Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
Just wondering ... suppose for the sake of argument that I've found a great strategy via genetic optimization. Since this strategy was found using backtest (historical) data, it is subject the same issues any backtested strategy is when run in real time, nes pas ? So I've got this optimized strategy, and now I've got to throw it into another environment (i.e. real time), that may or may not be anything like where it was optimized.
Does optimization really work ?
Can you help answer these questions from other members on NexusFi?
Everything is subjective. There is no short cut. There is no black or white answer. Yes optimization works. No optimization doesn't work. And everything in between.
If you have a strategy that has backtested well, the best thing you can do to learn from this is to test it in a live market (on sim) for several weeks or months. However, most people will end up making constant tweaks (curve fitting) when a trade goes badly, to avoid such a "mistake" in the future, which is worthless for the most part.
The enemy of backtesting is over curve fitting. Whereas the strength of not backtesting is to establish some basic principles or methodologies. You just need to fully understand out of sample and in sample testing, otherwise all of your testing will be for naught.
Weakness of backtesting = over curve fitting, lookback only
Strength of backtesting = establish basic principles, methodology
The only real, true, believable results from testing is forward testing. Your idea must be forward tested with out-of-sample data. Once a strategy has seen data (date range), it cannot be altered or modified (parameters changed, re-optimized) on the same data or it will be tainted -- out of sample data becomes in sample data.
Backtesting is highly dependent on the data series and indicators, if any, employed. Without knowledge of what you have in this regard, it's not helpful to offer an opinion, but Mike is spot on about curve fitting.