scooby-doo BOJ JPY pairs trading robot | Page 363 | 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

  • #7,241
  • Edited 11:10am Mar 7, 2010 10:59am | Edited 11:10am
Quote
1. ALL DAY events are not handled.
2. TENTATIVE events are not handled.
3. News impact is not allways correct.
4. The file is not always updated before Sunday open.
5. There is no support for different timezones when called by an EA.
6. The function is over complicated and difficult to modify.
7. If high impact news is enough to turn the trade against you by more than your ADR SL then you should just take the loss or kill the trade. Most likely a trend reversal anyway.
I'm going to write LibEconomicNews.mqh
which ought to fix the programming problems like #4, #5 and #6, assuming they are programming problems. If they are data feed problems then code can't fix them. So far it appears that the time in the FFCal machine readable is GMT.

If the FF calendar has formats for #1 and #2 then these could be incorporated, but I'll need examples.
  • #7,242
  • Mar 7, 2010 11:37am Mar 7, 2010 11:37am
Quoting mbkennel
I'm going to write LibEconomicNews.mqh
which ought to fix the programming problems like #4, #5 and #6, assuming they are programming problems. If they are data feed problems then code can't fix them. So far it appears that the time in the FFCal machine readable is GMT.

If the FF calendar has formats for #1 and #2 then these could be incorporated, but I'll need examples.
3 and 4 are not programming problems but down to FF news provider.

I have contacted FF about this but they are not interested in sorting it out.

Scoobs.
  • #7,243
  • Mar 7, 2010 11:43am Mar 7, 2010 11:43am
Attached is my uploaded results from running

The Holy Grail v1.6

for the past two weeks. I was a little conservative in my previous report. Quite honestly, it makes my head hurt to count that many pips. Let's just say it made a lot - 700+ over the past two weeks.

Remember, the forex gods will get you if do not show the proper respect by using only the initials instead of typing the entire name...

The Holy Grail v1.6

Scoobs as creator obviously excepted. Steve, I don't know, you decide if you want to take the chance...

T...
Attached File(s)
File Type: xls The Holy Grail through 030610.xls   34 KB | 423 downloads
  • #7,244
  • Mar 7, 2010 12:02pm Mar 7, 2010 12:02pm
Quoting scooby-doo
3 and 4 are not programming problems but down to FF news provider.

I have contacted FF about this but they are not interested in sorting it out.
I'm not sure about #3. Are there specific examples of incorrect impact for news events? If we can find a pattern of incorrect information, we can probably implement a fix for this issue.

There are solutions for #4. For example, when the XML file is not available on Sunday (or any other time), you simply would read the calendar page of the FF web site and parse the news data accordingly. This also works great for collecting historical news data for backtesting purposes.

It looks like mbkennel is already working on a solution to these issues so I guess we can discuss after his solution is provided.
  • #7,245
  • Mar 7, 2010 12:45pm Mar 7, 2010 12:45pm
Quoting zznbrm
I'm not sure about #3. Are there specific examples of incorrect impact for news events? If we can find a pattern of incorrect information, we can probably implement a fix for this issue.

There are solutions for #4. For example, when the XML file is not available on Sunday (or any other time), you simply would read the calendar page of the FF web site and parse the news data accordingly. This also works great for collecting historical news data for backtesting purposes.

It looks like mbkennel is already working on a solution to these issues so...
Sometimes the news provider says it is HIGH impact when it is MEDIUM impact and vice versa or things are just missing.

Anyway, I am not applying news rules anymore as there are very few which will actually create a trend reversal.

People need to stop looking for 100% win rate and concentrate their minds more on understanding that you will lose trades whatever you do.

Scoobs.
  • #7,246
  • Edited 1:05pm Mar 7, 2010 12:51pm | Edited 1:05pm
Quoting scooby-doo
Sometimes the news provider says it is HIGH impact when it is MEDIUM impact and vice versa or things are just missing.

Anyway, I am not applying news rules anymore as there are very few which will actually create a trend reversal.

People need to stop looking for 100% win rate and concentrate their minds more on understanding that you will lose trades whatever you do.

Scoobs.
Couldn't agree more. 100% win rate is a worthless subject.
  • #7,247
  • Mar 7, 2010 2:17pm Mar 7, 2010 2:17pm
Quoting scooby-doo
Anyway, I am not applying news rules anymore as there are very few which will actually create a trend reversal.
If a news filter is not condusive to higher profitability of a trading strategy, then there certainly is no reason to use it.

Quote
People need to stop looking for 100% win rate and concentrate their minds more on understanding that you will lose trades whatever you do.

So many of the strategies presented here have had such bad Risk:Reward ratios (2:1, 3:1, and higher), they had to strive for very high win rates. How many equity graphs have we seen slowly step upward to nice profits, only to see it all go away in 1 trade? That's no way to trade...those strategies will put a hole in your stomach the size of a jalapeno!!!

The more favorable your R:R ratio is, the easier it is to understand that a very high win rate is unnecessary.
  • #7,248
  • Mar 7, 2010 3:58pm Mar 7, 2010 3:58pm
Quoting scooby-doo
Sometimes the news provider says it is HIGH impact when it is MEDIUM impact and vice versa or things are just missing.

Anyway, I am not applying news rules anymore as there are very few which will actually create a trend reversal.

People need to stop looking for 100% win rate and concentrate their minds more on understanding that you will lose trades whatever you do.

Scoobs.
For what it is worth, here is what I think:

 

  1. news events come and go. Their effect on my trading is minimal to non-existent.
  2. s'r either holds or it is broken. When the movement is in our favour great; when not, we simply take the hit and move on.
  3. it seems to me news events that were earth-shattering when I first started trading have little impact these days. I cannot remember the last time even the dreaded NFP created a decent spike.

My latest robot here does not include a call to FFCal; there does not seem to be any point.

  • #7,249
  • Mar 7, 2010 4:29pm Mar 7, 2010 4:29pm
Quoting scooby-doo
Sometimes the news provider says it is HIGH impact when it is MEDIUM impact and vice versa or things are just missing.

Anyway, I am not applying news rules anymore as there are very few which will actually create a trend reversal.

People need to stop looking for 100% win rate and concentrate their minds more on understanding that you will lose trades whatever you do.

Scoobs.
actually, if TP vs SL is not dramatic, say, 1:5, or 1:10 like those scalping maniac, then news impact may well be a coin-flip, let it be, you never know this future news will be good or bad for this entered position.

I am not discouraging Mbkennel's creative efforts, but just want to make it known ultimately, it is ok with a reasonable TP vs SL ratio without considering news.

no 100% is needed or wanted, that's why i am more eager to see how much HG lost when reading a statement.
  • #7,250
  • Mar 7, 2010 4:46pm Mar 7, 2010 4:46pm
Incredible, you have no loss! How come? but looks like this is a live account, congradulations!

Quoting tsquad
Attached is my uploaded results from running

for the past two weeks. I was a little conservative in my previous report. Quite honestly, it makes my head hurt to count that many pips. Let's just say it made a lot - 700+ over the past two weeks.

Remember, the forex gods will get you if do not show the proper respect by using only the initials instead of typing the entire name...
  • #7,251
  • Mar 7, 2010 4:46pm Mar 7, 2010 4:46pm
Quoting charvo
much HG lost when reading a statement.
And why I was so delighted recently when 1.6 (I think) started with a couple of big losers. It was great to see the robot making up the losses and ending in a stomping great profit.

  • #7,252
  • Mar 7, 2010 4:56pm Mar 7, 2010 4:56pm
Quoting charvo
Incredible, you have no loss! How come? but looks like this is a live account, congradulations!
Please look again - this is a demo account and there were 3 losing trades in the group....

T...
  • #7,253
  • Mar 7, 2010 5:17pm Mar 7, 2010 5:17pm
Hehe. Having gone the whole second half of last week without taking any trades at all, the latest ADR_OB_OS robot has kicked off the new week with 6 'stop_and_reverse' trades.

I think I need a Sunday candle filter.
  • #7,254
  • Edited 5:38pm Mar 7, 2010 5:26pm | Edited 5:38pm
Quoting zznbrm
If a news filter is not condusive to higher profitability of a trading strategy, then there certainly is no reason to use it.
I was hoping that a news filter improves Sharpe ratio.

Quote
actually, if TP vs SL is not dramatic, say, 1:5, or 1:10 like those scalping maniac, then news impact may well be a coin-flip
That's my concern, increase in volatility of returns but zero expectation => lower Sharpe.

There's obviously a compromise, run half size with news filter.

In serious news against your position, you won't get filled in live trading anywhere near your intended stop loss. (demo will look much better than real).
  • #7,255
  • Mar 7, 2010 5:29pm Mar 7, 2010 5:29pm
Quoting SteveHopwood
it seems to me news events that were earth-shattering when I first started trading have little impact these days. I cannot remember the last time even the dreaded NFP created a decent spike.
In my non-expert opinion, this will change with the economic cycle, when more CB's start getting itchy on raising rates.

Right now, except for China, economies are so bad that there will be little movement for a long time.

When somebody breaks first to set up major carry trade imbalances then news will start to matter more as the "partial derivative of interest rates with respect to news surprises" will increase substantially.
  • #7,256
  • Mar 7, 2010 5:31pm Mar 7, 2010 5:31pm
Quote
People need to stop looking for 100% win rate and concentrate their minds more on understanding that you will lose trades whatever you do.

I agree in principle. However, what I like about the latest Holy Grail is that this was achieved in a week which was not ideal for automated trading (you even stated yourself that indicators would not help us on Thursday as the hedge funds would be making their move), and also that the SL to TP ratio is 1:1.

Since 1:1 basically means we can lose 50% of the trades and still break even, I figure that the bot will have to really fall along way from 100% accuracy last week to start being a loser. 100% in a below average week makes me think it should be able to make 60% in a really bad week, and almost never drop below the 50% it needs to keep the balance overall positive.

So personally I found the win rate to be encouraging for the above reasons.
  • #7,257
  • Mar 7, 2010 5:35pm Mar 7, 2010 5:35pm
Quoting mbkennel
In my opinion, this will change with the economic cycle, when more CB's start getting itchy on raising rates.

Right now, except for China, economies are so bad that there will be little movement for a long time. When somebody breaks first to set up major carry trade imbalances then news will start to matter more.
Sounds like a plan for worrying about this in the future rather than now.

  • #7,258
  • Mar 7, 2010 5:44pm Mar 7, 2010 5:44pm
Quoting SteveHopwood
Sounds like a plan for worrying about this in the future rather than now.

With regard to the UK in about 125 years time.

  • #7,259
  • Mar 7, 2010 5:51pm Mar 7, 2010 5:51pm
Quoting scooby-doo
With regard to the UK in about 125 years time.
Well, before that time it will trade like an "emerging market" ("demerging market"), where news on sovereign risk & capital controls matters as much as rates. GBP/ARS, turds of a feather.

The majors will be Chinese Yuan, Deutschemark and Arab Light.
  • #7,260
  • Mar 7, 2010 6:37pm Mar 7, 2010 6:37pm
Incredible results with maxou v516 despite a bug on 02/17!
Attached File(s)
File Type: pdf maxou516.pdf   72 KB | 350 downloads
4 traders viewing now