Forex Factory (https://www.forexfactory.com/forum.php)
-   Platform Tech (https://www.forexfactory.com/forumdisplay.php?f=69)
-   -   Eas under construction III (https://www.forexfactory.com/showthread.php?t=244225)

jiva34 Jul 2, 2010 3:52am | Post# 1

Eas under construction III
 
3 Attachment(s)

If a strategy is good a robot should be able to trade it successfully,
Taking stress and impulsive and costly decisions out of the equation.

I, for one, am thankful for totally mechanical systems of trading.

This thread is dedicated to developing a 3 position ea using hi lo channels and stochastics.

Master coder Sashasan has worked tirelessly to write the code and has run limitless tests. We will see how to proceed to develop the effectiveness of the ea and I am sure learn much along the way.

The most current version will always be posted here on this page.


JIVA 3EMAs DSS v1.5.ex4
JIVA 3EMAs DSS v1.5.mq4

sashasan Jul 2, 2010 1:16pm | Post# 2

firstable, the idea of this system and ea creation is completely initiated by jiva, so all the benefits go to her!!!

the original system is posted here -
http://www.forexfactory.com/showthread.php?t=242490,
this ea is a 3-orders version of this system with some other additions

important outline: all the values in the ea's code is set for 4-digit broker


a little explanation of the v1.3 options:

-- StartHourByGMT,EndHourByGMT,GmtOffset - hours when ea will "turn on" and "turn off". if your broker's time is different from GMT, please use an offset value (positive or negative), to convert hours to original gmt value
-- magic1,magic2,magic3 - manually defined magic numbers, you decide to use it as permanent values
-- GenerateMagicNumbers - set this option true, if you want to use new magics each time new set of orders opened, this option has only one limitation: it's strongly recommended to use it on vps because, for example, terminal will be disconnected from broker's server after orders filled (internet, electricity problems), es will not be able to locate the orders with magics.
-- MagicLength - length of symbols (digits) in magic if it's generated (if GenerateMagicNumbers will be false, this option is not in use). MT4 can use any value between 1 and 11, but if it's more then 9, will display a different digit although will locate a magic successfully. to avoid repeat in those numbers, please do not use this value less then 5.
-- OpenThirdOrder - if true, ea will open a 3rd order, if false, only 2 orders will be opened
-- OpenOrdersWithDelay - specific option, can be used if for some reason you need to open the orders with a pause between each other
-- Order2Delay,Order3Delay - value in seconds which ea will wait until will open 2nd and 3rd orders. calculated based on the open order time of the previous order (so, if 1st order filled at 1:00:00, 2nd one will be opened after 1:00:30 (if delay is set to 30 seconds), and 3rd will be opened after 1:01:00)
-- OrderExecutionTime - additional option, related to the previous ones, sets up a value in seconds while ea will open each of 2 (or 3) orders started from delay value, so if we have set orders delay to 30 seconds and order execution time is also set as 30 seconds, ea will open 1st order(lets say 1:00:00), calculate the delay for 2nd order(1:00:30), calculate execution time for 2nd order (1:01:00) and will send an order during the time window between 1:00:30 - 1:01:00. it's important to keep those values approximately the same (if order 2 delay is 30 sec so order execution should be around that), otherwise it will overlap with the opening of next order. the necessity of this "weird" option became important after some backtests when market moved in the opposite direction and closed 2nd or 3rd order earlier then a 1st one, so if for that moment entry conditions again were met, ea opened a "second" 3rd or 2nd order then collapsed all session.
-- OrderPlaceDistance - distance (in pips) from the open price of a new candle (after a previous one closed below/above the ema's channel and all other entry conditions are on) where a set of pending orders will be placed
-- ExitCandleForPendingOrders - after placing pending orders, if they will not become a market ones during the number of candles set here, they will be deleted and ea will start to look for a new entry
-- Order1TakeProfit2,3 - take profit for each of 3 orders
-- Order1StopLoss,2,3 - stop loss for each of 3 orders
-- lot - the same for each one
-- TrailOrder2,3 - if true, 2nd and 3rd orders will be closed by trailing stop, of false, there can be closed only by breakeven stop loss or takeprofit
-- Order2Trailing,3 - trailing stop value for each of 2 orders
-- ChannelEmaPeriod - period for 2 "fast" ema's channel
-- TrendSmaPeriod - period for a "long" sma
-- DssEmaPeriod,StochPeriod - 2 values for dss calculation

Magix Jul 2, 2010 7:32pm | Post# 3

1 Attachment(s)
Very Nice Work Sashasan....

Played around with it, see what it would do under not so optimal conditions and hammering through a 1m chart with a few changes...

Though not completely optimized, shows an 8% week with a 5.82% draw....

A true trend follower....

Of course the modeling quality on a 1m chart sucks, but, under the same settings it pulled just over 4% on a 5m chart, 90% quality, but with the exact same draw...Still not terrible as there were no hard losers in the mix....01 lot size.

Definitely something for me to muck around with on the weekend.
Click to Enlarge

Name: jiva1m.jpg
Size: 158 KB

Magix Jul 2, 2010 7:43pm | Post# 4

1 Attachment(s)
Same settings for EU on the 5m...the Cable one shows modeling quality of 90%, for EU it comes in as an unknown...but still nice results.
Click to Enlarge

Name: saseu.jpg
Size: 64 KB

Taiyakixz Jul 2, 2010 10:23pm | Post# 5

Very Nice Work Sashasan....

Played around with it, see what it would do under not so optimal conditions and hammering through a 1m chart with a few changes...

Though not completely optimized, shows an 8% week with a 5.82% draw....

A true trend follower....

Of course the modeling quality on a 1m chart sucks, but, under the same settings it pulled just over 4% on a 5m chart, 90% quality, but with the exact same draw...Still not terrible as there were no hard losers in the mix....01 lot size.

Definitely something for me to muck around with...

I want to know your settings....

Magix Jul 3, 2010 12:24am | Post# 6

2 Attachment(s)

I want to know your settings....
Sorry mango,

Was playing around and changed them, forgot exactly what I used for the previous 2 charts...

In any case, the actual settings though important are not nearly as important as the time frame...Keep this in mind....

The 1m/5m support and resistance are the 2 most commonly played and challenged in any daily and weekly setting....

I got this set together thinking it was the same as the ones I had used, but they actually preformed a little different...I am sure if you change a little in them you should be able to get a lot better on the overall results, and this is only based on a single week, haven't tested the week by week, longer term nor have I forward tested....I only came across the EA today.

Pic is test...set is in zipped folder.
Report based on EUR/USD 5m, .01 per lot, for the last week, on a 4 digit broker, fixed spread 3.

Again, if you spent a little time with this, you could probably get it tweaked a lot better, base your sets on fractal patterned movements for your tp on individual trade pairs.





It tests about the same on cable, but again, not the same as what I originally used.
Click to Enlarge

Name: sash2.jpg
Size: 65 KB
jiva3ema1m.zip

mikok730 Jul 3, 2010 9:14am | Post# 7

hi jiva,
is that v1.3 is the latest version?

Magix Jul 3, 2010 9:44am | Post# 8

1 Attachment(s)
[center][color=Black][b]If a strategy is good a robot should be able to...
I seem to have a bit of an issue with the 1.3 version...It only sells on the pairs and with settings I have tested(including defaults), not one buy even though, if I understand the logic, and the mechanics, it isn't taking the triggers....

Chart example below. Movement above the trend-line, no long order taken...
Click to Enlarge

Name: jiva13.jpg
Size: 237 KB

jiva34 Jul 3, 2010 3:42pm | Post# 9

I seem to have a bit of an issue with the 1.3 version...It only sells on the pairs and with settings I have tested(including defaults), not one buy even though, if I understand the logic, and the mechanics, it isn't taking the triggers....

Chart example below. Movement above the trend-line, no long order taken...
Yes there may be an issue with the mechanics. I think the order delay function may be unnecesary. It was an idea I had but I think it is more of a problem than I realized. And the source of some of the order send problems which I am also getting. Lets see what Sashasan thoughts are on this. Maybe he can just take it out altogether. I think it would also help the ea load and run faster.
Thank you for the modified stoch setting. (25,34) I like it a lot. And will be experimenting with it over the next little while.
Thank you for putting time into helping to develop the ea.

jiva34 Jul 3, 2010 5:12pm | Post# 10

Bugs for opening orders and some stop issues in 1.3 fixed.
version 1.5 is available on the main page.

Taiyakixz Jul 3, 2010 5:12pm | Post# 11

Sorry mango,

Was playing around and changed them, forgot exactly what I used for the previous 2 charts...

In any case, the actual settings though important are not nearly as important as the time frame...Keep this in mind....

The 1m/5m support and resistance are the 2 most commonly played and challenged in any daily and weekly setting....

I got this set together thinking it was the same as the ones I had used, but they actually preformed a little different...I am sure if you change a little in them you should be able to get a lot better...
Thank you. I wanted some hints to start test. This weekend spread is very large, so I will test after market open.

Taiyakixz Jul 3, 2010 5:13pm | Post# 12

Bugs for opening orders in 1.3 fixed.
version 1.5 is available on the main page.
only ex4 file?

jiva34 Jul 3, 2010 5:25pm | Post# 13

only ex4 file?
oops, check again

Magix Jul 4, 2010 8:01pm | Post# 14

1 Attachment(s)
Thank you for the modified stoch setting. (25,34) I like it a lot. And will be experimenting with it over the next little while.
Thank you for putting time into helping to develop the ea.
A little something ya might wanna check out if you play the following....

GJ/EJ -> 16, 32
AU/AJ -> 6/32
GU ->24, 32

Observed with a Hulls moving average and compared with the long/medium term moving average indicator, it helps to show strengths on buys and sells....

Chart below.

Indication of stronger uptrend as price is above medium/long term trend lines...Weakness of sells on chart as current trend is long...In consolidation mode.
Click to Enlarge

Name: dss.jpg
Size: 104 KB

reachjj Jul 5, 2010 4:24am | Post# 15

Bugs for opening orders and some stop issues in 1.3 fixed.
version 1.5 is available on the main page.
for some reason, backtest for the last whole month on GU and EU 5min didnt result in any trade opening. There were BuyStop and SellStop orders, but all got deleted.

AM I the only person seeing this or all are getting similar results. Few posts above I saw that, last week resulted in some 5-8% profit. SO there must be something wrong either with my broker or with the latest version.

Note: I am using liteforex.com account

jiva34 Jul 6, 2010 12:18am | Post# 16

for some reason, backtest for the last whole month on GU and EU 5min didnt result in any trade opening. There were BuyStop and SellStop orders, but all got deleted.

AM I the only person seeing this or all are getting similar results. Few posts above I saw that, last week resulted in some 5-8% profit. SO there must be something wrong either with my broker or with the latest version.

Note: I am using liteforex.com account
there is an expiration setting. they may have expired...
try setting the expiration for 5 or 10 candles or else make the distance for pending order less. instead of 30 try 15 or 20

sanyhuis Jul 6, 2010 3:40pm | Post# 17

Jiva,
What about settings for 5 digit brokers?

Thank you,
s

reachjj Jul 7, 2010 2:46pm | Post# 18

[center][color=Black][b]If a strategy is good a robot should be able to...
Anybody is testing this EA?

jiva34 Jul 7, 2010 3:43pm | Post# 19

Anybody is testing this EA?
welcome.
yes we are testing..
it is under development but runs very well as it is. We welcome your feedback. So far we have tested mostly GBPUSD 30 mins. various settings.
it can catch very big moves if setting for 3rd position is left at 1000. Maybe it's not going to get 1000 pips, but the channel will stop it out for 300 or 400 if it catches the move.

another interesting time frame is the 4hr with 1 candle expiration and fast channel like 2ema or 3ema.. the ea then acts like a 2 candle breakout strategy on the 4 hr.

good luck
and thank you for your time and attention.
jiva..

MrChips Jul 9, 2010 3:42pm | Post# 20

Can't get it to trade
 
Hi

I've been trying to test this version for the past 2 days and it hasn't places a trade yet. Is there something I need to do?


© Forex Factory