When to shut EA off? | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

FXLQ Shut down temporarily by NFA 323 replies

how to get my ea to shut up! 7 replies

Brokers who have been shut down already... 18 replies

markets have shut? 3 replies

CFG shut down by NFA 35 replies

Attachments: When to shut EA off?
Exit Attachments

When to shut EA off?

  • #1
  • First Post: Jun 21, 2016 6:08pm Jun 21, 2016 6:08pm
Here we have,
beautifull run of EA , say 4% drawdown simulated, and 8% live.
Or 9% simulated, when risking 2% of account - profits soar. [check biedarobot EA ]
Assume this is kept on demo account for some time.Say 100 trades.
1.
When would you leave your EA alone?
Shut it down - ok but, when?

I ve heard that when EA looses more than two drawdowns then it should be turned down.
What drawdowns? Simulated? Demo account drawdowns?

2.
How do end those hyper charged by cumulation effect accounts?
One year and paper milion.
What is reality here?



Attached Image (click to enlarge)
Click to Enlarge Name: TesterGraph.gif Size: 24 KB



Attached Image (click to enlarge)
Click to Enlarge Name: TesterGraph_c_a.gif Size: 22 KB
  • #2
  • Jun 22, 2016 5:36am Jun 22, 2016 5:36am
There is no fixed rule for that, because market changes and is not predictable. Whenever it leaves your comfort zone fits best in this case.

Hopping from one strategy to another or start fiddling in the setup is imho the main reason why 9 out of 10 traders loose money in the end.
  • #3
  • Jun 22, 2016 6:07pm Jun 22, 2016 6:07pm
Quoting Drolph
There is no fixed rule for that, because market changes and is not predictable. Whenever it leaves your comfort zone fits best in this case. Hopping from one strategy to another or start fiddling in the setup is imho the main reason why 9 out of 10 traders loose money in the end.
What about all those "5% weekly" EAs
Any of you had such a vehicle and it stopped working?


Every week below 5% income would be "below comfort" , optimise and search for new params or stick with it?
Till when?
  • #4
  • Jun 22, 2016 6:46pm Jun 22, 2016 6:46pm
If you have bought an EA off the shelf.. It will self destruct..

If you have built it and constantly update it then you may have a chance.

Ive never been one who relies solely on technicals but each to their own.

There isn't a right or wrong way to trade and don't let anyone tell you there is because its BS.
  • #5
  • Jun 23, 2016 2:45am Jun 23, 2016 2:45am
Quoting lasty
If you have bought an EA off the shelf.. It will self destruct.. If you have built it and constantly update it then you may have a chance. Ive never been one who relies solely on technicals but each to their own. There isn't a right or wrong way to trade and don't let anyone tell you there is because its BS.
It is hard to trust any EA if you are not a developer, moreover many EA are programmed to follow technical analysis so they are required to be monitored while news release session otherwise if not monitored they can blow the account.
  • #6
  • Jun 24, 2016 5:04pm Jun 24, 2016 5:04pm
I have simple lock against problems during news:
Trade only when spread is below, say 5pts on EURUSD.
During news it soars so either EA is stopped out - once or does not take any trades.

Also stoploss range based on volatility makes everything simpler.
Huge volatility - huge stop loss range - very small position.

One or two trades lost during news
does not make any difference when EA makes like 4 a day.
Just assure constant amount under risk, say 2% account.
  • #7
  • Jun 24, 2016 9:02pm Jun 24, 2016 9:02pm
The test is not equal to the facts.

init money: 10000
Attached Image (click to enlarge)
Click to Enlarge Name: TesterGraph.gif Size: 9 KB
  • #8
  • Jun 27, 2016 5:11pm Jun 27, 2016 5:11pm
Quoting lghr
The test is not equal to the facts. init money: 10000 {image}
You know why such thing is happening?
Try higher tf higher sl/ts/be.
Then move to vps <2ms
and go smaller tf/sl/ts/be [time frame/stop loss/ trailing stop/break even]

FInally how those guys make that scalping bots ?
  • #9
  • Jul 3, 2016 10:24am Jul 3, 2016 10:24am
The strategy tester stats not equal real stat it's only for basic principles testing of ur strategies and fx bots. Strategy tester don't calculate a commission and swaps, no spread jumps, requotes, have pricing errors. This is a very painful for intraday EAs.
Knowledge and skills turning to a technology
  • #10
  • Jul 3, 2016 1:02pm Jul 3, 2016 1:02pm
Really impressive results you got there, I don't know how it trades, but if it is based on Martin gale, then you should close and take profit from time to time, because those EA's will from experience come crashing down hard sooner or later, if it's not based on Martin Gale then keep it running forever as long as you have this uptrend in your balance.
Bulls are stupid Animals!especially when Im short!
  • #11
  • Jul 3, 2016 1:04pm Jul 3, 2016 1:04pm
Quoting mqldev01
The strategy tester stats not equal real stat it's only for basic principles testing of ur strategies and fx bots. Strategy tester don't calculate a commission and swaps, no spread jumps, requotes, have pricing errors. This is a very painful for intraday EAs.

1.
true

so how you deal with it?
2.
I made some paper earners and they broke at demo level.

​​how u select among all those optimized params?
how u use optimizer?
​one parameter at a time or two?

​

​3​.
​it did earn for some time and then one week drawdown occured.
a​t brexit see it went on zero.
​last week it was a loss.
​
as far as I understand marked conditions changed and
it need to change parameters.

all is based on simple balance algorithm from strategies revealed page.plus Ts sl be and stop orders based on
​based on volatility implied.
I ve been testing it on dukascopy data and it earns a lot.

​
​​then again when using mt4
imported history it takes losses and merely
keeps breakeven on monthly scale.

​
​​now I'm going to switch it on demo and observe.
​plus optimize against last week and a bit optimize
and also put it on demo.

what u suggest?
  • #12
  • Jul 3, 2016 1:09pm Jul 3, 2016 1:09pm
if you are going to optimize then make sure you are doing it on several pairs for at least 2 years each, you could curve fit it otherwice, and that is bad.
Bulls are stupid Animals!especially when Im short!
  • #13
  • Jul 3, 2016 5:23pm Jul 3, 2016 5:23pm
Quoting roughtrader
if you are going to optimize then make sure you are doing it on several pairs for at least 2 years each, you could curve fit it otherwice, and that is bad.
really?
you do optimisation and check params on two or more pairs?
one set of params for two sets of data?
interesting.
Pari EURUSD with .. USDJPY or what?
  • #14
  • Edited 12:47am Jul 4, 2016 12:27am | Edited 12:47am
The strategy tester in MT4 is utter crap. (it's a marketing tool used to drive the customer to put real money behind the system ....)
Only a forward test may be of some value, but that is still demo.
You can get some idea if your idea is valid via a backtest, but only way to be sure is do run the system live, in real time and with real money.
  • #15
  • Jul 4, 2016 5:06am Jul 4, 2016 5:06am
Quoting januszo
{quote} 1. true so how you deal with it? 2. I made some paper earners and they broke at demo level. ​​how u select among all those optimized params? how u use optimizer? ​one parameter at a time or two? ​ ​3​. ​it did earn for some time and then one week drawdown occured. a​t brexit see it went on zero. ​last week it was a loss. ​ as far as I understand marked conditions changed and it need to change parameters. all is based on simple balance algorithm from strategies revealed page.plus...
I see much pain in ur words. Keep doing and u'll collect needed experience, it's all about exp.
1. I understand how it's shout be in the real world. Just check tester bot movenment in visual mode.
2. I saw many fx bots coded by tester, some programmers don't calculate commissions or don't understand how brokers servers execute an orders, requotes, slippage, time sessions problems and many things.
3. Optimizer u shouldd use only for short periods were ur bot is failing with money management or risk management problems or wrong entries spotted. Amount of params u can choosing from max processors power of ur servers. Ducascopy quotes history don't cure the tester bugs mate.
Before putting ea in full automated mode u should take manual tests. Than demo forward testing with couple of accs with different settings/timeframes.
Knowledge and skills turning to a technology
  • #16
  • Jul 6, 2016 4:28am Jul 6, 2016 4:28am
I think you need to understand the rationale of your EA first and adjust accordingly, but i normally shut off my EA when there is NFP, FOMC, central bank interest rate's decision etc..
Pips Collector EA, No Martingale, No Grid
  • #17
  • Jul 9, 2016 3:09pm Jul 9, 2016 3:09pm
​the main problem
with ea i mentioned above posts
​was that It was exploiting mt4 1minute to tick fractal interpolation.
i had two instances of this ea

​
one on usdjpy 4h timeframe without trailing stop
but with break even
​
​Second on eurusd 1h with trailing stop around 0.5 implied volatility


​while eurusd still is suffering losses on demo
​usdjpy ​​is still performing.well because pair is trending maybe.still live trading


​firstly i wanna thanks for words of advidce still to be used.

​secondly
​i m using tick dukascopy data obtained by tickstory lite.
​anf optimizing for usjpy and eurusd on 15 60,and 240 minutes

​​albo i did modified ea to avoid trailingstop
when break even is not hit yet.

​
  • #18
  • Jul 25, 2016 2:49am Jul 25, 2016 2:49am
You should know what this EA actually does and based on this knowledge you should know when this EA is more likely to succeed and when not. You should turn it off when market conditions are not what this EA is designed for.
  • #19
  • Jul 27, 2016 11:49am Jul 27, 2016 11:49am
As a number of people have said, turn off the EA when the market changes in a way that will make the EA's actions unprofitable for you.
  • #20
  • Jul 29, 2016 10:31am Jul 29, 2016 10:31am
The best time to shut off the EA is when all the money is gone.
0 traders viewing now