Jiva Dragon

http://obiblog.files.wordpress.com/2008/08/data-pic.jpg

We are not dead!!
We may be buried in data files.
But still very much alive and kicking.
Will be coming out with new settings and a nice twist on the strategy.
It's looking good. Very good..

Hm, don't know. Yes indeed,1st position closed in one broker, not in another, you can see in my screens. Also I can remember the previous trade maybe 2 - 3 weeks ago (v11) - the same situation, 1st position was closed at BE and 2nd remained opened w/ initial TP and SL and when 1st position closed, the price went to oposite direction and hit SL of 2nd position and thus the result was loos of total 120 or even 200 pips (cannot remember exactly), I try to find that trade.

Does this symptom occur on only Colosseum....?
If so, it would be hard to fix for it

No Taiyakixz, I know, it wasn't only Colloseum. I run Dragon in 4 or 5 brokers, have to check at home later on to confirm that. So far to all traders: Please watch your trades closely, if they are processed correctly. I don't think I'm one and only, who has spotted it. I will update you as soon as I find another issue.

**********This the is archived page for the v11 series if anyone needs to refer back********

JIVA Dragon EA DSS T3MA testv11 mq4
jiva34, Taiyakixz |
Coded by Taiyakixz
Thanks to geoffb65 for inspiration:
http://www.forexfactory.com/showthread.php?t=231650

edit: templates have been added to all settings in the zip files 11/3/10
EURUSD 1h(optimized FXDD) 10/27/10 (50/50 test 61.11% win rate)
http://www.forexfactory.com/showpost...&postcount=737

EURUSD 15m(optimized FXDD) 10/23/10 (50/50 test 55.56% win rate)
http://www.forexfactory.com/showpost...&postcount=528

EURGBP 1h(optimized FXDD) 10/27/10 (50/50 57.14% win rate)
http://www.forexfactory.com/showpost...&postcount=738

GBPUSD 1h(optimized FXDD) 10/28/10 (50/50 test 58.97% win rate)
http://www.forexfactory.com/showpost...&postcount=739

EURCHF 1h(optimized FXDD) 10/28/10 (50/50 test 93.33% win rate)
http://www.forexfactory.com/showpost...&postcount=740

AUDUSD 1h(optimized FXDD) 10/28/10 (50/50 test 71.43% win rate)
http://www.forexfactory.com/showpost...&postcount=741

EURAUD 1h(optimized FXDD) 10/30/10 (50/50 test 57.14% win rate)
http://www.forexfactory.com/showpost...&postcount=742

AUDCAD 1h(optimized FXDD) 10/20/10 (50/50 test 59.52% win rate)
http://www.forexfactory.com/showpost...&postcount=743

NZDUSD 1h (optimized FXDD) 10/21/10 (50/50 test 63.64% win rate)
http://www.forexfactory.com/showpost...&postcount=746

NZDUSD 1h (optimized FXDD) with moving average shifts 10/24/10 (50/50 test 59.52% win rate)
http://www.forexfactory.com/showpost...&postcount=744

EURCAD 1h (optimized FXDD) 10/22/10 (50/50 test 78.57% win rate)
http://www.forexfactory.com/showpost...&postcount=745

Alpari settings

EURGBP 1hr (optimized Alpari) 10/21/10
http://www.forexfactory.com/showpost...&postcount=750

EURUSD 1hr (optimized for Alpari) 10/26//10
http://www.forexfactory.com/showpost...&postcount=747

EURCHF 1hr (optimized Alpari) 10/26/10
http://www.forexfactory.com/showpost...&postcount=748

AUDUSD 1hr (optimized Alpari) 10/29/10
http://www.forexfactory.com/showpost...&postcount=749

Trading algorithm for the 15 min:

1. Trades are only taken Tues., Weds., Thurs. and Friday(Friday may get phased out if it produces too many losing trades).
2. Trades are only taken between 1 hour after Tokyo open until noon NY.
3. macd default, dss default, channel is fixed at 2 ema.
4. Settings are optimized from the previous day to use for the next day.
So for example after close of day Monday optimization is run using Mondays data to set up for Tues.
5. Simple optimization for 2 moving averages. One slow, the other fast. Both on the 15 min chart.
6. At present 1st take profit, 2nd take profit, and stop are all fixed at 30. (We will be seeing how best to optimize the 2nd take profit).

Walk forward testing over 6 months for EURGBP using 99% modeling is showing a slight edge with win rate 53%. It has been consistent and we want to see how increase that. After studying many of these optimization graphs I am seeing that they can actually tell a lot about what is going on in the trading that can help for setting up for the next day.

A link for the optimized sets for each new trading day, will be posted on the title page shortly after the close of each trading day starting at close of Monday, NY, for Tuesday.
Close of Tuesday for Weds, etc..
These optimizations will be run on FXDD.

It is simple procedure and traders are encouraged to run their own optimizations for their respective brokers.
Take the basic setting from set file and check the box next to the moving average for fast (ma8) and check the box next to slow(ma7). Run the optimization for Monday's data and select the setting that shows the slowest for both the slow and the fast moving averages.

It's a pretty generic set file and I don't see why this strategy couldn't be applied to almost any pair. Practical application will show which pairs are best. The values of the fast and slow moving averages will change. Everything else remains the same.

Note: None of this is set in stone. It is just what we are presently working with because we are seeing some little bit of consistency and success.

The charts show an example of a backtest and walk forward...
The back test was done for the one day at the close of Thursday 8/20/09 to 8/21/09.
From the list in the optimization results (chart #1) the set with the slowest slow and slowest fast moving averages was chosen.

Chart #2 shows the optimization graph. They are not always this neat and clean. So it takes some practice to be able to chose the set that will perform the best for the next day.

That chosen set from 8/20 was walked forward (applied) for testing the next day 8/21/09 to 8/22/09.

Result was a continuation of the trend from the previous day's trading and good trade...

This is a great plan. But there may be a problem getting the data so soon after the market closes. It seems that MetaQuotes only has data available for up to about 1 month before the present day.

If I get the data for the previous day from Dukascopy that will be good for me and anyone trading FXDD becasue I will post the sets. But it's not going to help traders who are using other brokers. They won't be able to optimize their settings unless they can get Dukascopy data also and to use the FXDD settings can be awkward.
If I can get the previous day data maybe I can zip it and post it so traders can optimize their settings.
But I don't know yet if Dukascopy has data for the previous day available that soon. I opened a demo account in order to get the downloads because it is the easiest way but now they are calling me on the phone "When am I going to open a live account?" The demo is only good for 14 days. Now I can't log on. I think they may have blocked it because I didn't answer the phone and have been downloading nonstop day and night.
If we cannot get previous data in time for Tokyo open of next day then it's back to the drawing boards.

There is another way to access the 99% data without opening a demo and I will have to spend some more time figuring out how to do that. Since this demo thing isn't going to work.

hmmm. So many bug a boos...

Then again maybe we don't need to optimize at all. Just use the same set all the time. A slow moving average and fast moving average. Will look into running that just as it is. For Mon(maybe not), Tues, Weds, Thurs and Friday(maybe not)..
See how that compares to the test I ran for EURGBP with all the optimizing.
I have sneaky suspicion it really won't matter all that much optimizing or not optimizing... for the actual live trading.

That the data may only be important for back testing and forward walking to establish the algorithm but not for continually updating the settings.

http://www.thinkwebdesigns.com/images/light_bulb.png

Bare with me, we'll get there.
Sometimes it take a little time,
but it always takes lots of patience, strong determination and unshakable confidence of the ultimate success.

http://bloximages.chicago2.vip.townn...review-300.jpg

Sometimes it take a little time,
but it always takes lots of patience, strong determination and unshakable confidence of the ultimate success.

[/quote]

I have done some backtests on EU 60 min. long only, short only and long and short 30 pips SL very good results 86.99% success from 1.1.10 to 31.10.10 . But... When I changed the dates to 1.1.09 with the same set I get bad results.
So I agree with your post Jiva.
Good week all.

I have done some backtests on EU 60 min. long only, short only and long and short 30 pips SL very good results 86.99% success from 1.1.10 to 31.10.10 . But... When I changed the dates to 1.1.09 with the same set I get bad results.
So I agree with your post Jiva.
Good week all.[/quote]

Yes, backtest can look so good, but it is going forward we want.
So the ea is capable, the strategy is valid, now we have to see how manage things so there is more good trades than bad..

So yes this is going to take a little time.

I am still a little tied up in trying to get Dukascopy 99% data. There are a few more hurdles there and it is very time consuming.
I was using the download through their demo but that is not going to be along term method.
Now I have to figure out the php method which has no expiration date and no phone calls from Dukascopy "when am I going to open a live account"?
But it is a little complicated. So time time time.
Also we are working on a couple of small modifications to the ea.
More sensitive to moving average slope, and stopping out truly at breakeven and avoiding spread fee...

Sometimes just have to step back a little and survey the situation to get some insight for direction to go...

11/14/10
This week we will be testing a set strategy that walked forward very nicely in the few tests I was able to do.
We doing the unthinkable and increasing the stop loss. Because this has tested nicely in forward testing we are not going to disregard it because of a larger stop loss.
I still need to set up so that we are able to access the most recent tick data at 99%. That is going to take some to figure out but we are working on it.
So these tests were done 90% (I'm spoiled and don't trust 90%, as it has been said "that 90% figure in the upper right hand corner is just a figure in the upper right hand corner")...
But for now and for the sake of not just sitting and not testing anything.

EURUSD 15minute
Here is how we derive the settings:
Optimized for the previous week, starting Monday open, ending Friday close.
1. For now 1st take profit and 2nd take profit both equal 30.
1. optimize together for channel, stop loss, 1, 15, 60, and 1440 moving averages.
channel should be from 20 to 55
stop loss between 30 and 100
1m and 15m between 5 and 20
60m and 1440m between 20 and 55
3. chose a setting that looks reasonable. I like stop around 80, channel between 30 and 50.
4. run the setting in the week it was optimized for to see what the results look like. Were there a lot of trades but many of them losses? were there a few trades but all of them winners. (Prefer the latter) In general faster channels will produce more trades but a lot of them will be losses.
Use the setting from this optimization for the next week in a walk forward test.


The graphs show samples of testing results (which have been surprisingly consistent) with walk forward using this method:
1st graph is the backtest for the trades for 9/10/07 after optimization. 99%..
2nd graph shows trading results using the set from that test applied in a walk forward to the next week 9/17/07 to 9/22/07.

In further testing the success rate is very very good for this method.
Learning to know what setting from all the optimized settings is the trick.
It can be done. There are red flags, what to avoid and also what to look for.

I tend to get really excited and think wow this is it, we found it, the pot of gold at the end of the rainbow, so I don't want to fall into that with this, but this is the best I've seen yet!!!!

The success rate in walk forward is eye-popping!!

FXDD and Alpari tested the same for these settings.
EURUSD, EURGBP, AUDCAD, EURCHF (15min 99%)
For week of 11/15 thru 11/20

Trade today on new settings.
These settings will be intact for the rest of the week.
Then updated Friday close, and updated settings applied next Monday open...

Please always check the first page for most current settings...
and updates.

Experimental:
moving average strength indicator by Taiyakixz

Getting the data for the previous week is not going to work relying on Meta Quotes.
I really do not want to tackle php.. I envision computer crashes..
downloading from Dukascopy demo is easy. Only thing is a demo only lasts 14 days and they call you constantly on the phone, "open an account, open an account"...
Well I just may have to open an account. To be able to get that good 99% data for the previous week so that we can set up for next week... It just might be worth it. They don't offer mt4 or I would just switch altogether. (I hope they accept US citizens now that US is a 3rd world country).
$1000 to open. I really would rather do something else with the money but for that data....... Looks like I might have to. I'll keep opening demos and avoid answering the phone as long as I can....
We just need that previous week.
MetaQuotes has whole months of data missing.... And all sorts of errors and who knows what else, black holes.... 90% doesn't mean a thing....
You can run a test and get results but they're no good!!!
Only thing 90% is good for is to test the mechanics of the ea.

The new test setting for 15m has opened short EURGBP
must have been a fast moving market, the positions were opened far apart.
looks like AUDCAD has opened short too.
Not one of my top picks for a pair to trade, but it was one of the ones that I could get the 99% data for...

In further testing the success rate is very very good for this method.
Learning to know what setting from all the optimized settings is the trick.
It can be done. There are red flags, what to avoid and also what to look for.

I tend to get really excited and think wow this is it, we found it, the pot of gold at the end of the rainbow, so I don't want to fall into that with this, but this is the best I've seen yet!!!!

The success rate in walk forward is eye-popping!! Well maybe not that great. But we are seeing how to test this.

Please do not use a live account. This is under development only.

FXDD and Alpari tested the same for these settings.
EURUSD, EURGBP, AUDCAD, EURCHF, USDCHF, EURJPY(15min 99%)
For week of 11/15 thru 11/20
Attached Files http://cdn.forexfactory.com/images/attach/zip.gif 11-14eurusd15ma.zip (2.6 KB, 31 views) http://cdn.forexfactory.com/images/attach/zip.gif 11-14eurgbp15m.zip (3.1 KB, 6 views) http://cdn.forexfactory.com/images/attach/zip.gif 11-14audcad15m.zip (3.0 KB, 2 views)

sounds pretty cool
hope u guys 'll make it, dont have time at the moment to help to improve it
-.-

Taking a little break...
The tests are not proving successful..
Sometimes getting some distance for a little while helps to refocus...

Dragon may be resurrected.
We are working on a very novel strategy for the robot.
So let us see. I hesitate to post until there are some consistent results. More demo runs next week and hopefully we can post.
This is very fast scalping on the 1 min.

Quitting is not an option!!!