You, the reader, have read countless Seeking Alpha articles. When you read an article of interest, you are most likely prone to confirmation bias. That is, if you already hold a long position in Tesla Motors (NASDAQ:TSLA), Apple Inc. (NASDAQ:AAPL), or Amazon.com Inc. (NASDAQ:AMZN) you will tend to agree with articles which recommend you to buy these stocks but refute articles that give sell recommendations. One important facet of self-directed investing is to be able to objectively judge which articles speak the "truth", especially regarding company performance and future stock price.
Another important aspect of research is to determine whether the source of information is reliable. For instance, since you are reading this article, do you know how reliable the Seeking Alpha articles are? There are authors who truly know how to research and how to present their theses. Yet, on the other side of the coin, there are contributors who write controversial pieces just to earn a pretty penny. In the research presented here, I have compiled all the Seeking Alpha articles on Tesla Motors that were published in 2013, up to November 30th. Can we use some creative analytics to predict the future direction of TSLA share price?
In the first eleven months of 2013, there were 356 articles written on Tesla Motors. The first item that I examined was the long/short position of the authors who wrote the articles. Notice that prior to the release of 1Q 2013 results on May 8, authors were overwhelmingly long on Tesla. However, after this date, as we progressed to November, the number of articles written by authors who were short grew to a larger number (49) than those written by longs (30).
What about authors who did not have positions? In total, 230 articles (64.6%) were written by contributors who did not have a position in TSLA. Below is the graph showing the buy/sell/hold recommendations from these articles. A few points jump out. First, the 1Q 2013 event once again appears to be an inflection point that was populated by a cluster of sell recommendations. Second, there was a strikingly long period in June that lacked sell-side recommendations from authors who did not hold a position in TSLA. This was followed by a steady rise in share price. Third, a similarly long period in September lacked buy recommendations from authors who did not hold a position in TSLA, and this was followed by the gradual decline in share price to the present level. Remember, this figure omits data from authors who are long or short.
Unfortunately, this information is static and lacks obvious predictive powers. Thus I proceeded to examine a "window" of consecutive articles to gauge changes in author sentiment. Window sizes of 7 (target article + three articles before and after), 15 (target article + 7 before and after), and 21 (target article + 10 before and after) were examined. To calculate the "long" score, articles that were written by authors with a long position were given 2 points, while those with buy recommendations written by financially objective authors were given 1 point. The opinions of those who put money where their mouths are should be worth more. The sums were then divided by the window sizes. Conversely, the "short" scores were calculated by assigning 2 points for short positions and 1 point for sell recommendations. Shown below are the analyses using the three window sizes representing short-, medium, and long-term author sentiment.
In all three cases, especially the windows size of 21 articles, the divergence between the "long" and "short" scores gave extremely accurate predictions of TSLA share price. In the earlier part of this year, long sentiment dominated publications on Seeking Alpha. The stock went from $35 to almost $200. However, starting in July, short sentiment dominated Tesla articles on Seeking Alpha, and we witnessed popping of the TSLA bubble shortly thereafter.
If this is a good indication of where TSLA share price, then I have bad news for TSLA longs. At present, the "short" score is at its highest positive divergence versus the "long" score, indicating that additional correction to the downside is coming.
As a biochemist, the Tesla story is an intriguing one. In biology, if you put a single living cell in a closed system with limited resources, the culture will undergo 4 phases: lag, exponential growth, stationary, and death. The cells will die unless additional nutrients are added to the system. When I look at the year-to-date chart of Tesla Motors, I see the exact replica of these four phases. Its share price is in death phase right now and only a fundamental stimulus will bring it back to health. Lacking this, we will see further depreciation of TSLA share price.
I am also thinking about doing the same analysis for the mother of all stocks, AAPL. If you think this was a worthwhile exercise, send me a comment below.