Backtesting opinion - page 4

New comment
 
doshur:
Curve fit current period. Predict future change is not drastic but slowly change
That's right, maybe for the most of the time, but the problem are the black swans and the uncertainty of when they will arrive or leave.
Black swan theory - Wikipedia, the free encyclopedia
Black swan theory - Wikipedia, the free encyclopedia
  • en.wikipedia.org
The black swan theory or theory of black swan events is a metaphor that describes an event that comes as a surprise, has a major effect, and is often inappropriately rationalized after the fact with the benefit of hindsight. The disproportionate role of high-profile, hard-to-predict, and rare events that are beyond the realm of normal...
 
figurelli: That's right, maybe for the most of the time, but the problem are the black swans and the uncertainty of when they will arrive or leave.

A Black_Swan (Strategy) is a Trend following strategy. It spends allot of time taking losses ... waiting for a super_trend. Allot of traders do-not have the patience for such systems.

When the super-trend does arrive, it'll have to be strong enough for the trader to recover all their losses ... theres no guarantee thats gonna happen.

 
doshur:

My EA actually is a one pattern finder. I think I just need to "curve fit" to suit the current market changes.

So I do agree on the statement > in my opinion the answer depends too much in the EA algorithms

Like figurelli pointed out .... Forward_Testing is built into mt5. If someone believes they should optimize for Current_Conditions then they should be happy with their Forward_Testing analysis.

The forward test still has to be Large enough to be representative of different market conditions.

The back-tester is much more forgiven than real markets in terms of Slippage etc ... If a system isn't giving Great results for Forward_Testing what makes you think it'll work on Live_Trading.

Ideally, a system should be able to recognize the conditions its navigating and adjust accordingly  ... and NO you do-not need an Artificial Intelligence to accomplish this ....

If it fails the 10_Years && Multiple_Pairs testing ... Or Forward_Testing ... then its Failed as a trading System ... Back to the development cycle.

 
Ubzen:

A Black_Swan is a Trend following strategy. It spends allot of time taking losses ... waiting for a super_trend. Allot of traders do-not have the patience for such systems.

Nope,  it's a big unforeseen event that is big enough to kill an account . . .  in essence.
 
RaptorUK:
Nope,  it's a big unforeseen event that is big enough to kill an account . . .  in essence.


RaptorUK, indeed, this is the extreme impact, as stated by Nicolas Taleb.

By the way, and according to him, "a Black Swan is an event with the following three attributes":

1) "It is an outlier, as it lies outside the realm of regular expectations, because nothing in the past can convincingly point to its possibility"

2) "It carries an extreme impact"

3) "In spite of its outlier status, human nature makes us concoct explanations for its occurrence after the fact, making it explainable and predictable"

Most of traders think that this is not possible, since they use stoploss, and maybe they are right but in my opinion for me stoploss protect trades, don't protect accounts.

For sure you can create a good risk and money management to protect against some sort of black swans, but market has infinity uncertainty and, as RaptorUK stated before, "it's a big unforeseen event that is big enough to kill an account ".

 

 
RaptorUK: Nope,  it's a big unforeseen event that is big enough to kill an account . . .  in essence.

I was talking about Black_Swan Trading Strategy (should have made the distinction). I wasn't dis-agreeing with figurelli upon what a Black_Swan is || what it could do to an account. It says what Black_Swan(Event) is right under his post ... with the wiki article ... lol. Its not just an event which kills accounts ... it can also be an event which built accounts.

 

I do not know why some of you guys keep stressing that EA should show profit from other pairs as well. I believe each pair got its own moving pattern which differ from the rest.

https://www.mql5.com/en/forum/12298

https://www.mql5.com/en/forum/13862 

Looking back at this 2 threads, just a slightly more than 50% EA performs well on other pairs.

Does one trading strategy applies to one pair or multiple pairs?
Does one trading strategy applies to one pair or multiple pairs?
  • www.mql5.com
Does one trading strategy applies to one pair or multiple pairs?
 
doshur:

I do not know why some of you guys keep stressing that EA should show profit from other pairs as well. I believe each pair got its own moving pattern which differ from the rest.

https://www.mql5.com/en/forum/12298

https://www.mql5.com/en/forum/13862 

Looking back at this 2 threads, just a slightly more than 50% EA performs well on other pairs.

You do-not like the Multiple_Pair test for the same reason you do-not like the Multiple_Year test.

Your EA is Curve_Fitted to one Currency_Pair_Price_Movement @ a Particular_Point_In_Time. :) 

 
Ubzen:

Its not just an event which kills accounts ... it can also be an event which built accounts.

Ubzen, I do agree, good point, you can build accounts with black swans, like Taleb did, but you have to create a strategy and EA based on patience, as do the fishermen.
 
Ubzen:

You do-not like the Multiple_Pair test for the same reason you do-not like the Multiple_Year test.

Your EA is Curve_Fitted to one Currency_Pair_Price_Movement @ a Particular_Point_In_Time. :) 

Just what I was about to type,  so I won't 
12345
New comment