Mousetrap 7/6/2011

This is the first report (on this site) of my "Mousetrap" portfolio. The methodology used in this portfolio is broad market timing agnostic. Instead, the portfolio examines volume and breadth divergences in 98 industry groups and selects fundamental stocks within technically selected industries.

The "Mousetrap" is at present a live beta test. The fundamentals have been backtested, and the technicals have been backtested. But it was impossible to backtest them both together. The target return is 30% or better. While not as good as the Robot, it could, if successful, offer a possible strategy hedge.

When fully loaded, the portfolio will rotate around 10 stocks. As of today it is up to five.

I'm being a guinea pig with my own money here:

On 5/31 I bought BKI at 25.11 -- now up 11.03%
On 6/22 I bought CFI at 8.29 -- now up 16.54%
On 6/27 I bought SE at 26.36 -- now up 4.84%
On 7/5 I bought AWR at 35.02 -- now up .03%
On 7/6 I bought CLH at 106.30 -- now up 1.89%

I'll try to keep this updated in as close to real time as possible (i.e. within a few minutes of any given trade). If it works in real time, we'll all find out together.

Tim

PS:

On the sector rotation model the long and short sectors (depending on your timing source) are...

XLK (technology) if long, or
short XLE (energy) if short

This is basically a stock screen and one that employs no hedges / timing, would that be fair to say?

Also, may I ask what software was used for backtesting, and if there was any survivorship bias in the data used for backtesting (it can be expensive to acquire survivorship-free stocks data)?

And, what are the liquidity thresholds used (i.e. does every stock have to be above say $5 and trade more than $X dollars in dollar volume per day)?

Thanks Tim.

Yes, this is a stock screen that employs no timing for the [I]broad [/I]market, but only [I]relative [/I]timing between industries.

It's basically an expansion of my sector rotation model, which I've tested back to 1977. Since that test was on whole sectors, there wasn't much potential for survivor bias interference.

The idea is to find industries or sectors that have the highest breadth&volume / price ratios relative to the others, and then to select individual stocks with a known fundamental screen.

The known screen is Greenblatt's 12 month trailing earnings yield and return on total capital. Although it can be argued that there are better screens, this is one that Greenblatt tested on the Compustat point in time database (which negates survivor bias) and is currently being tested in real time on Validea.com. I also tested it in real time for both long and short selections from April 2008 through April 2010. During that time I found that Value Line's "timeliness" rankings actually have NEGATIVE synergy with Greenblatt's so-called "Magic Formula." But what was most fascinating was the Greenblatt outperformed Value Line 1s during my test period.

Since I've live tested Greenblatt, and Validea is running a live test as well, there is no survivor bias problem.

I'm using Validea as a control set to examine the synergy between my technical screen and Greenblatt's fundamental screen.

My own technical screen for nine sectors backtested an average of 25% returns over the 1977-2011 testing period.

Greenblatt claimed 30% returns on Compustat, but Haugen reproduced the test and only got 20% returns. Validea is currently showing it outperforming the S&P by 10% a year.

Hence my target of 30%. 30% would show positive synergy between my technicals and that simple fundamental screen.

I don't have a blog or a site or anything, and I'm just a private investor. But I'm sharing this experiment here and in another forum I'm a member of to keep my real-time test "honest." Not "honest to others" so much as "honest to my own rules." I've had a bad habit of violating my own rules in the past, and by using a portion of my portfolio as a live and visible test online I can hopefully keep myself from violating my own strategy and perhaps prove a concept that might help a few others as well.

Tim

PS -- I'm also using a sector ETF version of this model in Billy's forum to see if his pivot points and my sector selections work together. Right now my XLU is woefully underperforming the IWM Robot...

On 5/31 I bought BKI at 25.11 -- now up 15.99%
On 6/22 I bought CFI at 8.29 -- now up 22.55%
On 6/27 I bought SE at 26.36 -- now up 5.56%
On 7/5 I bought AWR at 35.02 -- now up 1.20%
On 7/6 I bought CLH at 106.30 -- now up 1.42%

No new trades happened today. I'll run the macros again tonight to see it any setups occur.