normalise strategy drawdowns

imagine that i have 100 different strategies generated with fixed size (0.1 lots) - so every strategy has different max drawdown, for example from 300 - 2000 USD



i want to run automatic retest and normalize the drawdown to the "same" value - so if i set, that i want to normalize the drawdown to 1000 USD, some strategy will have 3.33x bigger position, and strategy with highest drawdown will have half position


can this be added to for example automatic retest task?

Attachments
  • Votes +11
  • Project StrategyQuant X
  • Type Feature
  • Status Refused
  • Priority Normal

History

h
#1

hankeys

02.09.2020 12:53

Task created

h
#2

hankeys

02.09.2020 12:53
Voted for this task.
m
#3

mabi

03.09.2020 07:08
Voted for this task.
m
#4

MT

05.09.2020 00:13
Voted for this task.
MF
#5

Mark Fric

11.09.2020 12:19

Status changed from New to Refused

I don't think this is a task for SQ, it would be maybe possible to compute the sizes for every strategy externally.
h
#6

hankeys

11.09.2020 12:55
this is very usefull function for traders - not have unbalanced portfolio from different kind of strategies


and its simple - add another MM method in automatic retest which will recompute the original used MM to best match to some drawdown value nad retest all strategies

m
#7

Martin

11.09.2020 18:10
This would be a great feature however I think it's harder than it sounds depending on the your money management settings and your stop loss settings to get this even close to been exact without entering in an infinite loop. I have been working in a similar way to what you want, adjusting all strategies to the same DD before using portfolio master. I implemented this by having a set of automatic retest with a certain ammount risked starting  as high as 100 (this would be my highest expected value which would result in all of my strategies having a DD equal or higher to my desired DD) -> save strategies in databank  -> delete those which have lower DD than the ammount you want and then clone those task by setting lower ammounts in money management settings such as 91, 82, 74.... (the lower you go the lower the step should be to better approximate). This is not perfect but it saves you time and eventually you can improve this project a lot. With this base project I am now adjusting them between 0.5 and 1 times my max DD based on their other stats, and it works pretty well approximating their DDs. I hope it helps hankeys.
m
#8

Martin

11.09.2020 18:13
Voted for this task.
h
#9

hankeys

11.09.2020 19:18
make it simple for example 3 strategie trading fixedlot - 1. has DD 500, 2. has DD 2000, 3. has DD 1500


if i want to normalize the DD to 1000 - i will change the fixedsize for the 1. strat 2x, 2. strat /2, 3. strat /1.5 - simple, why not?


and the same for fixed money - everything i need to do is compute the ratio and make new backtest

b
#10

beppil

11.09.2020 19:38
Voted for this task.
JK
#11

Insanity82007

24.11.2020 08:06
Voted for this task.
JK
#12

Insanity82007

24.11.2020 08:13

Attachment Normalise DD chat 1.png added

Attachment Normalise DD chat 2.png added

Attachment Normalise DD chat 3.png added

Attachment Normalise DD chat 4.png added

Attachment Snippets.zip added

Attachment WhatIfBulk.java added

Normalise DD chat 4.png
(368.47 KiB)
Snippets.zip
(3.39 KiB)
Normalise DD chat 2.png
(384.25 KiB)
Normalise DD chat 3.png
(307.94 KiB)
Normalise DD chat 1.png
(509.04 KiB)
WhatIfBulk.java
(5.29 KiB)
WHY is this task refused?!?!?! This will be EXTREMELY useful for normalising strategy results prior to exporting strategies.


I've already written snippets / scripts that do this in Quant Analyzer (attached).


I've also written VBA code in Excel to modify the exported MQL4 scripts to change the MM settings in each MQL4 file depending on my calculated MM for each normalised strategy.


SURELY this can be done in SQ without too much trouble!

b
#13

bentra

24.11.2020 08:33
Voted for this task.
k
#14

Karish

24.11.2020 11:07
Voted for this task.
m
#15

microsporum

24.11.2020 21:06
Voted for this task.
MF
#16

Mark Fric

25.11.2020 09:00
sorry, it remains refused. 


Anything can be added to SQ, but I don't see this as a part of SQ core functionality, it should be done by some other tool, or maybe by a custom made plugin.

h
#17

hankeys

25.11.2020 10:11
10 votes is not enough to convince the devteam to add this usefull feature?


how many votes and how many users using for example martingale MM which was added some time ago...

MF
#18

Mark Fric

25.11.2020 10:28
votes are indication of interest, but please let us use our own discretion in deciding what to develop based on our plans with SQ. 

I don't know any other software developer who listens to their customers as much as we  do.


Resources are always limited and there are many other things that we want to develop. I simply think there are bigger priorities than this.

f
#19

fm663

25.11.2020 13:07
Agree Mark. It would be good if you can provide ways that people can configure/extend features themselves.
b
#20

bentra

25.11.2020 17:33

Attachment Screenshot 79.png added

Screenshot 79.png
(13.25 KiB)
Guys I think the MM is not stored in the individual strategy itself unfortunately, it is stored the settings of the task so actually this would be a bigger job that it may seem.

Although, maybe it could be calculated and put in the code when this drop down is changed ie a new option here called "from strategy DD." A crude normalizing calc is not hard something like new_risk_size =  last_risk_size * (Target_Max_DD/Last_Test_Max_DD)




h
#21

hankeys

25.11.2020 18:51
everything is stored in SQX file, so there is no problem at all
b
#22

bentra

25.11.2020 20:48
Yes you're right, I see "define own size" can be inserted in to strategy by modding the strategy. It looks like it would be static setting though unrelated to account equity size and no longer adjustable by global settings from that point forward. Still maybe useful for just before putting live.
g
#23

geektrader

14.02.2021 13:43
Voted for this task.

Votes: +11

Drop files to upload

or

choose files

Max size: 5MB

Not allowed: exe, msi, application, reg, php, js, htaccess, htpasswd, gitignore

...
Wait please