From theory to practice - page 181
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
A TS with regular over-optimisation may well be tested on a three-month history, but a three-month history tests the validity of specific parameters, and the system itself with different parameters should also give consistent results over a longer period, so a few years is a reasonable requirement.
By the way. Some years ago I decided to resurrect my very old system from 3 years ago, which showed good results. It didn't work at all, not with any settings.
So does it make sense to test a system with old data, if you already know that it doesn't correspond to the realities of today's market? My guess is that it doesn't.
From my observations, the market (its behavior) changes significantly in about 2 years. Even the use of one year old data is already highly questionable. Imho.
Within a year, systems that have been fine-tuned for 3 months do not need to be reconfigured. Another question is who will give it to them to work without readjustment.)
Why about a year? Depends on the operating pattern. I, for example, have had it for about three years. The pattern that Alexander is unknowingly trying to exploit, I also think it will last for many years, until the structure of the currency market and the composition of its participants change (which is unlikely).
If only to see how the system survives Brexit, the SCB surprise in 2015 and other flush crashes. These are rare (but significant) events, they happen about once a year.
So Alexander's trades are rare. I don't think months will be long enough to accumulate enough statistics.
Why about a year? Depends on the operating pattern. I, for example, have had it for about three years. The pattern that Alexander is unknowingly trying to exploit, I also think will exist for many years until the structure of the currency market and its participants change (which is unlikely).
I was referring only to 3-month old systems. I am not talking about other systems in this post.
Here are the results for two weeks:
Not all trades are shown, of course.
On a demo account, but using quotes from a real NDD account.
In 2 weeks +100% to deposit.
This week I will test it on a Poisson flow of quotes. If everything is OK, I will start it on real account. If it's the same there - what else do I need?
And yes! I do not want to offend anyone with my posts! It's just a way of communicating. I sincerely feel for the people who like lions have fought with Forex, but fell in an uneven battle. I want to help them.
You will be surprised when you go to the real market and the quality of quotes is not the issue, but the reality is very different and the only proof of your rightness is a signal from a real account, of course with a positive dynamics of equity. Earnings in the market are formed not by single trades but by series of trades with a cut-off date of a week, month, etc. You will see that the real situation will be different from your expectations. Just because the investor institutions could not think of it, and you could ..... Don't you think it's odd?
A long time ago, there was a program called Chaos Hunter. Chaos Hunter. The result of the work of this program was exactly the above-mentioned market formula, which, as I understand it, people are trying to find. But unfortunately this formula does not exist, for many reasons, but the main one is that the FUTURE is NOT DEFINITELY possible! After all, if we are talking about a universal formula that can work for a long time, say a year or more, there is no guarantee that the formula will work the same way in the future......
The market exists in the moment, here and now, and history in the past has no bearing on the future......
If only to see how the system survives Brexit, the SCB surprise in 2015 and other flush crashes. These are rare (but significant) events, they happen about once a year.
So Alexander's trades are rare. I don't think months would be enough time to accumulate enough statistics.
I guess so.
I only deal with intraday, sometimes overnight.
My trade with Alexander is rare, but by duration (what I've seen) it is a typical intrade. All sorts of brexit, etc. intraday does not threaten it in any way).
In all likelihood, it is either over-sizing the trade or over-lossing. Both end up with a poker face over the long haul. That's why we need a test over several years.
Look at GBPUSD 24.06.2016 and EURCHF 15.01.2015 :)
In all likelihood, it is either overbidding the trade or overlapping losses. Both end up with a poker face over the long haul. That's why you need a test over several years.
I don't even touch the TS. It's not familiar with such concepts as overshooting. A model approximating the Wiener process with demolition in action. You yourself know that you can't think of a better one. The main thing is to correctly count variance and mean as a measure of central tendency. And an additional parameter is of course necessary. I am currently using asymmetry, but am gradually switching to non-gentropy.
In all likelihood, it is either overbidding the trade or overlapping losses. Both end up with a poker face over the long haul. That's why you need a multi-year test.
it all depends on the entry/exit signal
if there is such a signal and there is no manual intervention (start/stop program), then quite a decent result
the risk is of course overestimated, which the real does not like