scooby-doo BOJ JPY pairs trading robot | Page 180 | 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

  • #3,581
  • Nov 28, 2009 11:50am Nov 28, 2009 11:50am
Quoting Fredella
Hi Steve,

just checked the code for 24H 2.6 and I might have spotted an issue:

//Cancel the basket if GJ is wrong side of 200 sma
if (UseSma && (Bid < LtfSMA + 0.2 || TrendDirection == down) )
{
GlobalVariableDel(OverboughtGV);
return;
}//if (Bid < LtfSMA)

//Lower low check
if (Low[1] < Low[2])
{
GlobalVariableDel(OversoldGV);
return;
}//if (Bid < High[1])

Should the OversoldGV not be OverboughtGV?
Nice piece of proofreading. Cheers.

  • #3,582
  • Nov 28, 2009 12:11pm Nov 28, 2009 12:11pm
Quoting SteveHopwood
Nice piece of proofreading. Cheers.

Welcome. I hope to give some more feedback on the running of the EA in the near future now that i'm aiming to run it in 24hr mode.

I'm not entirely sure what the reason was for incorporting this higher high and lower low filter. I must have missed a page in this thread. Is there not a risk we're filtering out too many valid signals now?
  • #3,583
  • Nov 28, 2009 12:24pm Nov 28, 2009 12:24pm
Quoting Fredella
Welcome. I hope to give some more feedback on the running of the EA in the near future now that i'm aiming to run it in 24hr mode.

I'm not entirely sure what the reason was for incorporting this higher high and lower low filter. I must have missed a page in this thread. Is there not a risk we're filtering out too many valid signals now?
The problem is that this method gives a great deal of false signals outside of the Tokyo session as we are basically trading OB and OS levels and waiting for the retraces. The higher high's and lower low's play a big part of it.

You can of course remove this extra filter but then you will end up trading false retraces and end up with large DD's if the retrace doesn't come or is too small for your TP settings. This happened allot on Thursday and Friday this past week.

However, i have given Steve details of a further filter which could replace the above one and be more accurate.

Scoob.
  • #3,584
  • Nov 28, 2009 12:46pm Nov 28, 2009 12:46pm
Quoting scooby-doo
The problem is that this method gives a great deal of false signals outside of the Tokyo session as we are basically trading OB and OS levels and waiting for the retraces. The higher high's and lower low's play a big part of it.

You can of course remove this extra filter but then you will end up trading false retraces and end up with large DD's if the retrace doesn't come or is too small for your TP settings. This happened allot on Thursday and Friday this past week.

However, i have given Steve details of a further filter which could replace...
Hey Scoobs
Suppose we are to trade it manually outside the tokyo session, what specific rules should we be looking at since you mentioned there are a lot of fake-outs in that period. Worked like a charm in Tokyo session on Thursday and Friday but did sent on 2 trades I took manually down the drain outside tokyo session.

Kaz
  • #3,585
  • Nov 28, 2009 12:48pm Nov 28, 2009 12:48pm
Scoobs, I have incorporated the tp into both bots and AO into the 24H version.

The Orignal bot has code designed to avoid repeated basket sends during the kind of chaos we experienced a couple of days ago. It also forces the bot to wait for a move out of oversold/bought before placing a wave 2 basket.

Next on my to do list is to adapt the code in the Original bot so it can trade all 3 waves with the same basic code - will make reading it easier and so reduce the chances of bugs creeping in. I shall do this early in the coming week.

The latest versions are in post 1, so proofreaders can have a look.

  • #3,586
  • Nov 28, 2009 12:58pm Nov 28, 2009 12:58pm
Quoting Fredella
I'm not entirely sure what the reason was for incorporting this higher high and lower low filter. I must have missed a page in this thread. Is there not a risk we're filtering out too many valid signals now?
The 24H version is an experiment will little chance of surviving through to live trading.

By comparison, the Original version is based on a proven trading strategy. It has already expanded to a 2-wave bot, and will expand further this week to incorporate the third wave. This means 6 trading waves a day, apart from Friday.

Given this, the 24H will probably recede into irrelevance. We shall see.

  • #3,587
  • Nov 28, 2009 1:00pm Nov 28, 2009 1:00pm
Quoting SteveHopwood
Scoobs, I have incorporated the tp into both bots and AO into the 24H version.

The Orignal bot has code designed to avoid repeated basket sends during the kind of chaos we experienced a couple of days ago. It also forces the bot to wait for a move out of oversold/bought before placing a wave 2 basket.

Next on my to do list is to adapt the code in the Original bot so it can trade all 3 waves with the same basic code - will make reading it easier and so reduce the chances of bugs creeping in. I shall do this early in the coming week.

The latest...
If you have incorporated the AO stuff then you can remove the higher high and lower low filter.

cheers
  • #3,588
  • Nov 28, 2009 1:08pm Nov 28, 2009 1:08pm
Quoting scooby-doo
If you have incorporated the AO stuff then you can remove the higher high and lower low filter.

cheers
Done and uploaded to post 1.

Cheers

  • #3,589
  • Nov 28, 2009 1:11pm Nov 28, 2009 1:11pm
Quoting SteveHopwood
The 24H version is an experiment will little chance of surviving through to live trading.
By comparison, the Original version is based on a proven trading strategy. It has already expanded to a 2-wave bot, and will expand further this week to incorporate the third wave. This means 6 trading waves a day, apart from Friday.

Given this, the 24H will probably recede into irrelevance. We shall see.

Got my answer to previous question for Scoobs. I'll just stick with the proven strategy for manual trades.
Cheers
Kaz
  • #3,590
  • Nov 28, 2009 1:13pm Nov 28, 2009 1:13pm
Quoting scooby-doo
If you have incorporated the AO stuff then you can remove the higher high and lower low filter.

cheers
Ok: here is my confession

i have so far played with the 24h in manual mode mostly: but with incredibly good results!

what i did make sure of was that the trip from overbought to oversold and vice versa was quick, and that after the close of a 15m candle we at least had one overbought or oversold level, i.e. not a retrace within the same candle. Maybe these 2 criteria could be added but i do not want to stop you from your views as you have obviously far more experience than i have.

I also mainly ran the manual entries in the London morning hours, partly because of time available but also because i feel that that is the most 'busy' time on the market and hence would lead to quick results.

I will continue with my observations and post again when i spot something of interest.

In the mean time, thank you again for coming up with this idea and steve for the implementation as it has given me and i'm sure many others some great opportunities and knowledge
  • #3,591
  • Nov 28, 2009 1:17pm Nov 28, 2009 1:17pm
Quoting SteveHopwood
Done and uploaded to post 1.

Cheers

Sorry to be a pest but because of the inclusion of the AO we can change the OB and OS thresholds to 90/10 (LTF) and 70/30 (HTF).

We will get allot of signals but the AO averages should weed the bad ones out, i think!

Cheers

Scoobs
  • #3,592
  • Nov 28, 2009 1:23pm Nov 28, 2009 1:23pm
Quoting scooby-doo
Sorry to be a pest but because of the inclusion of the AO we can change the OB and OS thresholds to 90/10 (LTF) and 70/30 (HTF).

We will get allot of signals but the AO averages should weed the bad ones out, i think!

Cheers

Scoobs
Be a pest as much as you want. The rest of us will worry if you stop.

Updated versions in post 1. Nobody else appears to have downloaded yet, so the repeated uploads shouldn't be a problem for anyone.

  • #3,593
  • Nov 28, 2009 2:21pm Nov 28, 2009 2:21pm
Quoting scooby-doo
Sorry to be a pest but because of the inclusion of the AO we can change the OB and OS thresholds to 90/10 (LTF) and 70/30 (HTF).

We will get allot of signals but the AO averages should weed the bad ones out, i think!

Cheers

Scoobs

I am not a programmer, so please: What does AO stand for?

Also I understand that new versions of say the original robot are mainly for programmers to vet first, before us mere mortals are encouraged to download to demo or even later to live account.

Can you therefore please tell us which version of the original is now suitable for downloading to demo (presumably now incorporating wave 2) and which version could be used live?

I personally prefer to have the original robot on demo and than mimic the trades live, using myself as the ultimate, mostly unreliable, filter. I am still in awe on how this robot works, maybe because I grew up with wind-up alarm clocks!
  • #3,594
  • Nov 28, 2009 2:29pm Nov 28, 2009 2:29pm
Quoting ikdenk
I am not a programmer, so please: What does AO stand for?

Also I understand that new versions of say the original robot are mainly for programmers to vet first, before us mere mortals are encouraged to download to demo or even later to live account.

Can you therefore please tell us which version of the original is now suitable for downloading to demo (presumably now incorporating wave 2) and which version could be used live?

I personally prefer to have the original robot on demo and than mimic the trades live, using myself as the ultimate,...
http://ta.mql4.com/indicators/bills/awesome
  • #3,595
  • Nov 28, 2009 2:56pm Nov 28, 2009 2:56pm
Quoting SteveHopwood
Be a pest as much as you want. The rest of us will worry if you stop.

Updated versions in post 1. Nobody else appears to have downloaded yet, so the repeated uploads shouldn't be a problem for anyone.

There's still 27hrs before the trading session opens-- there will be at least two more versions by then...LOL...

As for hedging- I use hedging as part of my strategy- but I don't hedge every trade and sometimes the hedge can become the larger liability. Sometimes it's just better to get stopped out and then re-enter with a better entry-- probably sounds weird. And sometimes it's more important to retain equity (thus hedging).

On the EU I have an older EA that I use and hedge every trade at 30 pips- with a tp of 50 pips- (basically price is against you entry by 80 pips at this point (on very low % trades this works well- as you can have multiple hedges and orders and still be under 2% acct dd. This works on EU for several reasons -1. consolidation- 2. Too many traders are holding shorts (The main players won't trend price too much down unless the sentiment turns bullish) 3. Daily LTL hasn't been breeched since last March so keeping trades within this area (~400 pip range) is easily manageable. The EA will heged itself by default (having 3-5 orders open) as long as DD for the EA is less then the profit is banked in a 48 hr period it works.. I also will hedge trades at -300 pips if the daily trend has made a change.

Waiting til 5% (for small accts isn't that much) this could put traders in a position of 20-30% dd if they TP the hedge and price continues against them. (BTDT) or worse lose their accts. Anyone with experience knows the GY can kill acct quick- but also give great rewards. hedging works- but it can be very mentally taxing as well- or you might need to leave trades open for longer periods of time to find those larger "bounces" to rid both for BE or small profit. When I hedge against the main trend I start at 25-50%. I have 4x hedged a trade before- as I realized I was "going the wrong way" and profited.
As you have pointed out a hedge trade is a "dead trade" I have left hedge trades open for literally months on GU before and closed both for profits- but it can be very nagging to have open trades (not to mention the swap %) While I continued to trade- I had to use a SL because I didn't want multiple hedges opened- nor was I just going to sit around and not trade...LOL...

For me I would prefer a set pip hedge (as I like to set my lot sizes too) say 30-60 pips on xxxJY and manage both trades once a hedge is opened- and also either a set SL/TP (maybe a TP of 50 pips with a TS of 7 pips) and of course the option to not hedge at all. A good hedging EA would need to determine the the major S/R lines off a 4hr or daily chart to catch theses major bounces and close both the hedge and original trade in profit (and sometimes this might take a few days) Again the GY would be a perfect example- it bounced at 139 ish --anyone holding a long and short GY could have removed a hedge of even a couple hundred pips had they known this was a major support level. Same with GU (OT) 1.6250ish has always been a major support as well(good to know if looking to remove a hedge)
  • #3,596
  • Nov 28, 2009 3:07pm Nov 28, 2009 3:07pm
my alteration to scooby doo
Attached Image (click to enlarge)
Click to Enlarge Name: New Bitmap Image.jpg Size: 216 KB
  • #3,597
  • Nov 28, 2009 5:36pm Nov 28, 2009 5:36pm
Everyone

Because of the new way we calculate TP you should not set the TpMultiplicationFactor higher than 200.
  • #3,598
  • Nov 29, 2009 8:09am Nov 29, 2009 8:09am
Sorry if I missed something. but I noticed the correlation code is present in the 24H robot but not the "original." Group correlation is not necessary in the original strategy?

Thanks, Jack
  • #3,599
  • Nov 29, 2009 10:19am Nov 29, 2009 10:19am
Quoting SteveHopwood
The 24H version is an experiment will little chance of surviving through to live trading.

By comparison, the Original version is based on a proven trading strategy. It has already expanded to a 2-wave bot, and will expand further this week to incorporate the third wave. This means 6 trading waves a day, apart from Friday.

Given this, the 24H will probably recede into irrelevance. We shall see.

1) In the Original V1.1
void LookForOversoldOpportunity()
:
:
GlobalVariableSet(PreCloseTradingDoneGV,0);
GlobalVariableDel(PreCloseTradingDoneGV);

I think you meant GlobalVariableDel(PreCloseOversoldGV);

2) This is something I am not too sure of. But looking at your code in the Original V1.1, the 2-wave might not be sent if the 1-wave is done, and GlobalVariable TradingDoneGV is set.
eg:
if (!GlobalVariableCheck(TradingDoneGV) && !GlobalVariableCheck(PreOpenTradingDoneGV) ) LookForOverboughtOpportunity();

But again, I'm not too sure about this one. Please confirm.

Good job, Steve!
  • #3,600
  • Nov 29, 2009 1:01pm Nov 29, 2009 1:01pm
Quoting Kling
1) In the Original V1.1
void LookForOversoldOpportunity()
:
:
GlobalVariableSet(PreCloseTradingDoneGV,0);
GlobalVariableDel(PreCloseTradingDoneGV);

I think you meant GlobalVariableDel(PreCloseOversoldGV);
Super piece of proofreading again, Kling; thanks. I have saved the fixed version as 1.2 so everyone knows they have the correct version.

Quote
2) This is something I am not too sure of. But looking at your code in the Original V1.1, the 2-wave might not be sent if the 1-wave is done, and GlobalVariable TradingDoneGV is set.
eg:
if (!GlobalVariableCheck(TradingDoneGV) && !GlobalVariableCheck(PreOpenTradingDoneGV) ) LookForOverboughtOpportunity();

But again, I'm not too sure about this one. Please confirm.

Good job, Steve!
If wave 2 has already happened before the robot wakes up, then there will be no wave 2 trades; we certainly don't want wave 2 trades once BOJ has intervened and the robot's trading is over.

2 traders viewing now