Last week, I presented the state of author sentiment on Tesla Motors (TSLA) using strictly data from the Seeking Alpha domain. The premise was simple. I asked the question: can author sentiment at Seeking Alpha be used to identify entry points for long and short positions. Tesla Motors was the ideal candidate stock to analyze because there is a good balance of longs and shorts who publish on Seeking Alpha. In this article, I will present some interesting findings for a similar analysis on Apple Inc. (AAPL).
The first item that I want to present is the number of articles written at Seeking Alpha in the last 2 years or so. I have drawn a baseline at just over 3 articles per day for comparison. When the share price of Apple hit $600 in April 2012, the number of articles written about the company also exploded to 9 per day. It seemed like everyone had something to say about the delicious Apple pie. Strangely enough, the volume of articles vanished when Apple hit $700 in September 2012. Was Seeking Alpha clamping down on quantity to increase quality? Or was this an indication of the things to come?
In addition to the number of articles written, we can also monitor the number of comments in each article. There are two points that immediately jump out, and I've marked them with an asterisk. Incidentally, both occur at the base of the bowl in Apple share price. A caveat is that articles that are less than 1 month old are most likely still accumulating comments but older articles will tend to be more static in that regard.
From January 1 to December 11, 2013, 355 unique authors contributed a total of 924 Seeking Alpha articles. Of those, 396, 15, and 513 articles were published when the author held a long, short, or no position, respectively. However, the 15 articles wherein the author held a short position all belonged to a single author - Michael Blair. This brings me to the most important point that I want to hammer home: Seeking Alpha articles written on Apple are heavily biased to the buy side.
Having said that, there were many instances of authors who were long the stock but gave a hold recommendation. If we look at only the articles wherein the authors had no position in Apple, the buy/sell/hold recommendations were distributed as follows.
In order to dig deeper into this data, I scored each article in 2013 using this simple algorithm: +1 point if the author was long Apple or had a buy recommendation in the article (+2 if both), 0 points if the author was neutral on Apple, or -1 point if the author was short Apple or had a sell recommendation (-2 if both). For each publication, its score, as well as the scores from the previous 20 articles, were tallied up. Here is what it looks like. Thus we get a running average of 21 articles.
As I have pointed out earlier, there is a clear bias toward Apple ownership in the author pool; thus, the author sentiment score always ended up being a positive number. What we need to focus on are the extremes in positive sentiment, which I have marked using green and red lines. If we had bought and sold AAPL shares based on this information, we would have made money 8 times (marked by "*") and lost money 5 times (marked by "!") within a one-month time window. Now, I am in no way suggesting that you should trade AAPL using this simple methodology. It is as reliable as any other technical indicators. Also, this is a relatively small sample size, and it would be interesting to back-test this method all the way to 2011 to include the rise to, and fall from, $700.
There will be doubters who will immediately discard all this information and tell me that I have wasted 5 minutes of their time or a single mouse click. Yet, there will be others who will find this information intriguing, and maybe even helpful. I believe that at Seeking Alpha, we, as authors, have a mandate to try and keep one step ahead of the professionals who do this for a living. I hope this analysis was one step in that direction.