New EA - page 3

 
RaptorUK:
tonny,  all you are doing is showing your ignorance . . .  if you don't understand just ask and I will try to dumb it down so you can grasp this simple concept.
Are you sure ? Theory is one thing, how do you use this theory to build a winning strategy isn't so clear. As you need a sufficient number of trades to know the RR and WR of a strategy, you already known if it's profitable or not. And if you change a parameter there is a great chance that RR AND WR are changing. So yes, can you explain how you are using this curve to build a winning strategy ?
 
RaptorUK:
tonny,  all you are doing is showing your ignorance . . .  if you don't understand just ask and I will try to dumb it down so you can grasp this simple concept.
Maybe ill understand better in signal.
 

Hi Raptor !

Thanks for your detailed illustration, so nice of you. Okay, so what I get from the diagrammatic representation of Win Rate and RR, if your RR is 2, the BE Win Rate must be between 60 and 70% ?

 

Thanks 

 
angevoyageur:
Are you sure ? Theory is one thing, how do you use this theory to build a winning strategy isn't so clear. As you need a sufficient number of trades to know the RR and WR of a strategy, you already known if it's profitable or not. And if you change a parameter there is a great chance that RR AND WR are changing. So yes, can you explain how you are using this curve to build a winning strategy ?
It is not necessary to "need" sufficient number of trades to know the RR of a strategy, when you already have a predetermined strategy. If I decide that my SL will always be 100 pips and TP 300 pips, and no trailing stop and no interference during the operation of the EA, how I would need a number of trades to determine this. Yes, for the WR, we do need a number of trades, on that I agree with you, but not the RR. :)
 
angevoyageur:
Are you sure ? Theory is one thing, how do you use this theory to build a winning strategy isn't so clear. As you need a sufficient number of trades to know the RR and WR of a strategy, you already known if it's profitable or not. And if you change a parameter there is a great chance that RR AND WR are changing. So yes, can you explain how you are using this curve to build a winning strategy ?

 

Any strategy can be profitable for a day,  a week maybe even a month.  The question is how do you determine if a strategy is behaving the way it is due to it's design or due to chance and the nature of the markets ?

This concept (shown by the WR vs R:R chart) is not a strategy it is an analysis tool.

A "coin toss" type strategy can have a good run . . .  so why can't any strategy have a good run also ?  how do you know if your strategy is actually performing due to it's design or is just having a good run ?

 

This is from a "coin toss" type strategy . . .  

 

 

A "coin toss" type strategy is exactly the opposite of what everyone tries to achieve with the strategy they design,  it does not seek to predict the market in any way,  it simply tries to place trades in a way that is random and uses fixed SL & TP positions.  It will end up with equal long and short trades.  One thing that will surprise most people is that this "coin toss" strategy can have a Win Rate that is not 50% . . . in fact it's WR can be any value you like.

So how does a "coin toss" end up with a WR greater or less than 50% ?  simple,  by adjusting the SL & TP. 

This is the key fact . . . .   the WR for a "coin toss" type strategy is driven by the ratio of it's SL:TP or Risk:Reward.  SL:TP of  60:6 gives a WR of 91%  SL:TP of 6:60 gives a WR of 9%  (these figures assume a spread of zero) 

 

How does this help ?  

If your strategy's WR is also determined by it's Risk:Reward then it is no better than a "coin toss".   

 

How does one determine this ?

Determine the Risk:Reward for your strategy,  find it's WR,  plot these figures on the WR vs R:R chart  (or use the calculation method to determine the BE WR)  if your plotted point is close to the theoretical line then your strategy is as good as a "coin toss".   If it is not close to the line then your strategy is more likely to be profitable.

 

What are the complications in using this analysis ?

For many EAs this analysis is simple,  for manual trading strategies it can be difficult to determine a representative WR and Risk:Reward figures especially if the number of trades is small. For a strategy that does not use a fixed SL the risk may not be directly proportional to the average loss,  in some cases the average MAE should be considered as the true Risk for the strategy.   

 
Shunmas:

Hi Raptor !

Thanks for your detailed illustration, so nice of you. Okay, so what I get from the diagrammatic representation of Win Rate and RR, if your RR is 2, the BE Win Rate must be between 60 and 70% ?

Not quite . . .  the closer your WR is to the theoretical line,  66.6% for a R:R of 2,  the closer it is to a "coin toss" so you want a WR of more than 66.6%,  and the higher above 66.6% the better.
 
Shunmas:
It is not necessary to "need" sufficient number of trades to know the RR of a strategy, when you already have a predetermined strategy. If I decide that my SL will always be 100 pips and TP 300 pips, and no trailing stop and no interference during the operation of the EA, how I would need a number of trades to determine this. Yes, for the WR, we do need a number of trades, on that I agree with you, but not the RR. :)

If you have a fixed SL and TP then yes,  you can almost determine your R:R from these figures . . .  I say almost because the Spread will potentially have a part to play.  It may seem counter intuitive,  but the spread will actually make the WR for a "coin toss" higher . . .  

 

Consider a SL of  10 and a TP of 10 and assume a spread of zero,  this give a R:R of 10:10 of  1.0 and this gives a "coin toss" WR of 50%

Now consider a SL of  10 and a TP of 10 and assume a spread of   1.0  this changes the SL to 11 and the TP to 9,  this give a "coin toss" WR of 55%

 
RaptorUK:

If you have a fixed SL and TP then yes,  you can almost determine your R:R from these figures . . .  I say almost because the Spread will potentially have a part to play.  It may seem counter intuitive,  but the spread will actually make the WR for a "coin toss" higher . . .  

 

Consider a SL of  10 and a TP of 10 and assume a spread of zero,  this give a R:R of 10:10 of  1.0 and this gives a "coin toss" WR of 50%

Now consider a SL of  10 and a TP of 10 and assume a spread of   1.0  this changes the SL to 11 and the TP to 9,  this give a "coin toss" WR of 55%

Superb. Thanks for explanation. :)
 
Shunmas:
It is not necessary to "need" sufficient number of trades to know the RR of a strategy, when you already have a predetermined strategy. If I decide that my SL will always be 100 pips and TP 300 pips, and no trailing stop and no interference during the operation of the EA, how I would need a number of trades to determine this. Yes, for the WR, we do need a number of trades, on that I agree with you, but not the RR. :)
You are right for such strategies with fixed SL/TP ratio.
 
RaptorUK:

 

Any strategy can be profitable for a day,  a week maybe even a month.  The question is how do you determine if a strategy is behaving the way it is due to it's design or due to chance and the nature of the markets ?

This concept (shown by the WR vs R:R chart) is not a strategy it is an analysis tool.

A "coin toss" type strategy can have a good run . . .  so why can't any strategy have a good run also ?  how do you know if your strategy is actually performing due to it's design or is just having a good run ? 

...

Thanks a lot for this useful post. All the difficulty is now to draw the curve for a strategy.