Opinion of this method of developing strategies

[Sky123987]

I was wondering what you guys have noticed regarding the following situation.


* You have done a backtest and you have noticed that if you use JUST rule1 you get a profit factor of 1.2 going long.
* You also notice that if you use JUST rule2 you get a profit factor of 1.2 also going long

Generally Speaking
if you ONLY allow trades when Rule1 and Rule2 are BOTH present your profit factor when going long...

A) will be > 1.2
B) will be < 1.2
C) no relationship

 

[QQQShort]

The correct answer is (c), because the outcome of rule combinations differs by combination.

 

[Sky123987]

thanks kinda the conclusion I'm coming to, you'd think it'd be A, but yeah I think C

 

[telozo]

Quote from Sky123987:

I was wondering what you guys have noticed regarding the following situation.


* You have done a backtest and you have noticed that if you use JUST rule1 you get a profit factor of 1.2 going long.
* You also notice that if you use JUST rule2 you get a profit factor of 1.2 also going long

Generally Speaking
if you ONLY allow trades when Rule1 and Rule2 are BOTH present your profit factor when going long...

A) will be > 1.2
B) will be < 1.2
C) no relationship

Why don't you backtest using Rule1 AND Rule2 and see what the profit factor is?

 

[Sky123987]

the question was asked to see if a principle is true or not, thank you for your advice on backtesting the following Rule1 and Rule2, but I was hoping to make a conclusion w/ a sample size > 1. ~hence that is why I proposed the ?

 

[fearless9]

Quote from Sky123987:

I was wondering what you guys have noticed regarding the following situation.


* You have done a backtest and you have noticed that if you use JUST rule1 you get a profit factor of 1.2 going long.
* You also notice that if you use JUST rule2 you get a profit factor of 1.2 also going long

Generally Speaking
if you ONLY allow trades when Rule1 and Rule2 are BOTH present your profit factor when going long...

A) will be > 1.2
B) will be < 1.2
C) no relationship

rule 1 and rule 2 bear no necessary relationship with each other.

 

[MGJ]

Art Collins runs a thread here ( http://www.elitetrader.com/vb/showthread.php?s=&threadid=83627&highlight=odds+czar ) and wrote a book (http://www.amazon.com/Beating-Financial-Futures-Market-Strategies/dp/0470038659/ ) based on the same premise. You might want to check it out.

 

[thetrendfollowe]

Quote from Sky123987:

I was wondering what you guys have noticed regarding the following situation.


* You have done a backtest and you have noticed that if you use JUST rule1 you get a profit factor of 1.2 going long.
* You also notice that if you use JUST rule2 you get a profit factor of 1.2 also going long

Generally Speaking
if you ONLY allow trades when Rule1 and Rule2 are BOTH present your profit factor when going long...

A) will be > 1.2
B) will be < 1.2
C) no relationship

Are these different systems?

If they are then you should look for the lowest correlation possible, which would be C.

There is much to consider when evaluating a system.
Profit factor alone is insufficient.

You also need to know the win % and opportunity (trade frequency).

Nizar.

 

[syswizard]

Why bother with a strat with a profit factor < 2 ?
Exception: high frequency system (20+ trades per day)

 

[TrueStory]

Quote from syswizard:

Why bother with a strat with a profit factor < 2 ?
Exception: high frequency system (20+ trades per day)

I would encourage people to totally disregard this statement as it is completely fallacious. I'm not going to get into a long-winded discussion about system statistics as it's been done too many times on these forums, but suffice it to say one could be missing out on many extremely profitable strategies by following the advice above. Myself and many others I know earn a living trading strategies with profit factors less than 2.0 that don't trade more than once a day.