EAs under construction II

I placed a few trades manually the last few trading days. I have to admit the quality of the experience and the results in trading with an ea is so far superior to trading manually for me. I never want to trade manually again.

re: chart errors in tester

Have noticed a quirk that can really throw results of a test off.
Chart errors.
What I have been doing and it is making a huge difference:
Download data from history.
Run the test, not for the full time, just a few trades, stop it, open the chart and refresh.
Then run the test for the full time period.
this seems to reset the chart for the full test.
The example below shows what can happen when the chart has errors, blackholes, time gaps etc, to the tester results.
This is a before and after comparison.

The problems I had with v4.1 were related to this chart error quirk in the tester not in the ea. It has just been revealed what the actual problem was. So yes v4.1 is very good.

That makes 4 out of 4 successful trades in 7 days using the 4hr settings.
They act as a 2 candle breakout on the 4hr.
Here is the most recent, I was sleeping when it placed and closed.

USDJPY is setting up, but not every pending order is filled. These settings have been tested at least in backtesting for optimal results. Like 100% success for the last 12 months.
The pending is place quite a distance from the recent closing prices. So It would have to be a significant move for it to be filled. The pair has 3 candles to reach it.
Not likely to fill. This is a 30 min setting.
However these can and do produce winning trades. There was a winner for this pair with this setting 1 trading day earlier.
It is not how much is being missed but rather how much is not being risked. Although it is a rather large stop loss 80pips, the success rate is 100%. Of course there will be losses at some point. But the idea is they can be absorbed for sustained profitability.

Didn't have ea set up on chart.
Missed this one..

new set file for USDCAD 4hr.

jiva what version
I'm forward testing 7.2 no trades yet

will backtest when get some time

v7 series is a no go...There was some buggy stuff in it..
v4.1 posted 1st page is very neat and clean.
We tried some tricky stuff in v5 and v7 but none of it was really any good and it just clogged up the ea.
So back to v4.1. The best version we have to date.
v4.1 does the job.
presets are there too.
Now we have 5 with great settings. That we know can produce 2 or 3 winning trades 2 positions per trade per month each.
It is obviously a 2 candle breakout strategy that is being executed on the 4hr.
So far so good. 4 out of 4 winning trades in the last week.
Different pairs. So eggs not all in one basket.
This ea can do other things too but for now lets get this set. I would like at least 8 or 10 pairs set up for the 4hr... We have 5...
So 10 pairs set for the 2candle breakout on the 4hr and each trading say 2 trades per month, averaging between 30 and 50 pips per trade.
Figure high, 50 per trade (total 2 positions) that's 1000 pips/per month. If I have my math right... These are all set for 100% wins over the last 12 months. If there are some losses they would be absorbed for overall profit.
Set em and forget em...
They're sleepy... I mean 2 trades per month isn't going to keep anyone sitting on the edge of their chair for hours on end. But add it up at the end of the month... Nice...

do you know if the presets i did for v5 work in v4

also once you get a good batch of pairs going we could incoporate Steve Hopwoods MPTM ea to do batch trading

Any settings for v5 should load the same into v4.1 However they may test differently.
Also found a deadly bug in the tester. see my post http://www.forexfactory.com/showpost...&postcount=282
Pretty significant...

These chart errors are very troublesome.
Here is another conflict.
Immediately after downloading data from history the test was run.
Producing no losing trades at 90% modeling quality.
After refreshing the chart to eliminate gaps and blackholes
the test was run producing a losing trade on 9/16/09
at modeling quality n/a.
So pick your poison so it appears.
Test using data from the history with 90% modeling quality but may and does have serious chart errors, gaps and black holes that produce erroneous trades.
or
refresh the chart and clean up the chart errors, gaps and black holes, but have a test with modeling quality n/a.

I had heard there were problems with the tester. Now I see. There may be more I haven't encountered yet, but this is more than enough. Thank you.

hi jiva34 and all others who have keep this thread interesting

i am new in this system

sorry ,, for Q
how can i change the trading times on this EA ?

You need to know what gmt offset your broker is.
Then you put that number in gmt offset.
I use FXDD and they are +3 offset from gmt which = 0
So I put 3 in the settings.
Then the times for trading are calculated from gmt.

At gmt = 0 that is the first hour for the new day and if I enter 0 for trade start hour I am allowing trading from the start of the new day. It is very messy really. Because there is the time where you live, the time for your broker, and then there is the absolute time gmt.
But it is all calculated from gmt and for trading all you need to know is your brokers offset and what times you want to allow trades.
I know that explanation really helped.
Others can perhaps answer more clearly.

the dragon caught USDJPY.

Here's another set file EURJPY
So now we have 6.

Oh I almost forgot.
The ea has a name.
The Dragon..

EURJPY setting up for short.
New setting, let's see how the dragon does..
One 4hr candle to reach the sell stop

okay, trade was good. These are quick, reliable trades...
The 2nd position somehow closed along with t he 1st position, will have to keep an eye on the why and wherefore, the 2nd postion was set for 40 pips and the 1st for 15, so why the 2nd closed along with the 1st is a bit of puzzle. No harm, they are both gainers.

We moved:
http://www.forexfactory.com/showthre...79#post3937579

what do you think of this??

http://www.forexfactory.com/showthread.php?t=225915

we can create a perfect ea??