Jiva Dragon | Page 33 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Enter The Dragon - GBP/JPY 2 replies

Info on the Borihei-Dragon Scalping System? 15 replies

Jiva's 5,5 Trade Journal 5 replies

The Dragon 441 replies

***Dragon II*** 15 replies

Attachments: Jiva Dragon
Exit Attachments

Jiva Dragon

  • #641
  • Oct 28, 2010 9:39am Oct 28, 2010 9:39am
well, I tryied those settings at three brokers and I didnt come close to your results. The best was 54.43 profit trades.
  • #642
  • Oct 28, 2010 12:41pm Oct 28, 2010 12:41pm
Is anyone trading this on a live account? or is this just all still in demo testing. Was considering putting it on micro account ..
  • #643
  • Oct 28, 2010 1:47pm Oct 28, 2010 1:47pm
Quoting tommycz
Is anyone trading this on a live account? or is this just all still in demo testing. Was considering putting it on micro account ..
well, I would be very careful with that. This EA have large stops and need more demo testing in my opinion. Jiva is all the time making improvements and optimizing settings.

I would like to know what Jiva think about this
  • #644
  • Oct 28, 2010 2:27pm Oct 28, 2010 2:27pm
Quoting tommycz
Is anyone trading this on a live account? or is this just all still in demo testing. Was considering putting it on micro account ..
It's tempting!
But I am not putting it on live account, not even micro, for myself yet. If that says anything.
I really want to have the risk management under control. There are still too many loose end details. Details. Details. Details...
Mistakes are expensive.
So, yes, I am confident of the basic strategy. Still just a little too developmental yet for the big plunge.

Maybe when jm gets fired up that'll be the go ahead.
  • #645
  • Oct 28, 2010 2:53pm Oct 28, 2010 2:53pm
Quoting jm2110
You illustrated my point beautifully here Jiva. The only reason you made money on this trade was that your SL was further from your entry than your take profit level.

In fact the entry signal was pretty damn bad due to the fact that very quickly after you put the trade on it went significantly against you....in fact if you had been given a sell signal there we would have been thinking "damn that thing is good". But all's well that ends well and as I said playing the statistical game due to the stops being further away than the profit levels...
Question: I am assuming that it is permissible in the "rules" to optimize for the longer time period back test for 50-50.
Only thing not allowed is trailing stop and profit and stops must be same value.
The actual values of the other settings are variables and can be optimized for the longer period backtest whereas the basic strategy remains the same as long as the stops and limits are of equal value and no trails.
Am I correct in assuming this?
  • #646
  • Edited 7:41pm Oct 28, 2010 3:38pm | Edited 7:41pm
Quoting inso
well, I tryied those settings at three brokers and I didnt come close to your results. The best was 54.43 profit trades.
The other settings, like channel period and moving averages, may need to be adjusted for the different brokers. This is a pretty sensitive setting using the 5 min.

update: I posted a note on the original post.
this test was done with v14 which no one has yet. I got a little confused between the volumes. v14 has macd so v11 will not give the same results for the test. So basically no else can test my test.

oops..
  • #647
  • Edited 6:31pm Oct 28, 2010 3:49pm | Edited 6:31pm
Quoting jm2110
Exactly.

Make sure when you are looking at macd you are using a decent macd indicator and not Mt4's in built one.

See here for discussion
forexfactory.com/showthread.php?t=69409
According to my understanding there is no significant difference in the calculations of the mt4 macd, it is simply presented differently visually.I n mt4 the osma histogram is treated as it's own separate indicator and is not present on mt4 macd.
In mt4 they have represented the ma line as a histogram rather than a line. It is a graphic difference only.
The calculations for signal line and macd crossing are not different. That point becomes clear towards the end of the discussion.

http://www.forexfactory.com/showpost...0&postcount=16

Our invaluable technician, mathematician, programmer, grounded in common sense Taiyakixz has researched the macd question: His response is:

I checked MACD True, the signal line is different from mt4 standard macd.
But we should not select right one, but usuful one.
If you like mt4 standard macd, I think that there is no problem.

So there we have it, if trying to do this has not already revealed to us that there is no such thing as stable matter, but rather EVERYTHING is in motion, is indeed dynamic and can only be considered as different types of energy, well then we must be dead.
  • #648
  • Oct 28, 2010 4:45pm Oct 28, 2010 4:45pm
There has been a confusion, I am afraid I have created it..
I am good at doing that. I don't even have to try. It comes naturally.
The testing for the 5 min posted earlier to identify a statistical edge for 50/50 stop loss and limit was done on a version of the Dragon no one else has yet.

v14 has the macd filter but we are not ready to post it yet. My mistake. So no one will be able to test my test unless they have v14 and no one has v14 except for Taiyakixz and myself.

Dumb, yes, but I am getting used to it.
  • #649
  • Oct 28, 2010 6:07pm Oct 28, 2010 6:07pm
maybe i could help somehow? I m newbie in trading but maybe I could use my two vps for optimization settings, if you want
  • #650
  • Oct 28, 2010 6:26pm Oct 28, 2010 6:26pm
Quoting jiva34
Question: I am assuming that it is permissible in the "rules" to optimize for the longer time period back test for 50-50.
Only thing not allowed is trailing stop and profit and stops must be same value.
The actual values of the other settings are variables and can be optimized for the longer period backtest whereas the basic strategy remains the same as long as the stops and limits are of equal value and no trails.
Am I correct in assuming this?
They aren't really "rules" more common sense I think. If you think about trying to gauge how accurate your entries are then you want your SL and TP to INITIALLY be the same distance from the entry price. Personally I do think it is fine if you want to add a stoplossflat variable that once the market moves towards your TP level you move your stop to the entry price (or entry price + a few pips profit if you prefer)as in this case you are moving the Stop closer to the entry rather than having it further away and if that makes your strategy have a higher percentage of profitable trades then great as it does show that your entry still has an "edge".

As I said this has nothing to do with the finished variables we will use when actually trading the EA it is just a process to go through to gauge whether a change in the entry logic (macd, rsi , dss etc etc) is having any material effect on the accuracy of the entries as otherwise when we optimize while moving around the SL and TP levels we are adding other statistically significant parameters into the mix.
  • #651
  • Oct 28, 2010 6:27pm Oct 28, 2010 6:27pm
Quoting jm2110
They aren't really "rules" more common sense I think. If you think about trying to gauge how accurate your entries are then you want your SL and TP to INITIALLY be the same distance from the entry price. Personally I do think it is fine if you want to add a stoplossflat variable that once the market moves towards your TP level you move your stop to the entry price (or entry price + a few pips profit if you prefer)as in this case you are moving the Stop closer to the entry rather than having it further away and if that makes your strategy have...
gotcha...
  • #652
  • Oct 28, 2010 6:37pm Oct 28, 2010 6:37pm
Please always check the first page for the most recent updated set files.

Updated AUDUSD 1h optimized FXDD 10/28/10
Attached Image (click to enlarge)
Click to Enlarge Name: audusdnoloss10-28.gif Size: 8 KB
Attached File(s)
File Type: zip audusdcons10-28 - Copy.zip   1 KB | 244 downloads
  • #653
  • Oct 28, 2010 6:52pm Oct 28, 2010 6:52pm
I think the MT4 MACD is just missing one line.

In it's originally devised form MACD should have a MACD and Signal line and then a histogram showing the spread between the two. Most MACD based trading logic (for example those strategies developed by Boris Schlossberg) will be referring to the slope of this spread between the macd and signal and when this spread crosses 0 being a guage of bullish and bearish behaviour and when we see lower peaks or higher troughs signalling bearing or bullish divergence.

As Taiyakixz said it is a matter of preference so when comparing if the MACD is a successful filter for the Dragon I would test both the MT4 MACD and then the histogram of the MACD True to see which is a better filter as technicians long before us have decided that it is this spread that is significant.
  • #654
  • Oct 28, 2010 9:34pm Oct 28, 2010 9:34pm
Here is a problem, jm, for trying to test for the 50/50 stop and limit probability edge.
The Dragon strategy is based primarily on price action in relationship to a hi/lo moving average channel.
The channel itself is a sort moving average and the value of the channel will change in relationship the market conditions.
The problem is: as market conditions change the value of the hi/lo moving average channel will also change. For instance in some market condition a 55 channel will be the significant value for the channel. In another market condition a 34 hi lo channel will be the significant value. So the strategy does not change but because conditions are fluid and dynamic the value of the channel will need to change within the ea. To try to test for 50/50 over say 5 years will not be realistic proof of a statistical edge or lack thereof for just one setting of a channel value.
I am still in favor of regular and frequent optimizations, starting with channel values as opposed to zooming out and trying to see all the activity over a large amount of time as one chunk.

Another way to try to see the situation: In another strategy that relies on candlesticks, say go long when there is bullish harami supported by macd. Just hypothetical of course. Bullish haramis and macd's are pretty consistent. There is no setting that needs to change with changing market conditions. So something like that could be tested for 50/50 over 10 years even.
But for the Dragon it will be necessary to adapt settings for the moving average hi/lo channel on a regular basis. This is not a shortcoming, indeed it may be a strong point. That it has that ability to adapt to the market.
there has been some suggestion made earlier about automated optimizations. I think that may be more relevant for the Dragon considering this fluid aspect than a rigid 50/50 test. Although 50/50 should still have some value if done over a short period of time.
Here is a very simplified 50/50 test for 12 months for EURUSD with only a few parameters, basically the channel and moving averages.
It tested +66.67%. FXDD This time I used v11.
Attached Image (click to enlarge)
Click to Enlarge Name: eurusd50-50b.gif Size: 8 KB
Attached File(s)
File Type: zip eurusd50-50b.zip   1 KB | 228 downloads
  • #655
  • Oct 28, 2010 10:55pm Oct 28, 2010 10:55pm
I understand what you say but I think if you could find a reason why the variables need to change over time (something like their relationship to atr for example) then you could build this logic into the ea so they are dynamically reacting to market conditions. If you see the only reason for them to change is because the optimization says so we are back to curve fitting on historical data that may be completely different in the coming week/month.
  • #656
  • Oct 28, 2010 11:45pm Oct 28, 2010 11:45pm
Quoting jm2110
I understand what you say but I think if you could find a reason why the variables need to change over time (something like their relationship to atr for example) then you could build this logic into the ea so they are dynamically reacting to market conditions. If you see the only reason for them to change is because the optimization says so we are back to curve fitting on historical data that may be completely different in the coming week/month.
The monitoring of multi time frame moving averages should be sufficient to identify these changes. If you will notice, a trade is only considered when the moving averages on mult time frames are in alignment. That alignment is significant. The value of the channel being correlated with the intensity of the trend. I believe it is acceptable to optimize for the these values and it is not simply curve fitting. If optimization for these parameters is done regularly and frequently the ea will be making decisions following the natural flow of the market. This strategy allows the ea to only enter trades in the direction of the trend in strongly trending markets. Which also means using a large stop loss is not as risky as it may seem at first glance since the trade is being entered in the direction of an established trend.
  • #657
  • Oct 28, 2010 11:57pm Oct 28, 2010 11:57pm
EURUSD could be setting up short... would need to pull back a bit more with a close above channel low for ea to take a trade.
Attached Image (click to enlarge)
Click to Enlarge Name: eu.gif Size: 32 KB
  • #658
  • Oct 29, 2010 3:10am Oct 29, 2010 3:10am
Please always refer to the first page for the most rent updated set files.

AUDUSD 1hr optimized for Alpari 10/29/10
Attached Image (click to enlarge)
Click to Enlarge Name: audusdalparina10-29.gif Size: 7 KB
Attached File(s)
File Type: zip audusdalpari10-29.zip   1 KB | 272 downloads
  • #659
  • Edited 5:35am Oct 29, 2010 4:48am | Edited 5:35am
jadefx strikes again

http://img178.imageshack.us/img178/4707/jade3c.jpg
  • #660
  • Oct 29, 2010 8:02am Oct 29, 2010 8:02am
Hello Jiva34, sorry I have been too bussy with school, and other thing a can not upload anymore pictures, my profile says I have exceeded uploading quota, do you know how can I increase this storage size? .

Is Jiva-Dragon EA making more trades now ? in the picture seems to be 2 GU and 1 EU, very good floating numbers

Thanks Jiva34

___________
Sequoia
0 traders viewing now