The Problem — 2014 and 2015

Have you had the experience of writing a screen, and having it get fantastic performance — except for the last two years?  Like this one.
 
Zooming out we see outperformance, except in the last two years.  Maddening!

Back to the past

The way we’re going to tackle this is to zoom into a specific period, and see what it seemed to want. We’re going to limit ourselves to January 2014 to January 2015. This is going to tell us what the market wants right now. We can’t just go with this new plan, and throw out our old established recipes, because the market will just switch back, defeating our efforts. But we can learn something and enhance our screen with what we learn. The first thing we have to do is take a baseline.  We plot one chart with four things on it:

1. A backtest of everything — what happens if we buy equal weighted shares of each stock (that conforms to our default universe)
2. The results of buying our strategy
3. The results of buying everything that isn’t our strategy
4. The S&P 500

What does this tell us?  That our strategy still works, for one.  We handily beat the pants off buying everything.  But we’re still nowhere near that S&P 500 line.  Just to confirm that our strategy still works, we try two more things:  Test it on large caps, and test the inverse of our strategy, both by itself, and on the large cap stocks.  This of course adds the need for the large caps themselves to be on the chart, as a baseline.

From this we learn that large cap stocks seem to be the only game in town, and this strategy doesn’t work on large caps.  The inverse seems to outperform, frustratingly enough, until September, when it all falls apart.

Time to Shift Gears

Now that we know one of the problems, we will try something different:  figure out what works on large caps.  After an initial encouraging sign that increasing inventory is liked in 2014, we try ranking the increase in inventory, and taking the top 30%, as a proxy for growth.  We add checks to make sure that the expenses aren’t rising faster than sales, and that sales aren’t rising faster than cash flow, and we have something that comes close to matching the S&P 500.  We also now see that growth seems to work better on large caps (at least in 2014).

The yellow line is the one we’re looking at.  How to get it to actually beat the S&P 500?  After trying momentum, industry groups, and a few others, I am about to give up and decide that 2014 is a giant pile of beans.  Then I get mad and throw our four favorite scores at it:

Now we have something.  It won’t hold up under high trading costs, but at least it outperforms in 2014.  And 2013.  And 2010.  And 2009.  And 2008…  The large cap variant also beats in a large number of the years, particularly 2008.  Good enough for me.