backtest for 3 years, blow up in 3 days,

Quote from jacksmith:

One common story I had heard about trading system is that,

Successfully backtest for 3 year's data, yet the system blows up in 3 days,

what causes this ?

Thanks.
More...

Maybe this is wrong, IMO however, trading is changing permanently and this transfer of wealth is signifigantly impacting how markets behave. Would use caution with any backtesting strategy but that's just me. Good luck to you!
 
Quote from jacksmith:

One common story I had heard about trading system is that,

Successfully backtest for 3 year's data, yet the system blows up in 3 days,

what causes this ?

Thanks.
More...
Part of the problem, especially for HFT, automated systems, and scalpers, IMO is that you can't realistically backtest fills.
One solution that you may want to have a look at is Cyborg Trading's forward testing sim environment.
A Sterling Demo account can not help you forward test because it's fill engine is not realistic either. Anyone who has used it knows what I mean :)
It is just a small part of what Cyborg Does but it may help with this particular issue.
here's a link.

http://cyborgtrading.com/solutions/simulation
Forward test until your green and that should give you some confidence to go live.
We believe that fills are what trading is all about.
 
Quote from CyborgTrading:

Part of the problem, especially for HFT, automated systems, and scalpers, IMO is that you can't realistically backtest fills.
One solution that you may want to have a look at is Cyborg Trading's forward testing sim environment.
A Sterling Demo account can not help you forward test because it's fill engine is not realistic either. Anyone who has used it knows what I mean :)
It is just a small part of what Cyborg Does but it may help with this particular issue.
here's a link.

http://cyborgtrading.com/solutions/simulation
Forward test until your green and that should give you some confidence to go live.
We believe that fills are what trading is all about.
More...

Why, in particular, do you fill that for automated systems you can't realistically backtest fills ?

I do not have the same perception. Clearly a simulated fill may differ in several ways from a real fill (partial fills, slippage, latency, etc) but usually this is not something that prevents from determining at backtest time if a strategy will be robust and profitable.
Especially working with very liquid instruments.

Rather, I see the problem in the strategy itself, which can be more prone to overfitting (like signal based strategies) or unsuitable definition of performance stats or leaky filtering system, etc.

____________________
Tom
My <a href="http://www.datatime.eu/public/gbot/2009Oct19/default.htm" target="_blank">auto trading</a> journal
 
Quote from iuykcif:

Why, in particular, do you fill that for automated systems you can't realistically backtest fills ?

I do not have the same perception. Clearly a simulated fill may differ in several ways from a real fill (partial fills, slippage, latency, etc) but usually this is not something that prevents from determining at backtest time if a strategy will be robust and profitable.
Especially working with very liquid instruments.

Rather, I see the problem in the strategy itself, which can be more prone to overfitting (like signal based strategies) or unsuitable definition of performance stats or leaky filtering system, etc.

____________________
Tom
My <a href="http://www.datatime.eu/public/gbot/2009Oct19/default.htm" target="_blank">auto trading</a> journal
More...

Yeah, your definitely right. For a lot of traders unrealistic fills is not a huge problem.
Realistic fills become more and more important as the time frame of your trade becomes less and less.
Do you use Sterling Trade Pro?
 
if you arent getting positive slippage on your trades then your trading the wrong direction. words spoken from a counter trend trader = me. the knife catcher i am.
 
Quote from CyborgTrading:

Yeah, your definitely right. For a lot of traders unrealistic fills is not a huge problem.
Realistic fills become more and more important as the time frame of your trade becomes less and less.
Do you use Sterling Trade Pro?
More...


Right, with a small time frame it's crucial to work with liquid instruments.

Sterling Trade Pro: I did not know it I just browsed it on the internet. Is it good?

I am with IB, and i write all the software I need, when a need arises. (I enjoy reinventing all the wheels for personal achievement, deeper understanding and full control).

Tom
 
Quote from iuykcif:

Right, with a small time frame it's crucial to work with liquid instruments.

Sterling Trade Pro: I did not know it I just browsed it on the internet. Is it good?

I am with IB, and i write all the software I need, when a need arises. (I enjoy reinventing all the wheels for personal achievement, deeper understanding and full control).

Tom
More...

Yeah Sterling is good. It's one of a few platforms that are made specifically for prop traders. Sterling is very similar to anvil, prosper pro, laser.
I only mention it because it is the equities platform that Cyborg Trading currently connects to, although we will soon be releasing on genesis.
If you want to see a bit more of what we do pm me, or check out our site http://www.cyborgtrading.com
 
Quote from CyborgTrading:

Yeah Sterling is good. It's one of a few platforms that are made specifically for prop traders. Sterling is very similar to anvil, prosper pro, laser.
I only mention it because it is the equities platform that Cyborg Trading currently connects to, although we will soon be releasing on genesis.
If you want to see a bit more of what we do pm me, or check out our site http://www.cyborgtrading.com
More...


Thanks I will.

How is your product organized? Do you make available a few strategies among which the user can choose?

[I am currently with IB, and while rewriting the robot to connect to another broker too would not be impossible, I think I will undertake such effort only if the broker is among the top ones and has nice .NET API available. ]

____________________
Tom
My <a href="http://www.datatime.eu/public/gbot/2009Oct19/default.htm" target="_blank">auto trading</a> journal
 
You must log in or register to reply here.
Back
Top