Thank you very much @nonetheless!
Thank you all. I have included my favorite MM model in the latest version. Thank you
Hello @volpiemanuele,
I have been testing your MM strategy and I really like it.
Can you explain a little more about what each variable does?
prof = 0
ddpp = 212// max dd per point
ddm = 2 // multiples of max drawdowns to factor into positionsize
capital = 5000 //starting bank
dpct = 5 //% margin for deposit
deposit = (open/100)*dpct //margin per point
pri = 1 // percent to reinvest as a decimal
Also I was wondering how you arrived at the “212” number for Max DD per point.
It would be great to know a little more about how it’s calculating the increments.
Thanks so much in advance.
@StingRe Below I explain the code:
ddpp = 212// max dd per point I RUN THE BACK TEST WITH MIN SIZE OF INSTRUMENT AND I INSERT THE DD SHOWN BY BACK TEST
ddm = 2 // multiples of max drawdowns to factor into positionsize IN THIS SECTION YOU MUST CONTROL THE RISK; I SUPPOSE THAT THE FUTURE DD IS DOUBLE OF THE CURRENT BACK. YOU MAST INSERT OTHER FACTOR. IF YOU INSERT 1 MEANS THAT YOU SUPPOSE THE SAME DD FOR THE FUTURE
capital = 350 //starting bank
dpct = 5 //% margin for deposit
deposit = (open/100)*dpct //margin per point
pri = 1 // percent to reinvest as a decimal IN THIS SECTION YOU MAST INSERT THE % OF THE PROFIT THAT YOU WANT TO REINVEST.. FOR EXAMPLE 1 ALL PROFIT REINVESTED AND 0.5 IF YOU WANT TO REINVEST 50% OF PROFIT (ADMITTED VALUE 0.1 TO 1)
positionsize = (capital +((strategyprofit + prof)*pri))/((ddm*ddpp)+deposit) THIS FORMULA CALCULATE THE SIZE. YOU CAN REBUILD THIS FORMULA IN EXCEL WITH OTHER VARIABLES IN ORDER TO UNDERSTAND HOW THE LOTS VARY ACCORDING TO THE VARIATION OF THE OTHER VARIBLES
if positionsize < 0.25 then CHECK THE MINIMUN LOT (0.2 FOR EXAMPLE FOR DJ) OR YOU MUST DECIDE THE SIZE BEYOND WHICH THE MM START (IN THIS CASE THE MM START AFTER 0.25)
positionsize = 0.2
and the NAS short. They all got a small bump in performance with the new TS but the main thing is that it should work. Let me know if there’s any further problems.
PaulParticipant
Master
I created a basic version to start optimisation from.
It’s like this.
if check=1 then
condbuy=c1
condsell=c2
elsif check=2 then
condbuy=c1 and c3
condsell=c2 and c4
elsif check=3 then
condbuy=c1 and c3 and c5
condsell=c2 and c4 and c6
elsif check=4 then
condbuy=c1 and c3 and c5 and c7
condsell= c2 and c4 and c6 and c8
elsif check=5 then
condbuy=c1 and c3 and c5 and c7 and c9
condsell=c2 and c4 and c6 and c8 and c10
elsif check=6 then
condbuy=c1 and c3 and c5 and c7 and c9 and c11
condsell=c2 and c4 and c6 and c8 and c10 and c12
elsif check=7 then
condbuy=c1 and c3 and c5 and c7 and c9 and c11 and c13
condsell=c2 and c4 and c6 and c8 and c10 and c12 and c14
elsif check=8 then
condbuy=c1 and c3 and c5 and c7 and c9 and c11 and c13 and c15
condsell=c2 and c4 and c6 and c8 and c10 and c12 and c14 and c16
elsif check=9 then
condbuy=c1 and c3 and c5 and c7 and c9 and c11 and c13 and c15 and c19
condsell=c2 and c4 and c6 and c8 and c10 and c12 and c14 and c16 and c20
elsif check=10 then
condbuy=c1 and c3 and c5 and c7 and c9 and c11 and c13 and c15 and c19 and c21
condsell=c2 and c4 and c6 and c8 and c10 and c12 and c14 and c16 and c20 and c22
elsif check=11 then
condbuy=c1 and c3 and c5 and c7 and c9 and c11 and c13 and c15 and c19 and c21 and c23
condsell=c2 and c4 and c6 and c8 and c10 and c12 and c14 and c16 and c20 and c22 and c24
endif
Because if what’s not checked, in part the indicators are not activated and optimisation goes quick. And build it up. Maybe the order is not right or it will fail all together. Just a try.
Hey,
I’m running DJ 5m MOD v5.1 live with a startposition of 0.2 instead of 1. It went short today (October 30th) at 3:30 (European Time) and exited at 6:55 with a 49.94pts gain.
However, it doesn’t show on the backtest with the same algorithm as on live.
I have to say though that I changed the time indications in Nonetheless’s original code to fit European Time (as opposed to UK time) by adding one hour on lines 231 and 241.
Has it also happened to you or is it only me?
@BobOgden No trade today. I have change the time in the same lane. (I insert Italian time…1530 – 22.00….1540)
Thx volpiemanuele
I’m so confused…
Does anybody know if there’s any way to retrieve the code from a live system so I can compare it to the actual code?
Yes, in autotrading you can click the version number of the strategy you want to retrieve, then use Copy & Paste.
I created a basic version to start optimisation from.
@Paul, this looks really interesting but you have to explain it to me like I’m 4 years old – what do I do with it?
Thanks roberto.
So I’ve compared both algorithms and they’re exactly the same. Still no idea as to why it went short.
PaulParticipant
Master
@nonetheless Where to start optimising on this strategy? I thought why not add layer on layer and see what happens. It’s important to have many trades in the beginning, not necessary with the highest payout. Uses same sl & pt & ts and no other exit criteria.
So I enter a range for each variable and it automatically optimizes them one at a time?
PaulParticipant
Master
Not sure if I read that right, but I think not.
Since check3 uses a change in curve to go to true condition, I would say go up from there.
In other words, you make a choice for the check number and you can optimise the last condition or up to every condition the check number covers.
All conditions related to a higher check number are ignored.
So you take check3, optimise parameters for c1/2 c3/4 c5/6
Having found them, put them in, and go to check 4
then take parameters for c7/8 to optimise and/or you can take some parameters from above and optimise together. Found the best mix, go to check5 and now the focus is c9/c10 and it’s parameter.
The early results with a low check number will be bad and you need lots of trades which will be filtered out on a higher check number
Well that’s the rough idea and I hope it’s coded right, but it will be very tough to find something similar to your result if starting from scratch!
edit; I don’t think it was needed to reset conditions to 1.
