I’m not sure where Id even begin with developing discretionary strategies. I started going down the mechanical route as I have a maths, stats, programming type of background, but I failed, and inevitably ended up as a discretionary trader.
I dont know what to suggest other than screen time, and you’ll probably need a couple of years full time.
These days, Im back to trading purely mechanical strategies. The development methodology that I use, is actually my edge so its not something Id want to disclose. However if your interested in this route, Van Tharps stuff isnt a bad introduction. His designing winning systems course is OK, its not fantastic, and it lacks practical examples, and its definately short on details, but its a start. He also proposes a simple methodology in his appallingly titled book “trade your way to financial freedom”. As the book and the course are available at torrent sites etc it might be worth taking a look to see if they suit you.
It always helps to see a couple of examples of how people go about this process. There’s a 2 part series at trade 2 win of the development of a break out strategy (its a crap system, but the design process that they follow is OK), there’s also a nice article at elite trader where some guy develops a trend following system based on trade station thats pretty good. I dont have the links, but if your struggling to find them I could probably dig them out. The booklet on the original turtles system is OK in that it considers most of the main elements that make up the design of system, market selection, entry set ups and triggers, exit, position sizing, trade management etc.
If you can get hold of a copy of John Murphys book Technical Analysis of the Financial Markets there’s an appendix in the back by a guy named Fred Schutzman who does a pretty good job of describing a basic development methodology.
There are a few decent PhD thesis knocking about where development methodology is discussed, a good starting point is a thesis by a guy named Bruce Vanstone who used to post at the wealthlabs forum. The thesis concerns the development of a system based on ANN, and most of the references pertain to an artificial intelligence approach, but there are one or two gems hidden within.
There are no shortcuts, its watching charts in real time, or its stepping through charts by by bar and its more work than you can possibly imagine. Believe me, there are far easier ways of getting rich, but if you find this stuff fun, go for it.
On a positive note, in my opinion you’ve identified exactly what you need to do, and that improves your chances of success enormously. Deciding that you’ll take responsibility to design and impliment your own method is a giant leap forward and one that I’m sure you wont regret.