You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
"Universality" of your trading system?
- Thread starter gmailer
- Start date
Quote from jcl:
I'm not this familiar with Metastock, but even if it does not allow you to modify price data, can't you just replace f.i. h with 1/l, and l with 1/h in your strategy script, and inverse the trade direction?
Are you sure this would work? If EUR/USD H=1.30 then 1/h would be 0.76...no chance EUR/USD could cross such level. Do not hesitate to correct me if I am missing something.
Hmm, when a price c crosses a level h, then a price 1/c will cross a level 1/h at exactly the same bar, just in the opposite direction. At least this seems to me plain math.Quote from gmailer:
Are you sure this would work? If EUR/USD H=1.30 then 1/h would be 0.76...no chance EUR/USD could cross such level. Do not hesitate to correct me if I am missing something.
I don't know how your strategy works, but my suggestion to test the inverse price curve was meant for a quick test in 5 minutes. If it is very complicated to do, due to some shortcomings of Metastock, it's probably not worth the hassle. If your strategy is 100% symmetric and only compares prices, without filter functions, fixed thresholds or the like, the inverse price test will anyway generate the same winning and losing trades. In that case the test is meaningless and you know not more than before.
Quote from jcl:
Hmm, when a price c crosses a level h, then a price 1/c will cross a level 1/h at exactly the same bar, just in the opposite direction. At least this seems to me plain math.
I don't know how your strategy works, but my suggestion to test the inverse price curve was meant for a quick test, such as done in 5 minutes. If it is very complicated to do, due to some very complex strategy or some shortcomings of Metastock, it's probably not worth the hassle. If your strategy only compares prices at bars, and has no filter functions, averages or something like that, the inverse price curve will anyway generate exactly the same winning and losing trades. In that case the inverse price test would be meaningless.
The strategy you're describing cannot possibly be profitable long term, but the nature of it is to reflect the inverse.
Only one of my strategies will have this property, but what you've described is only relevant to forex and not anything you can test on futures.
OK guys. I have made additional tests with my system on other pairs than EUR/USD. I never do it but I used optimization and with specific "periods" setting my system did bring on historical data good results on any currency pair.
So I want to make things more clear and ask again, summarising everything.
Should system (that has exact periods settings) work on many markets if "periods" is fixed? Example:
ENTER when C is above HHV(100) - works good on EUR/USD
on GBP/USD it sucks until I change periods to 50(instead of 100)
Can such system be any good? Or does it have to work on many markets under the same conditions/period settings("100" for every market)?
So I want to make things more clear and ask again, summarising everything.
Should system (that has exact periods settings) work on many markets if "periods" is fixed? Example:
ENTER when C is above HHV(100) - works good on EUR/USD
on GBP/USD it sucks until I change periods to 50(instead of 100)
Can such system be any good? Or does it have to work on many markets under the same conditions/period settings("100" for every market)?
If you using a system that works only on some markets and only recently, that means you exploit inefficiency that appeared on those markets recently. It's up to you if you want to put your money on that, but I would never do it, as there is high chance that the inefficiency will disappear (assuming system is not just a curve fit to noise in the first place).
On the other hand, it's natural that some markets (let say illiquid penny stocks) don't give profit with your system. The rule I use is that 70% markets must be profitable on 10 years of data, 80% on 20 years, 85% on 30 years and so on. Don't expect to get system profitable on absolutely every possible market in the world (including Bangladeshi stocks).
That applies to one system - no parameters' changes.
On the other hand, it's natural that some markets (let say illiquid penny stocks) don't give profit with your system. The rule I use is that 70% markets must be profitable on 10 years of data, 80% on 20 years, 85% on 30 years and so on. Don't expect to get system profitable on absolutely every possible market in the world (including Bangladeshi stocks).
That applies to one system - no parameters' changes.
Quote from luckyputanski:
The rule I use is that 70% markets must be profitable on 10 years of data, 80% on 20 years, 85% on 30 years and so on. Don't expect to get system profitable on absolutely every possible market in the world (including Bangladeshi stocks).
That applies to one system - no parameters' changes.
I understand...so you have system with fixed parameters and it should work on 70% of markets?
1. May I know what markets you test your system on? Are we talking about currencies too? Does your system work on EUR/USD and GBP/USD at the same time?
2. Also - when you say 10 or 20 years what time frame you are thinking of as far as charts are concerned? Daily charts or other? I test my systems on 1h charts, 8y period.
I haven't been trading forex that long, but I can tell you what ever works on anything won't work on GBP. I don't care if you are automated or trading by the seat of your pants.Quote from gmailer:
Hello guys,
How do you deal with "universality" of your systems? Do you think system should be good at all markets(or at least as many as possible) or can it be good only at one?
I have few trading systems designed for forex. But I have following problem - some of them look great on EUR/USD but completely suck on GBP/USD.
Should I worry about that? My intuition tells me that system should work good on many markets - this means it is not "tailored" for one market/curve. But my experience shows me that my systems tend to look different on different pairs. I use 5000 periods - intra.
What is your opinion on that? Is it normal that system works in different way on different pairs? Or should I worry that system is not as good as I thought?
I think there's a guy in London who's job it is to trade GBP, and if it's too high he sells and it it's too low he buys.
He just works for the CB part time, his full time job is at the art museum. Bean, I think is his name.
Quote from gmailer:
I understand...so you have system with fixed parameters and it should work on 70% of markets?
1. May I know what markets you test your system on? Are we talking about currencies too? Does your system work on EUR/USD and GBP/USD at the same time?
2. Also - when you say 10 or 20 years what time frame you are thinking of as far as charts are concerned? Daily charts or other? I test my systems on 1h charts, 8y period.
Ad1. Indexes, currencies, commodities, bonds. It also work on most stocks, but I don't trade them.
Ad2. Daily