Forex Lasers Forum

Forex Lasers Forum (https://www.forexlasers.com/forums/)
-   Forex Systems (https://www.forexlasers.com/forums/forex-systems/)
-   -   Steve Hopwood's new AshFX V2 auto-trading robot (https://www.forexlasers.com/forums/forex-systems/382-steve-hopwoods-new-ashfx-v2-auto-trading-robot.html)

aharon.amir 26-10-2009 07:19

Quote:

Originally Posted by SteveHopwood (Post 5744)
Ignore them. They are strategy tester thingies that are nowt to do with the robot.

:)

Still,
I can't test your robot, I am getting only errors and no trades are done.
can u show me a configuration that work's with any pair?

aharon.amir 26-10-2009 07:26

Quote:

Originally Posted by SteveHopwood (Post 5744)
Ignore them. They are strategy tester thingies that are nowt to do with the robot.

:)

Still,
I can't test your robot, I am getting only errors and no trades are done.
can u show me a configuration that work's with any pair?

aharon.amir 26-10-2009 09:22

sorry,
Form resend error due to Internet connection failure.

SteveHopwood 27-10-2009 12:46

2 Attachment(s)
* Fixed a few bugs.
* Added ECN thingy - relevant input is CriminalIsECN. Set to 'false' if yours is not a 2-stage crim.
* Altered the screen update code so it only changes once a minute instead of every tick. This dramatically reduces the processor power eaten up by the robot.

:)

vff 27-10-2009 14:14

Quote:

Originally Posted by SteveHopwood (Post 5783)
* Fixed a few bugs.
* Added ECN thingy - relevant input is CriminalIsECN. Set to 'false' if yours is not a 2-stage crim.
* Altered the screen update code so it only changes once a minute instead of every tick. This dramatically reduces the processor power eaten up by the robot.

:)

Hello Steve and thanks for the new update.One question, in post #78, you said you added code for slippage? Is this hard coded, or is an input you can configure? I don't see a parameter for it?
Thanks again and have a nice day!

Decided to add to this post instead of posting another.I see a possible problem in using the longer time frames such as 4hr or especially on the daily.
If using the daily tf and you have let's say five pending trades ready to fire off at new candle.When new day starts, there is a possiblility you will get the trade did not happen due to busy context.I know I have the wording wrong, but I think this happens when you are trying to send more than one trade request at a time.
This just happened to me while watching the ea.One trade took place, the other did not.
Wonder if there is a way to pause between sending trade requests so that not all are sent at the same time?
Have a nice day!
Vince

SquareRoot 27-10-2009 18:15

Quote:

Originally Posted by SteveHopwood (Post 5783)
* Fixed a few bugs.
* Added ECN thingy - relevant input is CriminalIsECN. Set to 'false' if yours is not a 2-stage crim.
* Altered the screen update code so it only changes once a minute instead of every tick. This dramatically reduces the processor power eaten up by the robot.

:)

Why update every minute? Why not every hour or end of day?

SteveHopwood 29-10-2009 09:28

2 Attachment(s)
Over at FF, we found a few bugs. These are squished in the latest version. There is also a lot of code added to show why impending trades have been cancelled.

:)

Chills 29-10-2009 11:45

Hi Steve,

I don't know what I am doing wrong but the robot is not taking any trades at all.
As I have not used something like this before I am not sure how it works. I wondered if it was because it needed to be on 24/7 so the past couple of days I have done this but there are still no trades even though there seems to be a couple on my FXDD demo eg gbpjpy.

I ticked 'allow live trading' on the common tab. What else could I be missing?

I haven't downloaded your latest as if I can't get it working what is the point?!

Any suggestions?

Thanks heaps.

SteveHopwood 29-10-2009 11:53

Quote:

Originally Posted by Chills (Post 5848)
Hi Steve,

I don't know what I am doing wrong but the robot is not taking any trades at all.
As I have not used something like this before I am not sure how it works. I wondered if it was because it needed to be on 24/7 so the past couple of days I have done this but there are still no trades even though there seems to be a couple on my FXDD demo eg gbpjpy.

I ticked 'allow live trading' on the common tab. What else could I be missing?

I haven't downloaded your latest as if I can't get it working what is the point?!

Any suggestions?

Thanks heaps.

Yep. Read the documentation, especially the stuff about candle wick lengths; you will probably find the answers you need.

I leave the robot on 24/5 but it probably isn't necessary unless it is managing trades as well as looking for them. If you do not require it to do any stop-loss managing of open trades, then you can probably leave it on for about an hour either side of the start of the new candle. Leaving it running has little impact on your computer's processor use.

'Tis up to you whether you download the latest version or not. The previous one threw up a couple of bugs, so you are currently running a slightly faulty one. The bugs are not serious, but they will drive you nuts if they surface.

:)

vff 29-10-2009 14:18

Post #8
 
Hello Steve and thank you for the new update.I was wondering if you saw my post #85 reguarding trade context is busy message? When you have so many possible pending orders and if they all trigger on next candle, this causes the server to send busy context message therefore, there will be some trades that should trigger will not trigger.
Also which is the main forum to post for the ea? This one or on FF?
Thanks again for the new update.
Vince

SteveHopwood 29-10-2009 15:09

Quote:

Originally Posted by vff (Post 5851)
Hello Steve and thank you for the new update.I was wondering if you saw my post #85 reguarding trade context is busy message? When you have so many possible pending orders and if they all trigger on next candle, this causes the server to send busy context message therefore, there will be some trades that should trigger will not trigger.

I saw the first part of your post but missed the edit. Given that the robot can only send trades after the first tick of the new candle, this shouldn't be a problem. Let's see how it goes and I can always code in some sort of delay if need be.

Trades are more likely to be cancelled because one or more of the trading conditions fails. In this case, there will be an alert saying why.

I had an idea whilst I was tying the above. If the pending trade passes all the filter tests but the send fails because of, for example, trade context busy, I can code in a 10 second sleep that will shut the robot down. After it wakes up, it will be waiting for the next tick as usual and can try again. I will do this later (actually, I just did it, just can't be bothered uploading it right now).

Quote:

Also which is the main forum to post for the ea? This one or on FF?
Thanks again for the new update.
Vince
I upload updates to both. I started here because I assumed most interested traders would find their way here. I started a thread in FF in case there is a wider interested trader-base, and this turns out to be the case.

:)

vff 29-10-2009 15:41

Hello Steve..
Sorry about the edited message, I was trying to conserve space.Guess in the future it will be better just to do new posts.
Thanks for adding the sleep function.I had it on a 15m tf when I saw on the next tick that both trades were trying to trigger at same time and one made it and the other one got the trade context message.On lower time frames this won't be too bad since you can wait 15m etc., but on a daily, this could be problem if all try and none make it due to the trade busy context.But as you said you already fixed the problem.
I want to stress again how much I appreciate all the work you do.It is a gift being able to code.I spoke to you before about coding and I found the web site with java and tried to do some coding with that.That is also a great program.I am trying to figure out how to add items such as breakeven etc.
Also, I forgot to ask about the slipage again.Is that hard coded or can it be adjusted? I didn't see where it can be adjusted?
Thanks again and have a great day!
Vince

SteveHopwood 29-10-2009 16:06

Quote:

Originally Posted by vff (Post 5856)
Also, I forgot to ask about the slipage again.Is that hard coded or can it be adjusted? I didn't see where it can be adjusted?
Thanks again and have a great day!
Vince

Hard coded. You can have it as an input if you want.

:)

vff 29-10-2009 16:34

Ok Steve thanks and not sure if it is needed as an input?
Here is a question I have which I would like to ask?
I am testing the new ea on a 15m frame.I am just using the majors for the test and I guess just using four pairs could be why I am seeing what I am seeing.I see the four pairs saying psar flip and all the conditions are great but it is waiting for the next candle to confirm.Next candle comes and trade is not taken for reasons like wick too long ac color red etc. so the trade does not occurr due to the fact at new candle it did not meet all the conditions it did meet when you did have psar flip or at a time after psar flip and new candle opens.You did state that psar flip can occurr at any point on the candle even on a daily one.I would think each of the parameters are also on tick per tick bases?
Wonder what the probability of a trade never happening unless we have a strong up trend or down trend?
On a daily chart it could happen when no trade might occurr since all the parameters may not line up? But I guess if you are using all the pairs available to you that one of those should open a trade?

SteveHopwood 29-10-2009 17:24

Quote:

Originally Posted by vff (Post 5858)
Ok Steve thanks and not sure if it is needed as an input?
Here is a question I have which I would like to ask?
I am testing the new ea on a 15m frame.I am just using the majors for the test and I guess just using four pairs could be why I am seeing what I am seeing.I see the four pairs saying psar flip and all the conditions are great but it is waiting for the next candle to confirm.Next candle comes and trade is not taken for reasons like wick too long ac color red etc. so the trade does not occurr due to the fact at new candle it did not meet all the conditions it did meet when you did have psar flip or at a time after psar flip and new candle opens.You did state that psar flip can occurr at any point on the candle even on a daily one.I would think each of the parameters are also on tick per tick bases?
Wonder what the probability of a trade never happening unless we have a strong up trend or down trend?
On a daily chart it could happen when no trade might occurr since all the parameters may not line up? But I guess if you are using all the pairs available to you that one of those should open a trade?

The robot is coded to be as accurate a representation of Ash's system as possible. Timing of looking to send a trade etc are in accordance with my understanding of his system.

Few trades survive the rigours of the selection process. You need a lot of open charts to have any chance at all of taking any trades.

Having said all that, Ash is trying to provide us with a trading tool, not a rigid system whose traders must do such-and-such at exactly the right moment. It is an easy thing to do to add a TradeAggressively input that will make the robot take trades the moment the indis line up - might be fun to try this and see how well it works.

Leave it with me.

:D
:D

vff 29-10-2009 17:50

Quote:

Originally Posted by SteveHopwood (Post 5860)
The robot is coded to be as accurate a representation of Ash's system as possible. Timing of looking to send a trade etc are in accordance with my understanding of his system.

Few trades survive the rigours of the selection process. You need a lot of open charts to have any chance at all of taking any trades.

Having said all that, Ash is trying to provide us with a trading tool, not a rigid system whose traders must do such-and-such at exactly the right moment. It is an easy thing to do to add a TradeAggressively input that will make the robot take trades the moment the indis line up - might be fun to try this and see how well it works.

Leave it with me.

:D
:D

Hello Steve and don't get me wrong. I appreciate it very much what you and Ash have done. Also, I do not what to have a trade be made just to have a trade. As you state the trade has to be merit of a trade. But again let us use a usd/jpy pair on a daily tf. Let us say your server has a 7 pm new candle. Ok it is 9 am the usd/jpy pair is at 80.70 and all the parameters are set.You will need to wait until 7:01 pm to confirm a trade to happen. At 7:01 one of the parameters for a trade to happen does not happen and trade does not take place. At that time the pair is now at 70.10. What if you have all the pairs available to you open and the above happens on all 27 pairs just for a click at 7:01 the trade does not happen.
I just started trading the ea so my logic might be off. You and the others have more experience as I do so would know better than me how the ea should work.
I did put the ea on all pairs available and also changed from 15m to 30 min tf, and same is happening so far?
Please don't take my comments as being negative, I appreciate all the hard work that is put into the ea.

Emman95 29-10-2009 22:40

Thank you Steve
 
Quote:

Originally Posted by SteveHopwood (Post 5783)
* Fixed a few bugs.
* Added ECN thingy - relevant input is CriminalIsECN. Set to 'false' if yours is not a 2-stage crim.
* Altered the screen update code so it only changes once a minute instead of every tick. This dramatically reduces the processor power eaten up by the robot.

:)

Thank you so much Steve for your efforts. I have downloaded the robot and will test it out for a week before posting any comment on it. Once again, thank you.

SteveHopwood 29-10-2009 23:05

Quote:

Originally Posted by Emman95 (Post 5868)
Thank you so much Steve for your efforts. I have downloaded the robot and will test it out for a week before posting any comment on it. Once again, thank you.

My pleasure.

Not being able to edit post 1 here so I can always put the latest version there is a real pain. Make sure you have the latest version - from http://forexmoments.com/forums/syste....html#post5844

:D

SteveHopwood 30-10-2009 17:38

I have a GJ sell that retraced far enough to kick in the first tranche of hedging. Now the market has resumed its downward movement, there is every chance the hedged pair will close out at breakeven.

My StartHedgingAtLossPips is not flexible enough - hedging kicked in a lot of pips before it would have hit a stop loss.

I will think about how to code a StartHedgingAtLossPips setting that calculates the value of this setting automatically and sets it at the same figure as a stop loss, had there been one.

:D

swissforex 30-10-2009 22:42

candlesticks
 
Hi Steve
thanx for the EA and Support

no entries from the EA in the first 2 days

I've read carefully your Manual,but I didn't understand 2 points about the CandleSticks

1-If I don't want the EA trades whit previous candlestick bigger then 100 pips,what value should I set?
2-Could u explain the point"CandleLenghtOverAverage"??I didn't understand AT ALL :(

10x

SteveHopwood 30-10-2009 23:09

Quote:

Originally Posted by swissforex (Post 5892)
1-If I don't want the EA trades whit previous candlestick bigger then 100 pips,what value should I set?

You can't nor should you. This is irrelevant.

Quote:

2-Could u explain the point"CandleLenghtOverAverage"??I didn't understand AT ALL :(
10x
Ash does not take trades if the trigger candle is too long compared with the average candle length.

He does this, I assume, by eyeballing the chart and seeing that the trigger candle appears to be too long when compared to the others - a great example of his 'feel' for the markets.

A software robot cannot be given this 'feel'. I have coded it to look back 100 bars and calculate the average length of each candle over this period. The CandleLenghtOverAverage input allows you to say how much longer than this average the trigger candle can be before the trade is cancelled.

For example, suppose the average candle length is100 pips. CandleLenghtOverAverage is a percentage so, with an input of 10:

Average candle length = 100 pips, so the trigger candle can go to 110 pips before the trade is cancelled.

With an input of 30:
Average candle length = 100 pips, so the trigger candle can go to 130 pips before the trade is cancelled.

Hope this makes it clear.

:D

swissforex 31-10-2009 22:16

now it's more clear
 
Quote:

Originally Posted by SteveHopwood (Post 5894)
You can't nor should you. This is irrelevant.


Ash does not take trades if the trigger candle is too long compared with the average candle length.

He does this, I assume, by eyeballing the chart and seeing that the trigger candle appears to be too long when compared to the others - a great example of his 'feel' for the markets.

A software robot cannot be given this 'feel'. I have coded it to look back 100 bars and calculate the average length of each candle over this period. The CandleLenghtOverAverage input allows you to say how much longer than this average the trigger candle can be before the trade is cancelled.

For example, suppose the average candle length is100 pips. CandleLenghtOverAverage is a percentage so, with an input of 10:

Average candle length = 100 pips, so the trigger candle can go to 110 pips before the trade is cancelled.

With an input of 30:
Average candle length = 100 pips, so the trigger candle can go to 130 pips before the trade is cancelled.

Hope this makes it clear.

:D

Thank u for your good explanation..Now it's really clear how I should set that parameter.

So if I wanna check the last 50 candles, I've to put 50 value in "CandleLenghtLookBackBars" box,and input into the box below(CandleLenghtOverAverage) the percentage I'd like to enter

About S\R
I've traced 4 SR Lines on each chart..How many lines can handle your EA??

About jumping StopLoss..what does it mean "jumping"?I just can't figure out

10x a lot

this is really a good thread

have a nice Sunday!

SteveHopwood 01-11-2009 20:56

1 Attachment(s)
Added the option to turn off trading during the Sunday candle, for those who would otherwise get caught out by DST. If your broker does not have a Sunday candle, or you do not mind trading the Sunday candle, then you do not need this update.

:)

forexforex3 02-11-2009 17:59

forward test results?
 
Hello,

Has anyone used this system for forward autotrading? How is the performance?
thank you.

moondog 03-11-2009 18:51

Quote:

Originally Posted by forexforex3 (Post 5928)
Hello,

Has anyone used this system for forward autotrading? How is the performance?
thank you.

Yes, Some of us have,
a) I have completed backtesting, scenario testing and analysis of 26 pairs over the period 01/01/08 to 18/09/09.
b) "Demo" forward testing, on and off, the same 26 pairs.
c) "Live" trading 6 preferred pairs, using Steve's Robot, since late September, and using two brokers. (and in profit)

The strategy works and the robot works.

Can post details of preferred pairs and other related stuff if anyone interested.

Moondog

padman 03-11-2009 19:39

VERY interested in your results and best pairs please moondog

Om_Salat 04-11-2009 11:58

Quote:

Originally Posted by moondog (Post 5957)
Yes, Some of us have,
a) I have completed backtesting, scenario testing and analysis of 26 pairs over the period 01/01/08 to 18/09/09.
b) "Demo" forward testing, on and off, the same 26 pairs.
c) "Live" trading 6 preferred pairs, using Steve's Robot, since late September, and using two brokers. (and in profit)

The strategy works and the robot works.

Can post details of preferred pairs and other related stuff if anyone interested.

Moondog

Hi Moondog,

Yes, I'm very interested! I'd like to know your 6 preferred pairs and your settings for each of those pairs.

TIA.

Salat

Egwig 04-11-2009 14:26

Originally Posted by Swissforex

About jumping StopLoss..what does it mean "jumping"?I just can't figure out

10x a lot

this is really a good thread

have a nice Sunday![/QUOTE]

Hi Swiss.

A jumping stop loss only moves after the price has traveled a set number of pips in the right direction.

For example, you could set the jump to be 10 pips then if the price moved 10 the stop would jump,

It would not move again till the price had gained another 10 pips.

A normal trail will move with each and every pip that the price gains.

Clear?;)

Egwig:cool::cool:

vff 04-11-2009 14:31

Settings
 
3 Attachment(s)
Hello everyone.
For this week testing, I am using all the pairs available on a 1hr tf. I am using a 1hr tf for the reason of troubleshooting. I have the followings pics and Steve, I wonder if the settings of the ea I have are not correct? I would think since I have jump after 50 pips a stop would occurr and a partial close would happen on the gbpjpy trade? Also still getting a lot of requote and buy context messages that are killing some trades that should of triggered.
Also in the second picture not sure why I am getting those errors. I did close the trades manually.
Dad had to go to Hospices so I am running back and forth from there and home.

Have a nice day everyone.
Attachment 1109
Attachment 1110
Attachment 1111

moondog 04-11-2009 16:09

1 Attachment(s)
Quote:

Originally Posted by padman (Post 5959)
VERY interested in your results and best pairs please moondog

Greetings Padman & Om Salat,

Here you go,

See attached gif for extracts from the a/c history statement.

My other a/c is too messy and not representative of continuous auto trading with Steve's Robot, and my demo accounts are even worse due to testing and experimental issues, so I feel that looking at the other data will cloud and confuse the issue here, so lets focus on the cleaner simpler example and scale up thoughts instead of the other way around.

All the yellow stuff is my data/labels and all the other numbers are extracted from my recent MT4 stats trading statement.
Because of the way the statement and trades are accounted for I have included a trade ref#. to tie up and locate the split transactions.

NOTE:
The following is strictly what I use for D1 pairs Auto Trading only. Due to the differing dynamics between the D1 and H4 timeframes, I will not trade the H4 on Auto until testing of the H4 is completed and similiar results are available. Successful settings on D1 do not automatically work on H4. The first pass H4 auto trading tests did not look good to me, so I expect a lot of data/scenario filtering and optimisation work (per pair) will be needed to identify final H4 Auto Trade pairs.

The overall D1 test period was 01/01/08 to 18/09/09. (26 Pairs)
Preferences (best 6 pairs) from period 28/12/09 to 18/09/09
Timeframe Daily (D1) only

Auto Trading - Preferred Pairs showing DD Budget.
Extract of findings/suggestions :-

For 200.00 Account and Total Lots traded = 0.02
Drawdown Budget
-----------b)----a)
GBPJPY 21 - 28%
EURJPY 26%
AUDJPY 32 - 41%
USDJPY 28%
AUDNZD 13 - 15%
GBPUSD 46 - 53%

For 500.00 Account and Total Lots traded = 0.02
Drawdown Budget
------------b)-----a)
GBPJPY 9 - 12%
EURJPY 10%
AUDJPY 12 - 16%
USDJPY 12%
AUDNZD 5 - 6%
GBPUSD 18 - 21%

and so on........

I have used the following core settings a) and b) successfully across both accounts.
The example gif attached only used a) below and no advanced trading tools enhancements were included as yet (Jumping SL/Hedging/Big Numbers/SR etc):-

a) An AshFxV2 Native Mode (test ref 12) where PartClose is enabled and Close Lots and Preserve Lots are split equally. (All trades in statement are in this mode.)

b) An AshFxV2 Special Mode (test ref 8) where PartClose is enabled but no Preserve Lots (a take all the money and run mode on first profitable bar)

As you can see from the above you have to "plan" for the DrawDown dynamics as well as the fixed StopLoss.

The lower DD figure applies to b) and the higher one to a), where there is only one figure then DD is same for a) & b).
(I should have actually shown the number of trades and profits/loss details as well....can expand above info in this regard if there is enough interest)

For multiple trade scenarios, extra caution is needed to enable a mixed basket, as the current list has a 66% bias to the JPY.

There are merits to using the differing modes a) and b) with the robot depending on the dynamics of the pair selected, account balance and other objectives.

Remember that the above preferred pairs are in my opinion "safe" selections for Auto Trading purposes only. (obviously good for manual trading as well).

---------------------------------------------------------------------------------------------------------------------------
Final Caveat/Comment,
I have three rules (short version) that guide me, in all aspects of all things.....

1) Do not believe everything that you read.

2) Do not blindly rely on what others ("experts") tell you to do or advise etc.

3) Always do your own research and use any other information gained as a contribution to your own analysis/decision process.
---------------------------------------------------------------------------------------------------------------------------

If you need other info settings etc to get started, just ask and I'll see what I can dig out.

Moondog

aharon.amir 09-11-2009 10:57

Help Please
 
Hi,
I am trying to test the EA, but I always get the error:
"2009.11.09 12:55:15 TestGenerator: unmatched data error (volume limit 93201 at 2009.05.19 00:00 exceeded)
"
Any Ideas?

forexforex3 12-11-2009 11:10

Thank you!
 
Thank you for sharing your experience using the expert advisor! :)

wrp123 14-11-2009 02:06

Hi Moondog

I read your post about the perfomance of the system with great interest - excellent work!

I dont intend to following the system blindly as an EA, but as a manual signal. Would it be possible for you to provide further performance information for example an excel export of the back test for a few of the major pairs. This would be incredibly helpful for me in helping to assess the system's power.

I realise this might be a big task so no worries if you have no time.

Kind regards

Dave

padman 14-11-2009 14:14

Thanks for the stats moondog - very much appreciated :)

moondog 16-11-2009 09:28

Quote:

Originally Posted by wrp123 (Post 6121)
Hi Moondog

I read your post about the perfomance of the system with great interest - excellent work!

I dont intend to following the system blindly as an EA, but as a manual signal. Would it be possible for you to provide further performance information for example an excel export of the back test for a few of the major pairs. This would be incredibly helpful for me in helping to assess the system's power.

I realise this might be a big task so no worries if you have no time.

Kind regards

Dave

Thats ok,
I am a bit busy mon/tue this week, so give me a couple of days and I will post some of the auto trading back-test excel data around weds.
Moondog

wrp123 16-11-2009 20:08

Brilliant

Many thanks Moondog

padman 17-11-2009 16:17

Started to trade the robot on 4hr trades this week in demo mode.

Taken so far this week, euraud and audcad both long and still open.

I will start a new thread showing my 4hr trades if that is OK with Steve.

Padman

rewing 18-11-2009 13:47

Robot not taking trades
 
Steve.... I have had your robot on H4 timeframe for almost 2 weeks now, and I have had no trades. I keep getting these messages a lot

2009.11.18 03:57:13 AshFX V2 auto-trading robot by SteveHopwood USDCHF,H4: Alert: USDCHF Short cancelled: Trigger candle was too long

It appears the setting for CandleLengthOverAverage is the setting controlling that. I think default is 10, & I have raised it to 50 & still no trades

BTW:I am using the robot posted on #103. Is that the one I should be using?

Can you advise?
Thanks
Ron

moondog 19-11-2009 14:14

1 Attachment(s)
Quote:

Originally Posted by wrp123 (Post 6121)
Hi Moondog

I read your post about the perfomance of the system with great interest - excellent work!

I dont intend to following the system blindly as an EA, but as a manual signal. Would it be possible for you to provide further performance information for example an excel export of the back test for a few of the major pairs. This would be incredibly helpful for me in helping to assess the system's power.

I realise this might be a big task so no worries if you have no time.

Kind regards

Dave

========wrp123==========

Here it is.
Apologies for lateness, just arrived back today and settled.
Attached is the test data table mentioned for the D1 Ash Strategy using Steve Hopwood's Robot.

TEST OBJECTIVES:-
a) Finding which pairs were potentially tradable on Auto on the D1 Timeframe from 26 available.
b) Finding which settings/scenarios (modes) best suited a particular pair and provided the lowest drawdown and/or most profit.

NOTES:
1) These results are for (my) Auto Trading preferences only and not a recommendation for anyone to do anything especially wth your live accounts.
2) The Ash strategy works across most pairs using manual filtering and selection methods, something Ash is very good at with all his experience and knowledge but something that is difficult to emulate in any algorithm.
(However, an interesting observation was that when all the P/L results were combined for all 26 pairs the account we still ended in profit, but that is not a practical result and does not take into account Balance , DD and Margin.
3) The Robot is written by a veritable Software Maestro and one that understands trading as well, the Bot works very well and has lots of features to be explored and tested in the future to discover more Auto tradeable pairs outside this intial list.

TEST LIMITATIONS:-
The attached analysis was done on broker demo accounts and the dynamics will not the same as broker live, so forward live results will differ somewhat.
Although testing was done in various blocks over nearly two years of data the attached table representas a short period 28/12/08 to 03/10/09.
Core testing only, without any of the advanced feature enabled such as Money Managment, Jumping Stop Loss and S/R etc.

EXPLANATION OF MODES:-
Account Balance chosen was 200 as this was considered the rock bottom along with 0.02 lots total in trade, then its easier to scale up the account balance and reduce the DD as required.
So all DD% results in the table are related to these 200 / 0.02 factors.

Mode 4 - PartClose = False, with Close Lots and Preserve Lots also set to Zero ! What this does is let the trade run until the A0 reversal signal.

Mode 8 - Part Close = True. This will close all profits on first available profit bar.

Mode 12 - Native mode for the Ash Strategy where PartClose = True and the Lots are equally split.

Mode 16 - Special Mode (8) where both Stoch_Main or Stoch_SIgnal was used to filter the trade, unlike Modes 4, 8, 12 where only Stoch_Signal was used.

The DD% used to filter the data table is 30%, so results that are negative (loss) will show up as Red and any DD% greater than 30% will show up also a Red.

INTEPRETATION:-
The top 5 or 6 pairs (Blue) were also found to be the most robust on Auto from other tests going back over longer period 01/01/08 - 03/10/09.

*** What you are looking for in each Mode column is a pair of black entries with a reasonable DD% and profit. (not a 26%DD chasing an $8 profit) preferably in the Mode 8 and Mode 12 colunms supported by profit in the Mode 4 and Mode 16 columns.

Thats all for now folks,

Moondog.

wrp123 20-11-2009 22:48

Brilliant

Many thanks


All times are GMT. The time now is 01:16.

Powered by vBulletin® Version 3.8.10
Copyright ©2000 - 2021, vBulletin Solutions, Inc.