Loading ...Don’t Call it a Comeback – LL Cool J
Introduction
Ideas come to me in weird times and spots. This one during my morning routine a few days ago and in between articulating my thoughts on KD and dealing with the normal chaos that is the life of a family with a 13 month old, I managed to find some time to work out, in spreadsheet fashion of course, the particulars of an interesting system.
In particular, I was thinking about using first derivatives as indicators. A classical form is the venerable moving average, which whether in breakout mode or in mean-reversion form, probably represents 90% of all retail investor methodology.
But, what if one were to, in a moment of sheer goofiness, apply a moving average to a moving average? Would that tell anything?
Let’s find out.
The Setup
The instrument: The Devil’s own “FAS”
The first derivative: the 2 day moving average
The agenda: Take the 2 day moving average of the closing price as one column. Next to it, add the 2 day moving average of the first 2 day moving average. (Madness no?)
The signal: in the last 30 minutes of trading, examine the 2DMA against the 2DMA of the 2DMA. When the 2DMA prime is lower than the 2DMA of the 2DMA, buy.
If you are holding and the 2DMA prime is going to close higher than the 2DMA of the 2DMA, then sell.
The Results
- Trading Range: 11-19-08 through 11-13-09
- Number of trades: 44
- Win percentage: .68
- Average Trade: $1.00
- Average Win: $5.96
- Average Loss: $9.62
- Expectancy: $1.00
- Total Trade Earnings: $44.12 per share
- Two Tailed P Score (against 50/50 chance): 0.0159
Summary
While this is not the strongest system I’ve presented, it does offer some food for thought with regards to finding different ways towards tweaking the normal metrics used in equity analysis. I have a suspicion this system, or a variant thereof actually exists somewhere and am putting this out there additionally for any feedback.
If my instincts are right, someone here will be able to point me in the right direction.
Theme Song
Cuervos, this is just smoothing, no?
Take a look in dvindicators.com at the forum for the DV double stochastic. I think the code will be exactly what you are interested in.