yet those who aren’t willing to sell can get crushed by those who sell fast. Yet at some point, there is almost always a capitulation of the most emotional and weakest hands, who panic sell, and push prices to a floor.

I’m not encouraging you to study TA. I don’t think you should unless you are willing to put blood, sweat, and tears into studying TA for years. Achieving a win rate of 55% is exceptional, and nearly impossible to replicate.

But it is worth highlighting that TA isn’t all bad. Think about fundamental reasons behind certain price patterns. Why do all bubbles look so similar? Why do bubbles happen? These questions are more intellectually stimulating and could reveal insights into mass psychology that, yes, could prove lucrative.

But you’re not likely to find those insights on someone else’s TradingView tweet.

Qiao Wang

Interested in more insights like this? Subscribe to Unqualified Opinions.