what is the difference between the finantic trailing stop and the one already on wealth lab?
Rename
I assume you are talking about Building Blocks and the finantic.PowerGenes extension.
With WL comes a Building Block "Sell at Stop Loss" which has a CheckBox "Trailing" this generates code like this:
The finantic.PowerGenes extension contains a Building Block "Sell at Trailing Stop" which has no separate option to switch between a fixed stop loss and a trailing stop loss. It generates code like this:
Both Building Blocks generate identical code.
A similar argument holds for all the other Building Blocks in the finantic.PowerGenes extension. It is possible to replace all PowerGene Building Blocks by a combination of existing blocks.
The main difference comes with Strategy Evolver:
(I assume both PowerPack and finantic.PowerGenes are installed)
With vanilla WL blocks the random evolver process needs to choose
1.) the "Sell at Stop Loss" Block (a 16:1 chance) and
2.) choose to set the "Trailing" flag (a 2:1 chance).
So one out of 32 randomly chosen exits will be a trailing stop.
With PowerGenes it has to just to choose "Sell at Trailing Stop (a 16:1 chance).
So one out of 16 randomly chosen exits will be a trailing stop.
The differences become much larger if you look at other concepts like
"Sell at Profit Target in ATR or Timeout".
This is a very useful (and profitable) exit. Chances (for Strategy Evolver) to find it without PowerGenes are very, very low however.
(I leave the exact calculation as an exercise for the interested reader... ;) )
With WL comes a Building Block "Sell at Stop Loss" which has a CheckBox "Trailing" this generates code like this:
CODE:
CloseAtTrailingStop(pos, TrailingStopType.PercentC, 10.00, "Sell at 10% trailing stop loss");
The finantic.PowerGenes extension contains a Building Block "Sell at Trailing Stop" which has no separate option to switch between a fixed stop loss and a trailing stop loss. It generates code like this:
CODE:
CloseAtTrailingStop(pos, TrailingStopType.PercentC, 10.00);
Both Building Blocks generate identical code.
A similar argument holds for all the other Building Blocks in the finantic.PowerGenes extension. It is possible to replace all PowerGene Building Blocks by a combination of existing blocks.
The main difference comes with Strategy Evolver:
(I assume both PowerPack and finantic.PowerGenes are installed)
With vanilla WL blocks the random evolver process needs to choose
1.) the "Sell at Stop Loss" Block (a 16:1 chance) and
2.) choose to set the "Trailing" flag (a 2:1 chance).
So one out of 32 randomly chosen exits will be a trailing stop.
With PowerGenes it has to just to choose "Sell at Trailing Stop (a 16:1 chance).
So one out of 16 randomly chosen exits will be a trailing stop.
The differences become much larger if you look at other concepts like
"Sell at Profit Target in ATR or Timeout".
This is a very useful (and profitable) exit. Chances (for Strategy Evolver) to find it without PowerGenes are very, very low however.
(I leave the exact calculation as an exercise for the interested reader... ;) )
The discussion here is closely related to the Infinite monkey theorem (https://en.wikipedia.org/wiki/Infinite_monkey_theorem)
Applied to Genetic Strategy Evolver it says:
Of course the Evolver can find the best thinkable strategy but it will take forever to do so.
The situation improves if the Evolver has more powerful Genes to choose from.
Back to the monkeys: Chances to produce a nice text are higher if you choose from sentences or paragraphs instead of single letters.
Translated to Strategy Evolver:
Chances to produce a decent trading strategy are higher if you choose from more complex building blocks (concepts instead of elementary operations).
Applied to Genetic Strategy Evolver it says:
Of course the Evolver can find the best thinkable strategy but it will take forever to do so.
The situation improves if the Evolver has more powerful Genes to choose from.
Back to the monkeys: Chances to produce a nice text are higher if you choose from sentences or paragraphs instead of single letters.
Translated to Strategy Evolver:
Chances to produce a decent trading strategy are higher if you choose from more complex building blocks (concepts instead of elementary operations).
QUOTE:
The differences become much larger if you look at other concepts like
"Sell at Profit Target in ATR or Timeout".
Does that mean that timeout is the change here or are any other additional topics between WL "Sell at ATR Limit" and "Sell at Profit Target in ATR"?
Regards Christian
Thanks!
QUOTE:Yes.
Does that mean that timeout is the change here
And again, the probability for evolver finding an Profit-Target+Timeout combination among the "elementary blocks" are rather low.
(about 15 times lower then using the PowerGene if it decides to combine two exits.)
Your Response
Post
Edit Post
Login is required