Thursday, August 31, 2006

Market Teetering Between a Crash and a Rally

My indicators derived from the stochastic were in exactly the same pattern as they are now twice in 2005. Once in mid to late February and once in early November.

In the first case there was an upturn in the stochs from a high level and a small spike in the stochastics and a small but essentially failed rally. That is what I am betting on here. But in early November the upturn turned into a massive overbought rally.

So distinguishing between the two situations is of utmost importance :)

I'm still short but the market has been rising. In reaction I have dug up a tool I developed to help forecast the most bullish of my indicators. I had been skeptical that this tool was useful. But maybe it is useful in just such a situation. It says to be long at the moment. An alternative is if, for example, your standard position is 3000 shares, to allocate 1000 shares to each of the three indicators. So on a day like today where two indicators are short and one is long you would be theoretically short 2000 shares and long 1000 shares and, therefore, in practice short 1000 shares. I have read people discussing allocating their trading capital across multiple models and indicators. It could make sense. It is an agnostic approach to market direction.

I also found a distinction between February and November 2005. In November the fast %K stochastic was very strong - over 90 on the day that is the best analog for today. In February it was declining and around 80. If today ends up not too much it is going to look more like February and the case for the failed rally hypothesis will be strengthened.

No comments: