Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

Nanningbob 4h trading system ver. 5.0 4,432 replies

Nanningbob 60x90 EA 3,920 replies

Require help amending Nanningbob v10 EA 0 replies

Help regarding NanningBob EA 27 replies

Nanningbob 4H trading system 2,017 replies

Attachments: BigBob: Nanningbob V10 auto-trading robot by SteveHopwood
Exit Attachments

BigBob: Nanningbob V10 auto-trading robot by SteveHopwood

Re the modelling quality guys, it is irrelevant.

There is a 3 stage process to getting single-pair trading ea's onto a live account:

 

  1. backtesting in strategy tester
  2. forward testing on demo
  3. live testing tiny lots on a live account


Backtesting in strategy tester has only two uses:

 

  1. initial debugging.
  2. checking that the robot is actually worthwhile taking any further. If it is not profitable in backtesting, it gets murdered on forward demo.

Anybody with more than a rudimentary knowledge of how trading EA's work - me, for instance - understands this and cares not a jot about modelling quality.

So, guys, remember the 3 stage development of an ea once the initial coding is done:

 

  1. if it fails backtesting it gets murdered in forward testing on demo.
  2. if it survives forward testing on demo, it may still get murdered on live account. Anybody who does not understand why needs to spend some time at BabyPips.com
  3. if it survives live testing using tiny lot sizes, most traders will still blow their accounts. They will become over-confident, increase their lot sizes to a point where the account cannot withstand an adverse move against their trades. Nowt I can do about them.

So, forget modelling quality; it is irrelevant.

Quoting winsteadglenn
Stellar performance today, OOTB. Bravo on the Hopwood/Myers/Osthafen fixes.
Yes, that recent flurry of activity helped a lot. It worked wonders for my demo manual practise trades.

OK gals and galesses, time to start insisting a a bit of discipline around here.

I am reading an increasing number of posts from contributors who clearly have not taken the trouble to work out what BB is all about, despite some of my comments at the start of that blasted great user guide in post 1. I find myself wondering if some of them even bothered downloading it, let alone reading it.

Contributors here with firing brain cells will be relieved to know that I shall revert to my customary practise of restricting the ability of readers to contribute, should people continue to post nonsense. It will only take a few more rubbish posts for me to do this.

Quoting SteveHopwood
Re the modelling quality guys, it is irrelevant.

There is a 3 stage process to getting single-pair trading ea's onto a live account:

 

  1. backtesting in strategy tester
  2. forward testing on demo
  3. live testing tiny lots on a live account


Backtesting in strategy tester has only two uses:

 

  1. initial debugging.
  2. checking that the robot is actually worthwhile taking any further. If it is not profitable in backtesting, it gets murdered on forward demo.

Anybody with more than a rudimentary knowledge of how trading...

sorry not correct! again:
on mt4 data you calc on daily high and lows, no ticks between
on 99% quality you calc on ticks too, so if your ea works with tick data, you must have 99% to debug tick errors in your ea!

example: if you use indi's with price_open, its useless in mt4, you cant see a ma cross between high and low! res=cross on next bar!too late

with 99% you can drive any price_open indi, your ea recognizes the cross correct, it calculates on each tick.

and if i remember correct! bigbob uses price_open?!

i dont want to insult any statement(human), but wrong is wrong, sorry guys

cheers

IX
Ok, there are clearly people around here determined to take no notice of anything I write, so I have restricted posting in this thread. Details in post 1.

Hehe. You would have thought I would have learned by now, wouldn't you?

Quoting SteveHopwood
Ok, there are clearly people around here determined to take no notice of anything I write, so I have restricted posting in this thread. Details in post 1.

Hehe. You would have thought I would have learned by now, wouldn't you?

Na, to say it with your words : mocking dimwits shouldn't be allowed to post. Good choice. They shall go to an well known east asia forum and leech some of the ripped 97$ EAs to be happy with.

My observation of the backtest data is that the least you have at 90% modelling quality is the high and low of the M1 candle, but no data in between. Every tick comes with 99%.

I have it running default on EU EJ GU GJ since y'day and was pleased to see that Bernies' sneeze today didn't harm it too much, so I think we might have one of the 1% of EAs that could make it over the distance

Cheers, George
"We are all just prisoners ... of our own device ..." (Hotel California)
Have put it on 10 daily charts, as "blind" (not analising market conditions, I mean), here's the output. Default settings, lots 0.02. Not a single error reported, pendings deleted where and when they should be. Will count exactly the win in $$, as this was a fresh 3k account, but some losses are from removing precedent versions of BB and closing positions as they were at that moment. Anyway, at least 120 $ profit up...
Nice one!
Attached Image (click to enlarge)
Click to Enlarge Name: bb 26.08.2011.jpg Size: 210 KB
I don't say
Also doing some more backtests which are looking BRILLIANT, even if they are kind of incomplete.

I said that, they're incomplete, if you gonna check the report order by order, you can notice that are several periods of entire months even, when BB has done nothing...for a celestial unknown reason, sometimes, in backtests (at least) it won't delete a set of pendings and close the original position (yeah, same glitch, seems it was just partially solved), so, until price reaches again there, BB is taking a well deserved holiday :-) Also, for a greater unknown reason, my Strategy tester is closing on 03 2011, even I got all history data up-to-date...
Anyway, I have traced all the 2008 activity, so, even with BB starting 31.01, with a big "holiday" between 18.03.2008 - 01.09.2008 (that's almost 6 months), it reaches 2008 year's end at 350% from a starting 1000$ capital. Also, if somebody will watch in detail, there are more little "holidays" of 5-10 days, because the bastard needs many little breaks too .
As you can see, 2009 and 2010 there was a "reverse working season" , meanig BB has been more in holiday mode than in working shifts...
But, come on, he deserves it! Just look at the equity curve

So, as a conclusion, we need to keep an eye on BB when on forward to prevent him going in vacantion again!

Steve, you have all my respect! 1000x

P.S. About the modelling quality, I got 90% usually, but that n/a comes from the sudden stop in 03 2011...just for the ones that need to know about it
Hint: BB should be much better in forward testing!
Attached Image (click to enlarge)
Click to Enlarge Name: StrategyTester_GJ_D1BigBob.gif Size: 8 KB
Attached File(s)
File Type: pdf StrategyTester_GJ_D1BigBob.pdf   969 KB | 572 downloads
I don't say
Quoting SteveHopwood
Re the modelling quality guys, it is irrelevant.

There is a 3 stage process to getting single-pair trading ea's onto a live account:

 

  1. backtesting in strategy tester
  2. forward testing on demo
  3. live testing tiny lots on a live account


Backtesting in strategy tester has only two uses:

 

  1. initial debugging.
  2. checking that the robot is actually worthwhile taking any further. If it is not profitable in backtesting, it gets murdered on forward demo.

Anybody with more than a rudimentary knowledge of how trading...



That is what I learned hard way during years trying to make some forexmoney .. Agree 100%
Well, that's what we doing atm, I also try a optimization for specific pairs regarding volatility values...but tommorow is another day, so good night all :-)
I don't say
Just dodged a bullet......Hurricane Irene missed me......Keep your head down Pippo, it's headed for you......

Have a great weekend

whip
Hope everyone has a great week-end.

As the market is closed, I'm bored, so I'm reading the BB code, trying to learn something..:-) I'm really intrigued by that strange behaviour that BB sometimes goes in pause mode...
And I found this, right after pendings closure:

if (TradeEveryTick) OldRangeTradingBars = 0;//Look for trades every tick, not every candle
//Range trading opps need checking even though there may a trade already open, else the open trade never gets closed
if (trend == ranging && AllowRangeTrading && OldRangeTradingBars != iBars(NULL, RangeTradingTF) && Volatility <= MaximumRangeTradingVolatility)
{
OldRangeTradingBars = iBars(NULL, RangeTradingTF);
LookForRangeTradingOpps();
}//if (trend == ranging && AllowRangeTrading && OldRangeTradingBars != iBars(NULL, RangeTradingTF))

Is the blue phrase expressed in code somewhere else, as the statement that follows (if (trend == ranging...etc), doesn't sound to me that is taking care of that
Well, I'm not a coder,so I can't "think code", just asking...
I don't say
The comment will belong to an earlier feature that changed or removed and is no longer relevant; I forgot to remove it.

Re the long pauses, they are going to happen when the market is against the style of trading that volatility allows BB to engage in - especially when vol says the pair is a trender yet the market is ranging.

There may be bugs in the code that cause this as well, but these will emerge in time. There are some seriously clever coders who check over my code for my customary range of bloops; if they haven't spotted any then they either do not exist or are very hard to spot.

Mind, any of these guys will confirm that my spaghetti-like code makes bloop-spotting very hard - one of the reasons they usually end up with a vouch from me. Anyone who can read my code deserves it.

Well, I'm not a coder, but,if I pay attention, it's not so hard to read...just that I don't understand all the variables, but it has a certain logic.
That's why I asked earlier, as BB does exactly what you commented there, sometimes :-)

Aha, catch it this time! As running a backtest, journal says: Order close error 4108 :-) Pending order # cannot be closed
I don't say
The start of a new trading week. Happy trading everyone.

This week should see BB going live.

Hello, Steve,
I wish you a great one too! You going live with BB already?

May I have a question for you?

Do you consider implementing a version without SL for the pending orders?
According to my backtests, is much more profitable than the actual version.
(yes, I admit, I've played with code, :-), but I couldn't make it to close both open positions, due to the "black hole" represented by my coding skills)

But, seriously, if we got a good ranging pair, BB is a miracle. I have (almost) manually tested January-February 2009 and the results are incredible. More than 150%, with 0.02 lots, DD around 5%.
I don't say
Yep. I'm looking forward to this week too. I'm quite skint, so have to have BB on demo. If it does well, I'll remortgage my house and stick some real money in the account (i'll probably get around £500 for the park bench I am typing this message from).
Quoting dragosd1
Hello, Steve,
I wish you a great one too! You going live with BB already?
Not yet. Soon.

Quote
May I have a question for you?

Do you consider implementing a version without SL for the pending orders?
According to my backtests, is much more profitable than the actual version.
(yes, I admit, I've played with code, :-), but I couldn't make it to close both open positions, due to the "black hole" represented by my coding skills)
I coded the filter I was offered. If you want to try this, go to void SendGrid(). Find the two lines of code that begin with "stop = ". Delete them both and re-compile.

Well, I already did that :-)
But, as usual, a problem occurs: BB won't close the 1st market order, because there's another one open :-) One of the pendings that was filled. Only if I would know how to change the CloseTrade bool to manage them both...I was struggled all weekend, believe me, but nothing worked . Now I've found a MQL tutorial, but until I will be able to understand your coding perfection, I'll be old enough and not able to read anymore!

P.S, Hey, that was a compliment!!

By the way, I just realised maybe it's not fair to post here about different mods that I experiment with BB, as this thread must be focused on the original EA. At least until I find something really amazing...Sorry!
I don't say
Hi Steve

I have a question regarding user guide page 3:

"Imagining a sell trade, here is how Cat works:
• drag the robot onto the chart, and set TradeLong to true and TradeShort to false."

Shouldn't this be the other way round, or have I misunderstood it entirely?

Kind Regards
0 traders viewing now