Friday, March 02, 2007

January performance part II

[I started this piece last Friday, before the meltdown - so there has been a little 'updating' since...]

For the next step of my analysis I took a look at some of the technical signals and rated their performance to my earlier measures of 1%, 2%, 5%, and 8% risk - looking for a 3:1 reward to risk return.

The first figure gives the overall return (based on average return per trade, for all trades combined) for my free and paid-for picks from my newsletter (I should add, this does not include the stocks from the Trade Ideas scan listed on my blog - that will be a seperate article). The post-meltdown figures look like this:

The biggest change came from the performance drop in the use of 5% and 8% risk as those stops were most vulnerable to the weeks selling, whereas the lower risk sample either had reached their targets, or were stopped out prior to Tuesday. There was no change to the 1% risk, and minimal change to the 2% risk. When the figures are broken down to stocks which were free picks and those which were exclusive to members, the following returns were reported:

I then looked at the performance of the MACD trigger 'buy'. Twenty-six (26) percent of the stocks drawn from my scans had triggered a 'buy' in this indicator. The performance of those stocks which had a MACD trigger 'buy' outperformed those that did not with the exception of the 8% risk

Next I looked at a breakout in on-balance-volume. This was categorized by either a significant trendline or resistance break, and/or a new 6-month high for the indicator. Forty-six (46) percent of stocks enjoyed such a move. The returns were again greater for stocks which showed new highs in on-balance-volume versus those which did not. The 8% risk was the only group not to show an improvement:

The next group of stocks were for picks which had a trend measure strength (either up or down) of ADX > 20. Fifty-two (52) percent of my picks had a well defined trend in place at the time of their feature. There was clear performance advantage from stocks in an established trend as measured by ADX:

Finally, I looked at different combinations of indicators and the average return per trade for winners and losers combined. I excluded the 8% risk due to the low number of completed trades. The only thing I will say for it is the poor performance of the picks when none of the indicators showed a signal (i.e. a plain-Jane price+volume breakout); the sample size for this group was around 20 stocks per risk factor so it is a relatively accurate measurement. Data for ADX alone and ADX in combination with OBV also have a sample size above 10. The remaining categories had low numbers of completed trades so I wouldn't rate it as a valid sample size - but trends are apparent.