Here’s what we did while delaying our summer vacation: eight stock picks that reached the sweet spot of their targeted hold duration with an average 20% return. We needed that vacation, but twenty percent in six weeks is a nice start.
Very nice – but of course far from the whole story. So we’re veering away from big data and investing issues in this post to focus on “small data” – those oh-so-scrutinized data points that arrive immediately after we flip the switch to initiate a test.
On July 1, we began processing equity-related data and promised to share some of our “picks” real-time. (We post via our account at StockTwits because it’s free for anyone to access and because we are unable to modify or delete posts.) Since then, we’ve shared 47 positions (not counting the 8 we shared earlier today) that had a forecasted holding period of 1-2 months. We thought it might be useful to share some of our early impressions.
First, are the results in line with what we expected? Yes and no – these results are technically within bounds but far surpassed our target, and we know we won’t do this well very often. After the first 32 trading days (1.5 months), all eight positions had positive returns, six of the eight beat SPY, and five produced double-digit returns (long LNKD, WLT and YELP and short JCP and SRPT).
Second, are there any apparent weaknesses that we should begin to analyze now (that is, issues that are significant and unlikely to be anomalous)? Take a look at this mess of spaghetti that depicts trading returns for our long and short positions:
Clearly longs outperformed shorts (by 600 bps at the 21 day mark), and most of the positions were slow starters – again particularly the shorts. So we’ll be working on that part of the model and anticipate either fewer shorts or, possibly, a mix of shorts that are targeted for one-week positions along with the existing one to two-month positions. (We will, of course, refer to these shorts held for less time as . . . “briefs.” Cue the rimshot, please. Thank you very much.)
Lastly, are there any implementation surprises? Again, it’s early, but we think there are – and one of them is staring at us in that pair of graphs. When we checked with our affiliate, they had captured far less than what we hoped for based on these numbers. Some of that was due to position sizing as we gain experience and to portfolio management as we accumulate signals. But we could have given them better guidance for profit-taking; volatility on the shorts was high and many stop limits were triggered too early, so we’ll be working to help clarify appropriate risk/reward guidelines along with enhancing the short-term performance of the model. Hmmm. Guess that summer vacation isn’t happening….