As reported on HackerNoon, cryptocurrency investor Marc Howard has created a unique, but highly profitable, Bitcoin trading algorithm that uses “Sentiment Analysis.”
Sentiment Analysis, across all markets, is a tool to judge where the market is going. Many traders can get useful insights on the market’s “temperature” by talking with traders and seeing how they feel about a certain asset.
Now that it’s 2018, traders don’t have to talk to other traders to measure sentiment - they can go to the internet. Specifically, Google Trends.
Google Trends is a website that tracks how often a term is searched on Google. For example, here’s the Google Trends result for “bitcoin” for the past 12 months:
The chart above looks oddly similar to the Bitcoin price chart.
Using this knowledge, investor Marc Howard decided to see if he could create an algorithm that would trade based on changes in Google Trends.
The results are great: over 90 days, the algorithm produced a 29% return on investment. For a $100,000 model portfolio, that’s $28,839 in profit (minus trading fees).
Specifically, here’s how Marc designed his algorithm.
First, he searched Google Trends for “BTC USD” and “Buy Bitcoin” over 90 days. Both of these searches measure interest in Bitcoin, but when there’s more people interested in buying, Marc suggests the price will go up.
Next, Marc created a ratio of “BTC USD” to “Buy Bitcoin”. He found that if “Buy Bitcoin” divided by “BTC USD” was greater than 35%, more people were interested in buying. If it was lower, then buying interest was not rising.
To further refine the algorithm, Marc monitored daily price differentials.
Next I tested when the BTC price difference closes more than $80 above the prior day’s close price, this makes the pattern more consistent.
Marc began testing his algorithm, starting on July 6th, 2018. The results were strong. Even in a bear market, Marc was able to gain a 28% ROI. This hints that sentiment analysis could be a useful tool in playing the market.
Despite his early success, Marc admits there could be improvements. Marc’s buy/sell signals (35% ratio and $80 difference) were arbitrary numbers, and there were no tests performed to determine the best signals.
In addition, the market ranged from $6,000 to $8,000. These sentiment analysis tools might not work during a raging bull market, when everyone and their mom is looking to buy Bitcoin