How are the percentages of winning trades computed?
When the computer finds a winning pattern, it does so based on testing historical data for a limited a collection of stocks, testing to see what trades would have been executed if every buy and sell signal from a pattern had been executed on the market. We then simply look at the number of those trades that would have been winning trades compared to the total number of trades, that is our percentage of winning trades. Determining the interval is a bit more complicated. Odds are properly expressed as some percentage plus or minus some amount, such as 60% 5%. That tells us that the true odds are actually somewhere between 55% and 65%. This is also based on confidence level. On this site we use a 95% confidence level. The formula for the interval is 1.644854 * √((percentage winners)(percentage losers)/(number of trades)) Thus a 60% success rate with 100 trades would have a confidence interval of 8.1%, whereas if there had been 1000 trades, the confidence interval would be 2.5%. Since ther