How different are backtest results against demo against live? - page 2

Marco vd Heijden
Moderator
13595

Yes i know what you mean several things are possible, they find something profitable in tester and go live hoping to become rich quickly , and blow out their account in a few weeks, or, some of it makes it to market, and blows out other peoples accounts, or it becomes available as a signal,

I found that when you find something, you discover something else, that will be working against you, that has to be overcome.

Francis Dogbe
10924
Francis Dogbe  
Alain Verleyen:

Thank you, I also use the Strategy Tester only for testing logic and debugging purpose.

But a lot of people believe they can use for other purpose, which can be more or less true depending of the strategy. As JD4 said a real account has to deal with spread variation, slippage/requote, but also eventually loss of connection, low liquidity, etc...I would like, a day, to create a topic not only to tell these things but to demonstrate concretely the actual impact of all these parameters on the trading results.

i will be glad if you did open the discussion.
Alain Verleyen
42598
Francis Dogbe:
i will be glad if you did open the discussion.
I will try ASAP. :-)
JD4
838
JD4  
Alain Verleyen:
Can I ask why in private ? It will be good if we had a serious topic about this, as it's a recurring question people ask. There are some general advice as the one of JD4 which are correct, but a detailed explanation would be very welcome.
I also wonder on this as I was under the impression that this could also be a repository of information not just for now, but if in the future someone has a similar issue, and comes here looking for a solution.   It would be great to be able to have that available.
Oscar Ortiz
363
Oscar Ortiz  

Thank you all for your comments,

Of course in backtesting liquidity and order execution is granted.

What I would like to know is if results are similar,

Or if your broker plays a role in allowing you to either win or lose even if you made a good decision on trend. 

JD4
838
JD4  
Oscar Ortiz:

Thank you all for your comments,

Of course in backtesting liquidity and order execution is granted.

What I would like to know is if results are similar,

Or if your broker plays a role in allowing you to either win or lose even if you made a good decision on trend. 

The broker question is a big enough topic to get it's own thread, TBH.  And it also depends on the type of broker as well.  You might want to take a look at the thread at https://www.mql5.com/en/forum/156114 about the broker choices out there.

The results are all over the place in regards to being "similar", and I think if they happen to actually be similar at some point, it is more random chance than anything else.  If it comes down to a choice over backtesting or no backtesting, it is a good idea to backtest, for most strategies anyway.  But that will give you potentially a dangerous false sense of what will happen when you "go live."  As long as you are aware that while the backtest may look like it will get you every single cent out of every possible trade, the live trading is very likely to be very different.  It is a clear example of showing the saying "hindsight is 20-20."

How do I know my MT4 account is a MM, STP or ECN account? - MQL4 forum
  • www.mql5.com
How do I know my MT4 account is a MM, STP or ECN account? - MQL4 forum
Oscar Ortiz
363
Oscar Ortiz  
JD4:

The broker question is a big enough topic to get it's own thread, TBH.  And it also depends on the type of broker as well.  You might want to take a look at the thread at https://www.mql5.com/en/forum/156114 about the broker choices out there.

The results are all over the place in regards to being "similar", and I think if they happen to actually be similar at some point, it is more random chance than anything else.  If it comes down to a choice over backtesting or no backtesting, it is a good idea to backtest, for most strategies anyway.  But that will give you potentially a dangerous false sense of what will happen when you "go live."  As long as you are aware that while the backtest may look like it will get you every single cent out of every possible trade, the live trading is very likely to be very different.  It is a clear example of showing the saying "hindsight is 20-20."


Thank you for your answer.
I still dont have it clear why it would be different.
Why would there be an algorythm to generate ticks.
Why storing entire tick history and testing on that not be the same vs live?
It's like having a book with the results of football games from years ahead (Back into the future movies) and saying it is not gonna work.
Please explain in detail why it would not work
Marco vd Heijden
Moderator
13595

I just did do you want me to explain it again?

I bet you didn't even read it.


The generation of ticks is carried out based on support points

Intermediate ticks between support points are generated according to the following rules:
  • If the number of ticks is larger than the number of points between the support points, then a "saw" is generated.
  • If there are a number of points between the support points, then a linear sequence of ticks is generated.


And it is not storing the entire tick history, only the support points, and it just generates ticks in between, it will never be the same or a copy of live market.

Oscar Ortiz
363
Oscar Ortiz  
Marco vd Heijden:

I just did do you want me to explain it again?

I bet you didn't even read it.


The generation of ticks is carried out based on support points

Intermediate ticks between support points are generated according to the following rules:
  • If the number of ticks is larger than the number of points between the support points, then a "saw" is generated.
  • If there are a number of points between the support points, then a linear sequence of ticks is generated.

Thank you again for your answer,
I wasn't asking how does the algorythm works but WHY should there be one.

Marco vd Heijden
Moderator
13595

Because of data usage and execution speed.

Look at the file sizes of historic data can you imagine they had to incorporate volume as well...

And because it is a tester, to test things nothing more.

If it doesn't suit your needs i guess the only option is to find something that does replicate every tick, good luck with that.