Topic: EA Advisor Studio optimum settings for generating possible strategy's
Hello everyone, here a message for all the experienced users. Especially
Blaiserboy, Popov , Footon, Hannahis ,etc, are invited to post a reply
I'm generating strategies with the following settings. After that i select the top 5 which look to give consistent profit without an ongoing monthly drawdown. Then after a few weeks i check which strategies are profitable and i put these on another demo account. Somehow sofar my results are far from perfect so i'm wondering what i can improve.
I'm trying the 1m , 15m, 30m, 1h, 4h timeframe so far.
my settings are: (unfortunately i cannot post then one image)
Reactor
1 historical data; i select date from a mt4 account where i uploaded the data generated via the script, symbol , timeframe
2 strategy properties:
stop loss: allways use
type: fixed or trailing
min pips 10
max pips 200
take profit
take profit: may use
min pips 10
max pips 2000
3 generator settings
working minutes: 0 (so works xx hours till i stop it)
search best: return/ drawdown
out of sample: 30%
generate strategies with preset indicators
no entry rules
long exit: close a long position at bar open
short exit: close a short position at bar open
use the common acceptance criteria = min net profit 10, min count of trades 100
v Full data optimization
in sample
numeric values range +-20 steps
search best net balance
v optimize stop loss and take profit
v optimize preset indicators
validation
v use the common acceptance criteria
min net profit 10
min count of trades 100
v Walk forward optimization
segments 5
out of sample 30% OOS
numeric values range +- 20 steps
search best: net balance
v optimize stop loss and take profit
v optimize preset indicators
validation:
validated segments: 5
v use the common acceptance criteria
v Normalization
remove take profit
remove needless indicators
normalize values
reduce stop loss
reduce take profit
normalize incicator parameters
Optimization options:
numeric values steps: 20 steps
search best: net balance
not selected: normalize preset indicators
v use the common acceptance criteria
Perform Robustness testing:
v Monte Carlo validations
count of tests: 20
Monte carlo validation uses settings from Monte Carlo:
market variations
v randomize history data
v randomize spread
execution problems: none selected
strategy variations:
v randomize indicator parameters
v randomize backtest starting bar
validated tests: 80%
Thanks a lot for your help everyone!!