Sunday, April 01, 2007

Process Improvement

I continue to try to improve my processes. I learned that as a practicing systems engineer - no process can stand the test of time – there are always improvements that can be made. The problem is fining them.

Consequently I frequently make changes to my bread and butter filters to see what if anything can be done to make them more productive or more efficient where productivity is measured as win rate and efficiency is measured as ROI.

This morning I subjected BOB to several new approaches to see what, if anything, would happen if …

The if in this case was to see if adding a constraint regarding close being above or below a long-term moving average would improve BOB in any way whatsoever.

I made five test cases –

1. the Close is below EMA140
2. the Close is below MA200
3. the Close is above EMA140
4. the Close is above MA200
5. Original BOB

The results were interesting to say the least –

BOB the original outperformed the other test cases in a striking manner. But more importantly was this – the second place finisher was Close below MA200. In other words stocks that are considered to be taboo by the high priced stock selection site do better than the ones that are above MA200 – which they say is the only place to buy stocks.

At first I thought that was because they concentrate on the long term and I look at short term only. So I pulled the statistics on 1, 4, 10, 20 and 30-day net for a comparison.

And while BOB closing over EMA140 and MA200 is not bad at the 30-day level – only one beats original BOB and that is a close above EMA140.

So for now I'll stick with Original BOB and continue trying to improve it.


Dogwood said...

Marlyn, forgive me, but I'm too lazy to look for my answer on your blog right now.

The win percentage and ROI, is that from using BOB on weekly charts or daily charts?

I'm assuming weekly, but you know what they say about that....


Marlyn Trades said...

For this post - daily charts