While Al is not actively trading and the thread has had some unproductive debate, I thought it would be a good time to briefly go over some of the ideas behind Al:
The info in this post has in the main already been covered in the first 30 posts of this thread.
I have been bearish on the economy throughout 2001 and 2002. In the last few months of 2001 we had a considerable rally, while I was holding puts. At the end of the year, I was down sizeably on these puts and realised I needed to change tactics. I had done pretty well in predicting long term tech trends using fundamental analysis but timing was not great although I did OK from WEBX puts I was holding - that was lucky!
It was becoming clear that F/A worked in the very long term but not necessarily in the short or medium term. I had traded from T/A signals for some time before creating AL, but it was a "by eye" approach - using techmiques that appeared to work after looking at just a handful of examples.
Al was created as an Excel spreadsheet containing price data for a number of stocks, and some tests along with the results. So for example, a couple of simple rules might be:
1) A stock at its lower Bollinger band is likely to go up
2) An index closing at the day's high is likely to go down the next day
These rules can be tested by the spreadsheet fairly quickly. Of course, that is not enough to fully define a trade. A stock bought at the lower BB might go up 10c, then what? Do you sell there or do you hold out for 20c, 50c, a dollar? What if you hold out for more and it goes down? If you try to answer these questions on the fly, all sorts of emotional factors come in.
The idea I work on is to define the exit criteria in advance. Al started with a simple criteria that "seemed" right in my own mind, tested it and tested various adjustments either side, then reiterated until a superior profit system was developed.
The initial simple rules, clearly described here on this thread, involved a stop loss of 4%, and taking gains of 10%. Otherwise, the trades were closed out after 10 days regardless of results.
As Al gained an impressive profit from such simplicity, he was developed in various directions away from the simple short term trades into swing trades, whilst retaining and adjusting the daytrading rules. One of his favorite indicators is $VXN. This works incredibly well and Al's worst timed trades have been thos that paid the least regard to $VXN. A lesson there. Look how high $VXN was when Al made his current batch of unprofitable trades.
Al finds QQQ a very difficult stock to trade, and because of this concentrates on trading it. Trying to master something difficult is a good way of developing abilities.
One of the most important discoveries Al has made is that although overbought and oversold conditions are easy to spot, they can lead to an exaggeration of the condition. This can work against simple rules.
For example, a stock at the lower BB often means immediate upside is to come. However, when this indicator fails, it can lead to an explosive move downwards. A move downward of more than a range of 4% - 5% dependent on volatility of the specific stock has been found to be the point at which the BB indicator "fails" and the price starts to dictate the BB more than the BB bounds the price. This has been discussed on this thread before in terms of how Al's usual 4% - 5% stops are determined. When this particular failure happens, Al doensn't trade the stock again until it has returned to the middle BB.
While Al is tied up with his current trades, he is working on the following:
Experimenting with the use of very close stops that are less than 6c from the execution price for very short term trades.
Increasing the focus on support and resistance levels (esp intraday), both historic and predicted from non-horizontal trendlines.
Further analysis of $VXN.
Working on a dollar and time weighted put/call ratio.
Experimenting with clues from bid and ask sizes. |
