scooby-doo BOJ JPY pairs trading robot | Page 340 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Trading correlation pairs by using the other pairs 843 replies

Export All pairs or selected pairs to .csv with script 3 replies

Pairs of Currency Pairs 4 replies

Trending Pairs / Ranging Pairs 0 replies

Robot trading 120 replies

Attachments: scooby-doo BOJ JPY pairs trading robot
Exit Attachments

scooby-doo BOJ JPY pairs trading robot

  • #6,781
  • Feb 25, 2010 2:05pm Feb 25, 2010 2:05pm
Quoting scooby-doo
There is.

Its called studying the underlying economics of a currency.

You could have 1000 technical indicators all saying its a LONG setup only for it to reverse for a number of reasons. i.e. News, Market consolidation, Hedge funds, Trending/Ranging market, Buyers and Sellers.

Example:

You enter a LONG trade on USDJPY and without warning HSBC trading floor decides to sell 100 Million USD cashing in on their previous LONG trades.

This ofcourse is not enough to cause a trend reversal but will blow most retail traders LONG trades and/or accounts....
I agree scoobs.. makes me realise that there probably is no such thing as a 100% hands free EA!

On that note, how is your v1.9 coming along today?
  • #6,782
  • Feb 25, 2010 2:27pm Feb 25, 2010 2:27pm
Quoting scooby-doo
......

Dont forget also that your broker may well be trading against your position in the hope of moving the market to your SL's.

Scoobs.
That's why i and others use a Straight Through Processing with No Dealing Desk Broker........

....but this doesn't help against Scooby's Trading Floor....
  • #6,783
  • Feb 25, 2010 2:32pm Feb 25, 2010 2:32pm
Quoting Viv108
I agree scoobs.. makes me realise that there probably is no such thing as a 100% hands free EA!

On that note, how is your v1.9 coming along today?
There can never be 100% fully automated hands free EA or system. Banks and other financial institutions would have built one by now given their resources. But they are very close to doing so.

However, v1.9 is going ok at the momento.

What I have done is to remove RSI(2/3) as the indicator and replaced it with LaguerreRSIMA function which is allot less sensitive to noise etc with 0.85 and 0.15 being the OB and OS levels.

Latest trades attached in picture.

Scoobs.
Attached Image (click to enlarge)
Click to Enlarge Name: gbpjpy.gif Size: 50 KB
  • #6,784
  • Feb 25, 2010 2:45pm Feb 25, 2010 2:45pm
that is awesome scoob!
  • #6,785
  • Feb 25, 2010 2:52pm Feb 25, 2010 2:52pm
Quoting Viv108
that is awesome scoob!
If all goes well tomorrow I will tidy up the code and release it here ready for Sunday market open.

Scoobs.
  • #6,786
  • Feb 25, 2010 2:57pm Feb 25, 2010 2:57pm
cool. Look forward to playing with it.

Its shame we cannot backtest these mutiple pairs robots... forward testing is sooooo boring...
  • #6,787
  • Edited 3:28pm Feb 25, 2010 2:59pm | Edited 3:28pm
Quoting scooby-doo
There can never be 100% fully automated hands free EA or system. Banks and other financial institutions would have built one by now given their resources. But they are very close to doing so.

However, v1.9 is going ok at the momento.

What I have done is to remove RSI(2/3) as the indicator and replaced it with LaguerreRSIMA function which is allot less sensitive to noise etc with 0.85 and 0.15 being the OB and OS levels.

Latest trades attached in picture.

Scoobs.
rsi(2/3) boost entries with Lrsi as well: on trending market, Lrsi can stay a long time OB/OS before reverting (see today GU h4 for example), and using the 3 rsi steps is nice. Maybe you can make it optional instead of removed.

Sorry Scoob, i realise now that i've made a confusion with the 3 rsi steps in TCF2, which are a nice entries conditions, and your bot where rsi(2/3) is something quite different.
  • #6,788
  • Feb 25, 2010 3:01pm Feb 25, 2010 3:01pm
Quoting Viv108
cool. Look forward to playing with it.

Its shame we cannot backtest these mutiple pairs robots... forward testing is sooooo boring...
you should understand that backtesting is nothing else than loosing time (yours of course )
  • #6,789
  • Edited 3:29pm Feb 25, 2010 3:28pm | Edited 3:29pm
Quoting Middleastern
I´m running v1.6. Here it opened just 2 trades today, one closed in profit, the other is 4 pips away from TP as of now.
Thank you Middleastern. I have no errors in the logs so I'll just keep them running.

Sandy
  • #6,790
  • Feb 25, 2010 3:30pm Feb 25, 2010 3:30pm
Quoting scooby-doo
If all goes well tomorrow I will tidy up the code and release it here ready for Sunday market open.

Scoobs.
Thanks Scooby
  • #6,791
  • Feb 25, 2010 4:51pm Feb 25, 2010 4:51pm
Hey guys, I went a whole half-week without coding another robot.

I was determined to go a full week, but bit9ret's stuff about average daily range trading got my juices going.

Sorry, I just could not help myself; the fingers just sort-of took over. Scooby-doo ADR OB_OS auto trading robot has made a nice start on demo here, and is available in post 1 - details at the bottom of the post.

I am not certain that the half-close and move to breakeven code is quite correct, so sing out if there are any anomalies.

  • #6,792
  • Feb 25, 2010 5:03pm Feb 25, 2010 5:03pm
Quoting SteveHopwood
Hey guys, I went a whole half-week without coding another robot.

I was determined to go a full week, but bit9ret's stuff about average daily range trading got my juices going.

Sorry, I just could not help myself; the fingers just sort-of took over. Scooby-doo ADR OB_OS auto trading robot has made a nice start on demo here, and is available in post 1 - details at the bottom of the post.

I am not certain that the half-close and move to breakeven code is quite correct, so sing out if there are any anomalies.

Resistance is futile - the FX coding bug has well and truly bitten.
  • #6,793
  • Feb 25, 2010 5:10pm Feb 25, 2010 5:10pm
Quoting getdown71
Resistance is futile - the FX coding bug has well and truly bitten.
gibber gibber.....code......shops need coding......car needs petrol need a robot to automate the trade with the garage.......cats need feeding must code a stop loss into their wasted food......................

  • #6,794
  • Feb 25, 2010 5:33pm Feb 25, 2010 5:33pm
Quoting nubchai
Scooby,

If there is major news for any pair then you hold off on trades? I noticed you search trhough all of the pairs to see if there's news. Whereas Zoog checks for news for a single pair. If I'm reading the code right.

Thanks,
Sandy
Had 4 trades today Sandy (today being GMT time).
So far THG 1.6 is up 5% since Sunday open. 14 winners and 4 losers. Alpari UK Demo.
  • #6,795
  • Feb 25, 2010 5:41pm Feb 25, 2010 5:41pm
Quoting Jezza
Had 4 trades today Sandy (today being GMT time).
So far THG 1.6 is up 5% since Sunday open. 14 winners and 4 losers. Alpari UK Demo.
Thanks Jezza. It must be something to do with my fxdd demo set-up. The zoogle ea is working. I'm going to take a closer look at my logs tonight.

Thanks and congats on the wins!

Sandy
  • #6,796
  • Feb 25, 2010 5:50pm Feb 25, 2010 5:50pm
Quoting nubchai
Thanks Jezza. It must be something to do with my fxdd demo set-up. The zoogle ea is working. I'm going to take a closer look at my logs tonight.

Thanks and congats on the wins!

Sandy
Thanks. Well we are at opposites, because I can't get Zoogle's EA to trade Only 1 trade all week. Ah well
  • #6,797
  • Feb 25, 2010 7:36pm Feb 25, 2010 7:36pm
Steve,

The addition of the Laguerre RSI indy to THG looks promising.

Perhaps it's worthwhile considering this additional filter for TCF v2.1

Your thoughts ?

getdown71
  • #6,798
  • Feb 25, 2010 7:58pm Feb 25, 2010 7:58pm
Quoting getdown71
Steve,

The addition of the Laguerre RSI indy to THG looks promising.

Perhaps it's worthwhile considering this additional filter for TCF v2.1

Your thoughts ?

getdown71
Long or short answer?

  • #6,799
  • Feb 25, 2010 8:19pm Feb 25, 2010 8:19pm
Quoting Jezza
Thanks. Well we are at opposites, because I can't get Zoogle's EA to trade Only 1 trade all week. Ah well
Lol Jezza I think I only had 1 Zoogle trade today. Now Holy Grail 1.4 and 1.6 trades have triggered. So it looks like I'm o.k.

Sandy
  • #6,800
  • Edited Feb 26, 2010 12:39am Feb 25, 2010 9:19pm | Edited Feb 26, 2010 12:39am
Scooby-Doo ADR OB_OS auto trading robot

Steve,

This looks an elegant mean reversion EA (though not simple to code, I recognize) with no trend variables to clutter it up.

I see that you've:

- split the ADR around the day's open, ie. entry point is day's open +/- 0.5 of ADR, with the default stop another +/- 0.2 ADR beyond that, so the effective range is day's open +/- 0.7 of ADR (ie. capturing two thirds of ADR on either side of day's open). That looks robust.

- and set the adjustable default TP/SL as 40/20% of ADR for 2:1 RR. Good stuff, and esp. good that it's adjustable.

So with a default TP set at 0.4 and entry point to be at +/- 0.5, you are aiming for a mean reversion back to within 0.1 ADR of the day's open.

Our risk is that prices run 100%+ of ADR mainly on one side of day's open (not +/-50% on either side). If stop is broken at 0.7 ADR (0.5 entry point + 0.2 ADR stop), we may need a SAR function that reverses the system at that point ON THAT PAIR into a breakout trade (not a mean reversion trade back to open) targetting a trend ADR multiple beyond that (say 0.2-0.4 ADR beyond the broken stop at 0.7 ADR, meaning we target the whole 1.0+ ADR trend on one side of the day's open). May be wishful thinking/impossible to code/impractical complication. Not sure.

Three questions if you have time:

1 - can you explain the the half close and move to breakeven code/feature?

2 - what is the pending trades pip distance of 150?

3 - and the pending trades expire 60 minutes?

One suggestion or idea: can we customize the ADR entry point % or set a multiplier % of it for the entry, ie. so it won't always enter +/- 50% of ADR from the day's open? So entry could be at 70-150%+ of ADR, etc., if trend is running one way very strongly from London/NY opens (ie. much more than 50% of ADR from open, when it's not a mean reversion day but a running/big trend day).

Mind you, if the day trends one way, the user can just turn the EA off, and/or mischievously restart it at London close etc. when the one way trend has just about exhausted itself. There's room for some user smarts here.

Hope lots of trades fire as EA restarts (ie. at belated entry points > usual +/- 0.5 ADR entry threshold). As prices often retrace after huge daily move into the London close, there should be a chance to capture say a 20% retrace/mean reversion back toward open (ie. we can adjust the TP down from 40% at some times of day). With 1:1 RR at worst (if SL is set to 20% too) this could be a big aggregate winner (re multiple pairs), LOL.

Devious multi-pronged counter-attack on Mr. Market, here.

Quoting SteveHopwood
Hey guys, I went a whole half-week without coding another robot.

I was determined to go a full week, but bit9ret's stuff about average daily range trading got my juices going.

Sorry, I just could not help myself; the fingers just sort-of took over. Scooby-doo ADR OB_OS auto trading robot has made a nice start on demo here, and is available in post 1 - details at the bottom of the post.

I am not certain that the half-close and move to breakeven code is quite correct, so sing out if there are any anomalies.

20 traders viewing now