My technique is based upon observation and simple data analysis. Look for underlying concepts that may be valid and test a range of ideas/parameters based on those concepts, and see if there is any consistency. You can get ideas from books and articles, however one of the best techniques is just to record observations at the end of every trading day, and then examine them with data. If you do this every day for 6 months, you will likely start to discover patterns. You can then plan each day according to the analysis of where your edge is in that market for that day under different scenarios. If nothing is there, don't trade. This is the approach that has worked for me.