Results of a Four Month StockTwits Experiment

I’m a novice stock investor. My most in-depth analysis goes something like “I ordered twice as much stuff from Amazon this month. Let’s buy more $AMZN.”. But as a physicist, I love numbers, and StockTwits provides lots of them.

StockTwits is a social network that lets individual investors discuss stocks. Users can flag a post as being Bullish or Bearish with regards to a particular stock. If a sufficient number of people discuss a particular stock, it will have a “sentiment” option on the home page, giving the percentage of recent posts with a bullish sentiment for the stock.

Around November 7th, 2013, I did an experiment. I looked for six stocks with strong sentiments. I created a mock portfolio of the stocks in the table below.

The portfolio beat the S&P by almost 8%.

Some thoughts:

• Apparently there is something called “Beta”. (Sarcasm.) Each of these stocks, with the exception of $GDXJ, has a high beta. We would need a period of declining S&P in order to see if these gains are just a result of beta, or if there is truly influence from the StockTwits sentiment.
• Sentiment on StockTwits seems to be skewed largely positive. It would be great to see what the mean sentiment of all stocks is, so only stocks that are two standard deviations above the mean (or something like that) are chosen to buy (or, in the other direction, short). This data might be available via the pro API. Not sure.
• The experiment should have included periodic adjustments due to changes in investor sentiment. I didn’t do this. I got lazy.
• It’d be great to get a collection of historical sentiment data. I’d love to do an analysis of a few different algorithms over the last year.

Conclusion: StockTwits seems to be a great way for individual investors to benefit from the collective wisdom of the community. I’m going to start managing about 10% of my portfolio using StockTwits sentiment. Let’s see how that goes.

If anyone wants to float me a subscription to the StockTwits API, I’ll gladly nerd out on it for a while. ☺