08/24/2014 How the Model works... and sometimes doesn't...

Style Model	Small Value
Sector Model	XLI	0.00%
Large Portfolio	Date	Return	Days
TIVO	4/23/2014	13.50%	122
SHOO	4/28/2014	-1.46%	117
SR	6/2/2014	13.92%	82
CFI	6/9/2014	1.06%	75
FRAN	6/16/2014	-4.69%	68
RRD	7/21/2014	7.77%	33
CHFC	7/28/2014	2.28%	26
ESI	8/4/2014	-37.88%	19
BSET	8/11/2014	-1.75%	12
STRA	8/18/2014	1.65%	5
(Since 5/31/2011)
S&P	Annualized	12.86%
Sector Model	Annualized	25.89%
Large Portfolio	Annualized	23.92%

 

Rotation: selling FRAN; buying PBI.

The Sector Model had a good week, in spite of the XLF miss during an XLI whipsaw (you’ll note that it has returned to XLI).

As I noted on the blog, Gold Coast Advisors' Sector Fund actually MISSED the whipsaw because it was signaled so close to the end of trading, and picked up about 1.2% that the model itself missed.

Since I only report the model’s behavior, I am not showing that 1.2% gain on this chart:

 

I trade my own funds in parallel to Gold Coast’s trades, and so my own live returns are reflected here.

Not a bad chart, even with the miss.

Just a quick note on how these models work:

Breadth, volume, and price normally go together.  Momentum strategies invest when all three agree.  Mean reversion strategies invest when they do not agree.

As I said, I have a mean reversion strategy.  When the breadth and volume in a sector are outperforming price, I invest in the sector with the greatest long term divergence.

I’d try to make it sound more complex than that, and I’m always tinkering under the hood – but that disconnect is the heart and soul of the model. Majority vote between breadth, volume, and price, usually wins.  Occasionally price is right and breadth and volume collapse, but it works out in my favor more often than not.

Since I don’t follow momentum, I will NOT be in the greatest performing sector at any given time.  I just need to outperform the median for the market as a whole.  A black box algorithm may outperform by 30-40%.  I’m content with 20%.

I’ve tried to get greedy with the Full Model, to juice the returns through fundamental selection.  Most of my experiments have been abject failures – but 24% returns for the Full Model is the kind of “failure” that leaves you still kicking to experiment another day.

Interestingly, the goal has always been to outperform the Sector Model, which is the benchmark for the Full Model.

And to that goal I continue to work.

But the Sector Model – the technical core – keeps the Full Model afloat until I can find that sweet-spot between Technical and Fundamental indicators that allows them to work together, instead of against each other.

As for the trades, I’ll set limit orders over the weekend and let the trades strike at some point during the week.  For a sell, I’ll add 0.2% to Friday’s close.  For the buy, I’ll add 0.1%.

If there is a large shift in the futures, I’ll add the futures to Friday’s close and recalculate the limits accordingly.  So a 0.7% gain in the futures at 6am on Monday would leave me with Friday + 0.9% for the sell and Friday +0.8% for the buy.

Then I walk away.

The market would drive me crazy if I didn’t have the model to calculate for me.

In short:

Don’t get greedy, and don’t obsess.  Those black boxes out there are designed to play with your brain.  Ignore them.

Live!

Leave the black boxes to fight each other, and the leave psychiatric meds for the day traders.

Tim