Developing trading methodologies - phase two

Some of the old RA short-term trading models have been dusted down and programmed into an online charting programme that I have been looking at. So raw data from these systems is beginning to spill out. I have worked up a list of relevant statistics that I need to record and am slowly collecting this as well - unfortunately this has to be done manually.

As always, the possible approaches that could be used is virtually unlimited and so the selection process is going to be critical. My main theme is that prediction is best practiced over very short time periods and so the early models are being applied to data based on 5, 10 and 15 minutes. So lots of trades, small point wins and losses, but a rapid build up of statistics.

Lots of serious research questions beginning to form - mainly the sort of statistical question that I developed after reading two books by John Sweeney years ago. In the end, I will build up a collection of programmes which I will apply simultaneously. Each will work ok on its own, but will blend together better than any on its own. They will require substantial attention each day, and I have lots to do to work out the detailed day-to-day working schedule. But early results remain very encouraging.