automatic simplifier

I'll see strategies like this where TP's are never hit and I'll typically keep them to throw them in an optimizer later. So instead of automatically removing strategies like this by default, it may help to add a "Remove unnecessary blocks" toggle to allow us to decide if we want SQX to automatically remove these blocks/remove strats all together as recommended above. I think each strategy should come with some sort of efficiency index that measures the impact of each block on the strategy itself. You could come up with a weighting scale for each block and how manipulating it affects results (I'd find many strategies where one or two blocks did absolutely nothing to the strategies performance). This should be able to be calculated mathematically if some blocks prove to have close to 0 impact on performance. If an efficiency index is given to each strategy, users could remove blocks that rank below a certain weight of how effective/necessary that block is within the strategy.

However this would be an additional step after the strategies hit the databank between generations and may slow down the work flow. So I recommend implementing this in the robustness testing rather than the storing of the databank itself during genetic evolution/random generation.

Voted for this task.

Voted for this task.

i could agree, maybe the best way, where to implement this is a custom tasks, where we can think about task with a name "simplifier"

and we will be running this only on the final strategies - something like "last" task of a workflow

if the building block have no impact on net profit, it should be deleted

if the SL will be deleted (because it could happen too - you have to widen SL settings, so it will never be hit) - we must get some warning, that this strategy is actually withnout SL

Voted for this task.

I´ve also thought about adding this as a feature request, but as it demands a completely new module (it needs to run several backtests to determine which ones can be terminated) and given that we still have so many bugs with the current functionality, I decided not to request it yet. But you surely have my vote, I just hope they first fix all the bugs :-)

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

we need this function - we still have in strategies parts of a code, which are never taken

for example TAKEPROFIT - why i see in pseudo code, that i have strategy with takeprofit even if this TP was never used in list of trades

some simplification could be done very easily

is in pseudo code TP - are in list of trades some TP close types? NO - delete TP from the strategy, it was never used - have i turned on that i want TP required? dismiss this strategy, because TP in pseudo code was never reached

etc...the same for SL and other exit types

Voted for this task.

Voted for this task.

Voted for this task.

agree with hankeys upper

You could take this idea even further...and implement some kind of "parameter sensitivity test". For example sweeping across all parameters of a given strategy e.g.: using a simplified MC-like test and testing the outcome sensitivity vs variation of each parameter . This could give us an overview of all parameters and their impact on the final result. In the example above when TP is never reached, then it would have probably very small sensitivity figure since its change would not affect (much) the final result (at least on the positive side of parameter variation).

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

these are the reasons why we need this feature asap - i have set in build process, that i want TP and SL required...OK i am getting strategies with SL and TP in pseudo code, but from list of trades, they are never reached, so i am fooled by pseudo code, that i have strategy with SL and TP

and i dont think so, that my SL and TP settings are too high - for daily strategy....exit type exit bars or exit indicator were hitted sooner than any SL and TP

if i set that i want SL and TP required, i need strategy with these values, which are USED in strategy

second example:

i have set, that i want to build SYMETRIC strategies - OK, i am getting them, i see in the pseudo code - BUY and SELLS

but if i look at this strategy, it has only LONG SIDE, because the short side is never TRUE

why i see some pseudo code for SELL? when the sell conditions was never true

this strategy is not symetric, i have set, that i want symmetric strategies but i dont get one

DISMISS or SIMPLIFY

another CRUCIAL use for simplifier is a question of this kind

i am building multiTF strategies with filter for example on D1, i see in pseudo code that i have some filter in D1 chart - OK

but is this filter STRONG or not? what job is the filter doing in the strategy? isnt it useless?

here https://roadmap.strategyquant.com/tasks/sq4_5771 is strategy with D1 filter, if i remove the D1 filter from strategy i am getting better RDD - this kind of multiTF strategy is nonsense for me - i have in pseudo code something, which is worsens my results

with D1 filter i need better strategies than without it

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

any progress...crucial feature

Voted for this task.

Voted for this task.

I add (attached) a couple of examples of blocks that are never used and are in the pseudo-code.

It's very important to fix it.

devteam need to understand that this is a crucil feature and need to be implemented right to the builder, because if user wants anything as "required" this block will be in the strategy but must do some work in it

another example is that if i want to build multiTF strategies like H1 with D1 filter i need to know that D1 filter is doing some job in the strategy. i see many where D1 filter is always true so i dont have H1D1 strategy but only H1 strategy

Voted for this task.

any NEWS about this CRUCIAL feature for traders...we are still getting to the strategies building blocks which are not doing anything in the strategies

even if traders has STOP LOSS is required, this does not mean, that this stop loss is working in the strategy...

i could have exit signals in the strategy and from the list of trades i see, that it was never used

i could have trailing stop in the strategy, and from the list of trades i see, that it was never used

i could have take profit in the strategy and again, i can see, that it was never hitted

our strategies are full of building blocks and only few of them are making the edge for us...it could also happen, that some custom blocks could make strategy worse, etc. etc.

think about it to add asap, thx.

sending one strategy which is full with not used building blocks

to clean the pseudo code and keep the building blocks which are really used must be our basic task...

we cant see false positives in the pseudo code, like never used SL

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.

Voted for this task.