Dear TV, FF and others,
I'm done reading through the complete thread. It's been a pleasure to see that past members reappeared over time back and then. This shows ppl interest.
My thinking up to now:
- FF you started a great thread
- TV you provided non-coders here a "professional support and coding"
- There are way too many times the same questions on "I can't compile or attached the EA .... OrderReliable .... blabla" .... come on guys ... the only thing you have to do is "read" ... great ppl are doing the rest for you.
That said:
- The present code does not handle non-uniform grid but that is fine considering it does not allow to do so. Specifically, adding the "extra-increment" beyond the original TP would directly impact lot-size calculation .... at every level ... or it better be extra-pips you wanna suck in.
- I'm 100% on FF side when he said there were potential on "intermediate exit strategies"
, this is exactly what I meant when saying
- up-to-now, (using ViperEA .. and yes the latest version with backtesting improvements), and my own coding (with my own lot calculation), I'm not quite sure of the "starting from the outer band inward" to calculate lot size .... this does "over-weight" the last level ... however I do understand that it is statistically the most unlikely level for a reverse. Nevertheless, you'd better check ahead that this last level is not a support/resistance, pivot and else once you think of it ... obvious no
- I think what would work best is some kind of precalculated tables as we do in "distance transform calculation" to speed up the propagation of "a-priori known distances".... I have to find what entries would be best for the table though (level, increment, equity, nb of orders ....) and how to automatically "feed" the table to generate every cases with associated risk.
I'll keep you up-to-date with these concepts.
Great work.
XmPh.
I'm done reading through the complete thread. It's been a pleasure to see that past members reappeared over time back and then. This shows ppl interest.
My thinking up to now:
- FF you started a great thread
- TV you provided non-coders here a "professional support and coding"
- There are way too many times the same questions on "I can't compile or attached the EA .... OrderReliable .... blabla" .... come on guys ... the only thing you have to do is "read" ... great ppl are doing the rest for you.
That said:
- The present code does not handle non-uniform grid but that is fine considering it does not allow to do so. Specifically, adding the "extra-increment" beyond the original TP would directly impact lot-size calculation .... at every level ... or it better be extra-pips you wanna suck in.
- I'm 100% on FF side when he said there were potential on "intermediate exit strategies"
QuoteDislikedyou should work out the math to solve this problem, and that is something like 3 main exit strategies with each having 6 substrategies for loss mitigation.....so, total of 18 pure math exit strategies!
QuoteDislikedMy current mindset is to consider the grid as a predictor, that is you can calculate ahead what will happen .... so better handle now the worst case scenario based on your margin and think ahead to strategize "incomplete exit strategies" based on your current margin.
- I think what would work best is some kind of precalculated tables as we do in "distance transform calculation" to speed up the propagation of "a-priori known distances".... I have to find what entries would be best for the table though (level, increment, equity, nb of orders ....) and how to automatically "feed" the table to generate every cases with associated risk.
I'll keep you up-to-date with these concepts.
Great work.
XmPh.