Just on looking at charts around signals, yes agree that's not enough to test a system but it may give some insight into how price behaves post a signal that may lead to further ideas to backtest e.g. IF entry is triggered THEN do this. I don't know if that is a legitimate way to develop a system but an idea that comes to mind. I am not yet familiar with curve fitting and how complex or simple to keep a system and how that interplay works. Something that I look forward to exploring.
In terms of your questions, some thoughts:
- What markets do you want to focuson? Ultimately all equity markets, but to begin with ASX.
- What timeframes do you want to test? EOD initially.
- What ideas will you test? Happy to start with a MRS such as Marwood's, also interested in trend following.
- What kind of tests will you run? Backtest with optimisation.
- What parameters will you test? Follow other people's tests to begin with.
- What metrics will you use to measuresuccess? CAR/MDD, equity curve etc.
- When will you give up on an idea? If the optimisation leads to values for parameters that have poor results either side or the MDD is say >~20%.
I've decided on Amibroker, now just to check standard v pro and also whether to get the AFL wizard or not. Looking at data options, would prefer one with delisted stocks included, survivorship bias but also may just back test incorrectly since trades could have been taken on those stocks.
Thanks for the offer to assist if I hit a road bump on install etc! Look forward to testing soon - hope to be up and running by next week.
(20min delay)
