Trading Systems

QuantShare is an advanced (multi-databases) trading software with powerful back-testing and simulation features, advanced charting and drawing capabilities, many plug-ins, a social network website and a sharing server where application users can discuss, create groups, review, vote and share trading objects with other users. Application objects include : Trading systems, scripts, composites items, list of symbols, money management scripts, list of rules, ranking systems, neural network prediction items, genetic algorithm optimization items, download scripts...

The software is currently in a beta-testing phase. Beta testing is a special development phase that helps us track bugs, errors and improve the application before its official release. We suggest that only experienced users take part in BETA testing. If you are interested in testing the software just enter your email address and we will send you further information. The beta version of the software, the access to the web 2.0 website and the sharing server are FREE.



Wednesday, December 28, 2005

70% winners and an annual gain of 850%, but...

I developed two years ago a simple trading system that showed great results in back-test.
This trading system showed 70% winners and an annual gain of 850%.

I was very happy and was thinking i found the ultimate system.
Everything was correct and i double checked everything, but when i traded this system, it was clear for me that i the system cannot be traded :(

So first let me explain rules of this trading system.Its was a long day trading system.
The rules was simple :
- 10% of capital for each position.
- buy a stock when it's 5% down from the open price.
- Stock must have more than 100 000$ average dollars volume per day.
- some others rules but i forget them.

The important thing to focus on is the second rule.
Do you think its correct ?
Of course yes, but the back-testing results was completely incorrect because of something that happen all the time to stocks that don't have enough liquidity (stocks with 100 000$ average dollars volume per day is not a liquid stock).

With an example it should be clear :

A stock open a 10$, then after 1 hour it is at 9.98$, one minute after that it drop to 9.4$ (-6%) for a second then it come back to 9.98$.
What happen here is due to lack of buyers at that moment, a single sell order make the stock drop quickly but when that happen quickly a new buyer come and make the stock up again.
Of course you will not have enough time to buy at 9.4$, even at 9.5$ so you will miss the trade.
The back-test don't miss that trade and show a performance of about 6%.

When in a simulation lot of theses kind of trades are picked, this will mislead the back-test results.