scooby-doo BOJ JPY pairs trading robot | Page 228 | 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

  • #4,541
  • Jan 16, 2010 4:26am Jan 16, 2010 4:26am
Quoting xicecube
Steve, I just realized that the earlier versions of the AI had the "CriminalsECN" option while the latest version doesn't. Did that slip through or does the latest AI take care of this by itself?
Best polish your glasses, xicecube.
  • #4,542
  • Jan 16, 2010 4:32am Jan 16, 2010 4:32am
Quoting scooby-doo
A BUY would be correct but not just because all the RSI(2) are green.

1. You should look for it being OS in 3 TFs x 4 apart, so, 15M + 1H + 4H.
2. It is above the 200 SMA.
3. It is below the medium price channel line.

Well done, however, being GBPJPY you would need to have larger SL than other currencies and be careful of a trend reversal.

Scoobs.
Hi Scoobs,

Could you please explain how you draw your channel, by that I mean how you derive your initial start-point and then how you update it as PA proceeds?

Also, with respect to the BUY call in post #4522, was this based on the RSI(2) being OS in several timeframes or were there other considerations.

With thanks.
  • #4,543
  • Jan 16, 2010 6:15am Jan 16, 2010 6:15am
I have added an updated version of the Simple Entry robot to the zip in post 1. I haven't bothered changing the version number; there are not going to be multiple releases of this bot.

I have added the HedgeNotStopLoss input. When set to true, the robot will not send a stop with the trade. Instead, it will monitor the trade and send a 50% hedge at the stop loss point, removing the tp from the first trade. It will monitor the position and close both trades when their combined upl eventually returns to the profit percentage chosen in BreakEvenProfitPercent.

Backtesting showed that a sustained movement against the position would kill the account with a LotDivisor of 10000, so I have doubled this by default.

Backtesting invariably finishes with a hedged position in place. That apart, partial hedging works wonders for this robot. I shall add the same feature to the Cross Pairs robot.

Have fun, backtesters.

  • #4,544
  • Jan 16, 2010 8:18am Jan 16, 2010 8:18am
Quoting SteveHopwood
I have added an updated version of the Simple Entry robot to the zip in post 1. I haven't bothered changing the version number; there are not going to be multiple releases of this bot.

I have added the HedgeNotStopLoss input. When set to true, the robot will not send a stop with the trade. Instead, it will monitor the trade and send a 50% hedge at the stop loss point, removing the tp from the first trade. It will monitor the position and close both trades when their combined upl eventually returns to the profit percentage chosen in BreakEvenProfitPercent.

Backtesting...
It would be better to set the default lotdivisor to 25000 for extreme safety.

But hey what do I know.

Scoobs.
  • #4,545
  • Jan 16, 2010 8:30am Jan 16, 2010 8:30am
Steve, delete this if deemed inapproiate.
Inserted Video
Vincit qui se vincit.
  • #4,546
  • Jan 16, 2010 8:32am Jan 16, 2010 8:32am
Quoting scooby-doo
It would be better to set the default lotdivisor to 25000 for extreme safety.

But hey what do I know.

Scoobs.
Just an ignorant, no-knowledge, space-wasting ho-hoper, aren't you?

Consider it done.

  • #4,547
  • Jan 16, 2010 8:39am Jan 16, 2010 8:39am
Quoting SteveHopwood
Just an ignorant, no-knowledge, space-wasting ho-hoper, aren't you?

Consider it done.

  • #4,548
  • Jan 16, 2010 10:35am Jan 16, 2010 10:35am
Quoting SteveHopwood
I have added an updated version of the Simple Entry robot to the zip in post 1. I haven't bothered changing the version number; there are not going to be multiple releases of this bot.

I have added the HedgeNotStopLoss input. When set to true, the robot will not send a stop with the trade. Instead, it will monitor the trade and send a 50% hedge at the stop loss point, removing the tp from the first trade. It will monitor the position and close both trades when their combined upl eventually returns to the profit percentage chosen in BreakEvenProfitPercent.

Backtesting...
Steve,
Just a suggestion. Why not keep both optional? Many traders are with US Criminals (I am not one of them) so the 'hedging robot' may not work for them. Moreover, to be honest, I do prefer the 'stoploss' style of trading because if a trade goes South I may as well dump it and move on to new opportunities rather than nursing the bad trade.
As I said early on this is just a suggestion. Take it or leave it as you like.

Vinesh
  • #4,549
  • Jan 16, 2010 10:40am Jan 16, 2010 10:40am
Quoting SteveHopwood
Best polish your glasses, xicecube.
Hey, how could I possibly know that you were hiding the latest version in post 1?

Ok, sorry, my bad. Thanks!
  • #4,550
  • Jan 16, 2010 11:21am Jan 16, 2010 11:21am
Quoting scooby-doo
I work for a large trading floor.

"In retail Forex, technical analysis is a vital trading tool, as markets are "blind." What's a blind market? In a market where you cannot see the 'tape' and/or, order size, the market is considered 'blind.' In essence, you cannot see the order flow, or those actually making market, to see where the buying and selling is coming from."
Kudos on an excellent thread. I have a question for Scooby. I know that we will never capture Scooby's knowledge, experience and intuition in an EA, but I wonder whether we might be able to benefit from his use of information that retail traders don't see. Specifically, we might be able to write an indicator that approximated the order flow info.

If we use an ECN we have a small window into a large interbank network. Some ECN's provide level II quotes on their native platform which might be piped into MT4 using a DLL. Also we might be able to infer the order flow info using an indicator that kept a running sum of trades. (Up-ticks are buys and down ticks are sells. The sum of (delta-price)*(trade size) is similar to order flow?). Dealing desks mask this info, but ECN's cannot.

Suppose that we could do this. Would you be willing to suggest how this might be incorporated into the EA's, or is this info too subjective to be of use in an EA?

Thanks for this thread.
MadCow
  • #4,551
  • Jan 16, 2010 11:41am Jan 16, 2010 11:41am
Quoting MadCow
Kudos on an excellent thread. I have a question for Scooby. I know that we will never capture Scooby's knowledge, experience and intuition in an EA, but I wonder whether we might be able to benefit from his use of information that retail traders don't see. Specifically, we might be able to write an indicator that approximated the order flow info.

If we use an ECN we have a small window into a large interbank network. Some ECN's provide level II quotes on their native platform which might be piped into MT4 using a DLL. Also we might be able to infer...
You cannot trade just using order flow as pending orders can appear and disapper in milliseconds.

Central Banks will often place large pending orders at certain levels and it then becomes a fight between supply/demand if those levels are broken or held.

Order flow is just another type of global indicator like interest rates. It just becomes another input to your trading direction which YOU have to analyse and decide upon.

Knowing order flow just means you are no longer blind in the market. It does not however you can gaurantee a profit.

As I have stated previosuly, my trading floor has market pattern recognition software which incorporates order flow and a multitude of other inputs/indicators, it also scans incoming economic news and presents me with CLOSE, BUY or SELL advice. I then decide if it is correct or not. It is usually correct 68% of the time on current stats. That may sound low to some but dont forget we are not scalping 5 or 10 points per trade to get near 100% accuracy like some ebay solutions claim.

To answer your point, then yes you could probably build something but dont forget banks and other financial instituitions spend millions developing such systems and they are still very far away from being totally automatic or correct without some form of human interaction.

Scoobs.
  • #4,552
  • Jan 16, 2010 11:48am Jan 16, 2010 11:48am
Quoting vinesh
Steve,
Just a suggestion. Why not keep both optional?

Vinesh
Hehe. Vinesh, all I can say is that the clue is in this bit, from post 4545, where I alerted you all to the update:

Quote
I have added the HedgeNotStopLoss input. When set to true, the robot will not send a stop with the trade. Instead, it will monitor the trade and send a 50% hedge at the stop loss point,,,,,,,,,,,,,,,,,,,,,,,,,,
So, what do you think happens if HedgeNotStopLoss is set to false?

No prizes for a correct answer, I am afraid, but a shed-load of gentle mockery and scorn if you get the wrong one.

  • #4,553
  • Jan 16, 2010 12:02pm Jan 16, 2010 12:02pm
There some updated versions in the robot zip in post 1:

 

  1. Simple Entry: altered the default LotDivisor to 25000. If you know what scoobs is talking about, then you can make this change for yourselves and don't need the update.
  2. Crosspairs: added the hedging functionality copied from Simple Entry.
  3. Opposite bot: changed the hedging so it takes an immediate 50% hedge of both trades instead of doing an incremental hedge of the biggest loser. This is a radical change, so there is a new version number so people can hang on to the old version if they prefer. Users might consider moving the HedgeAtLossPercent, perhaps to 7.5% or even 10%. We shall see what happens on demo.

Next on my todo list is to add the hedging to the AI robot.

  • #4,554
  • Jan 16, 2010 12:04pm Jan 16, 2010 12:04pm
Quoting SteveHopwood
There some updated versions in the robot zip in post 1:
[list][*]Simple Entry: altered the default LotDivisor to 25000. If you know what scoobs is talking about, then you can make this change for yourselves and don't need the update.[*]Crosspairs: added the hedging functionality copied from Simple Entry.[*]Opposite bot: changed the hedging so it takes an immediate 50% hedge of both trades instead of doing an incremental hedge of the biggest loser. This is a radical change, so there is a new version number so people can hang on to the old version...
  • #4,555
  • Jan 16, 2010 12:10pm Jan 16, 2010 12:10pm
Quoting Pippopotamus
Steve, delete this if deemed inapproiate.http://www.youtube.com/watch?v=QXHUIpNCu2k
I don't have the ability to delete individual posts. Even if I did, I would hardly delete this. Thanks for posting it.

Growing up, I would have sacrificed a lot to have access to this kind of clip.

It is fascinating to listen to such a great pianist in his younger days. Interestingly, I would have to stop a young pianist playing like this if the performance were to be assessed for, say, a diploma exam. Benedetti's use of sustaining pedal and rubato is considered 'unstylistic' by most modern examiners and the player would run the risk of his or her interpretation being rejected.

Yet, the playing is wonderful. Strange world.

  • #4,556
  • Jan 16, 2010 12:27pm Jan 16, 2010 12:27pm
Just to let you guys know, I had a quick look at the AI code and decided immediately against adding hedging.

Breakout systems either work or they don't. I cannot see hedging being any help when the breakout turns to fakeout. I shall be doing no further development of this robot.

  • #4,557
  • Jan 16, 2010 12:28pm Jan 16, 2010 12:28pm
If you have a young pianist playing like this, please don't stop him.
Vincit qui se vincit.
  • #4,558
  • Jan 16, 2010 12:34pm Jan 16, 2010 12:34pm
Quoting Pippopotamus
If you have a young pianist playing like this, please don't stop him.
No, he probably wouldn't be taking a diploma so I wouldn't need to.

  • #4,559
  • Jan 16, 2010 2:20pm Jan 16, 2010 2:20pm
Steve

I dont know if you have an EA that does the following but just some thoughts.

Scan all currency pairs available from your broker:

Go Long

- if currency > 200 4H SMA and OS in 15M, 1H and 4H.

Go Short

- if currency < 200 4H SMA and OB in 15M, 1H and 4H.

Obviously you would need to set TP, SL and Spread size. I would use a lotdivisor of 100000 and allow a maximum of 7 trades to be opened at any one time.

Also, I would use an overall TP (close all) of 1.25% of account size. Personally, I think you should just scan EUR, JPY and USD pairs.

This is similar to the send groups if OB or OS but we only send those pairs that meet the conditions above.

Obviously you would need to decide if OB or OS is 90/10, 80/20, 70/30.

I am not asking you to build yet another EA but this is just some information.

Scoobs
  • #4,560
  • Jan 16, 2010 4:42pm Jan 16, 2010 4:42pm
Makesure you read this article at least 10 times.

http://www.investopedia.com/articles...-timeframe.asp

Scoobs
1 trader viewing now