A Framework for Sentiment Analysis

Sentiment analysis is one of the most important tools available to a technical trader. However, it seems that some traders, especially newer ones, have trouble understanding how to properly utilize sentiment analysis.

There are three primary ways an investor can use interpret sentiment:

1: To reveal what positions investors have already taken

2: To determine what positions undecided investors might take

3: To indicate a possible expansion of the investor pool

Method 1

One common mistake new traders make is they assume that an asset with very positive sentiment is likely to rise. The mistake is understandable; it would make sense that if an asset has a large number of enthusiastic supporters, there should be a lot of buying pressure. However, in reality, the opposite is usually true. Raw sentiment analysis is almost always a lagging indicator rather than a leading indicator. If someone is taking a bullish or bearish stance publicly, it is likely that they have already traded accordingly. Why would someone have such strong opinions about an asset they aren’t invested in? Additionally, those with high Twitter followings have the potential to move altcoin markets with well-timed tweets.¹ Why would they let other investors front-run them?

Investor confusion around the proper use of sentiment analysis has been used by bad actors for decades to manipulate markets. One of the most famous crypto-related incidents was in 2014, when a mysterious trader known as Wolong used this principle to orchestrate a pump-and-dump scheme in the Dogecoin market.²

[1] https://motherboard.vice.com/en_us/article/9knnpz/john-mcafee-twitter-coin-of-the-day-cryptocurrency-markets

[2] https://steemit.com/bitcoin/@joseph/wolong-the-game-of-deception-unedited-version

For example, see the tweet above. It’s extremely likely that the tweeter was already heavily invested in XRP before they posted that tweet. Additionally, it’s likely that many of those who engaged with that tweet (through likes, retweets, and replies) were also already significantly invested in XRP.

Does this mean that sentiment should be used strictly as a counter-indicator? Not always; using sentiment that way could cause you to trade against the trend. The next two methods discuss ways sentiment can be used to correctly predict trends.

Method 2

For any asset, there are three pools of traders:

  • Those who are bullish and already invested
  • Those who are bearish and already short (or not invested if short selling isn’t an option)
  • Those who are willing to enter a position but are unsure which direction to trade

Like in politics, the goal of each side is to persuade undecided voters/traders to join their side. Sentiment analysis could indicate which way these undecided traders could swing.

A good way to gauge whether this method is effective is first to assess how large the pool of undecided traders is. This pool is usually pretty small for large-cap altcoins, but it could be quite large for less well-known altcoins. Additionally, coins with significant emotional baggage (such as Ripple and Bitcoin Cash) are likely to have a smaller pool of undecided traders, and the persistently active social media chatter around those coins could make sentiment analysis harder.

Once the pool of undecided traders has been estimated, keeping an eye on tweeting trends, as shown above, is one way of gauging how likely an undecided trader is to swing one way or another. Often times, the most active, and vocal communities have the most success convincing undecided traders.

Method 3

In the previous section, we assumed that the pool of traders for a given asset is fixed. In real life, the size of this pool grows and shrinks over time. Analyzing public discourse and sentiment can indicate whether or not the pool of investors is likely to grow or shrink. This is extremely important to keep an eye on, because a large influx of new traders can provide significant buying power, which fuels price increases.

One prime example of this is when CNBC started covering cryptocurrencies in early August 2017. Until then, cryptocurrency always had a very unique investor profile. However, CNBC’s coverage spread knowledge and interest in crypto to a much broader, finance-oriented audience. In the months that followed, Bitcoin’s volume increased significantly, fueling a record-breaking rally to $20,000 per coin later that year.

Closing Thoughts

Although it might be tricky to balance these three methods, looking also at price action and fundamentals can help you determine which of the three methods is applicable. For example, Twitter chatter is usually only useful for determining what positions people already hold. However, higher levels of chatter overall could indicate an expansion of the investor pool. At the end of the day, there’s no one correct way to utilize sentiment analysis, and each investor has to figure out which method works best for them personally.

Ex-cryptocurrency trader fascinated with financial markets and the future of money.