Nanningbob 60x90 EA | Page 52 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

NanningBob's 4H Sellhigh BUYLOW 158 replies

Help regarding NanningBob EA 27 replies

Nanningbob 4H trading system 2,017 replies

Indicator needed for Nanningbob strategy 2 replies

Help to improve profitable EA - 1H NanningBob 40MA 200 BB 10 replies

Attachments: Nanningbob 60x90 EA
Exit Attachments

Nanningbob 60x90 EA

  • #1,021
  • Mar 8, 2009 11:02pm Mar 8, 2009 11:02pm
Still haven't noticed fxdd changing spreads at all, so I trade sundays.
  • #1,022
  • Mar 8, 2009 11:06pm Mar 8, 2009 11:06pm
Quoting tazmaan
Still haven't noticed fxdd changing spreads at all, so I trade sundays.
yea I run my settings when the markets are open too. So I guess it depends on what JT is doing as well.
  • #1,023
  • Mar 8, 2009 11:12pm Mar 8, 2009 11:12pm
Hi,

Can someone do me a favor? I can't seem to find the recovery EA you guys were talking about. Would you lend me a hand and give me the best recovery EA you can find. Thanks.
  • #1,024
  • Mar 9, 2009 12:14am Mar 9, 2009 12:14am
In uncalibrated state


SET ALL CHARTS TO 1H in metatrader4 not in EA
I recommend setting all of the charts to 1H (wish I could set them at 2H or 3H but I don’t have those buttons)
To get this look you need

12 charts and tile vertically. Zoom out as far as you can (get as many bars in view as possible. Do not lock or fix scaling. Set all charts to 1H
Depending on your screen size (I’m on a 24’’ )
EA am using

AllTF_nanningBob 60x90_hiddenTP
If you notice on the charts you can see the MAs all look different, some look zoomed out and some look zoomed in.
We are basing our entry on 90 from center MA.
we know that the smaller the TF the farther the entry to first step is.
Am I right so far daddy?
So on to rules.
Attached Image (click to enlarge)
Click to Enlarge Name: uncalabrated.jpg Size: 324 KB
the markets moving, time to get in!
  • #1,025
  • Mar 9, 2009 12:16am Mar 9, 2009 12:16am
Rule: No Yellow MA

No yellow MA

All you do here is scale down toward 1M chart until you just can’t see the yellow bar.
When you do this you will have a fairly aggressive and easy enough to manage
I like this area because I can get
more trades for my aggressive stuff.

In this chart I have a questions on the EA

In this I’m pointing out a yellow bar. But this yellow bar is just a little piece on my chart.
So my questions are can I set my TF to 20 for a closer setting ?
more charts daddy
Attached Image(s) (click to enlarge)
Click to Enlarge Name: no yellow bars calabrated.jpg Size: 303 KB
Click to Enlarge Name: no yellow mid calabration = 20 TF.jpg Size: 277 KB
the markets moving, time to get in!
  • #1,026
  • Mar 9, 2009 12:18am Mar 9, 2009 12:18am
More conservative

The charts height remains the same but the length grows, exposing more bars and in many cases this
Action Exposing yellow bars, Forcing me to move down the TF list.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: long same TF withyellow.jpg Size: 342 KB
Click to Enlarge Name: new calabrated long no yellow.jpg Size: 293 KB
the markets moving, time to get in!
  • #1,027
  • Mar 9, 2009 12:33am Mar 9, 2009 12:33am
Thanks Bob, this is a great system. my account was up 250% last week (very small account) on ibfx , has anyone had luck with gotrader?
cheers
darren
  • #1,028
  • Mar 9, 2009 12:38am Mar 9, 2009 12:38am
Rule: Make them look similar

This method you are trying to keep the overall look similar
One method is keep a similar or closest to similar distance inbetween the MAs
(look at pic notice the distance is similar in all pairs at least as close as you can get)
except the 100x80 pair

This you do by looking at it or you can count how many pip distances.

(this case i tried to stay 300-500 pips for chart calibration)

you can do long calibration for a more conservative effect.
Attached Image (click to enlarge)
Click to Enlarge Name: calabrated to look similar.jpg Size: 328 KB
the markets moving, time to get in!
  • #1,029
  • Mar 9, 2009 12:41am Mar 9, 2009 12:41am
Hi all,

For anyone interested, attached the Tradestation code to implement anywhere from 1 to 5 levels and any combination of settings. Although I use MT4 I find the testing ability inadequate and timeconsuming. TS very much better.

Also attached performance curves for 3L 90x60 with TF as per the first post for the main 7 currencies against USD for 1 year of trading each. I've optimised the stoploss for each pair which turns out between 300 for AUD to 490 for JPY. Of course you only know the optimum after the event so these curves represent the optimum over 1y and you wouldn't achieve as good.

Point of this is that the method appears basically untradable. You'd have to be pretty amazing to filter out the losing trades just by looking at news days.

Happy to test and post any combination of currencies and settings anyone wants. Each test takes around 5 sec.

cheers
Knoxy
Attached File(s)
File Type: txt SnapBack.txt   5 KB | 472 downloads
File Type: zip EquityCurves.zip   1.1 MB | 554 downloads
Aspice, officio fungeris sine spe honoris amplioris
  • #1,030
  • Mar 9, 2009 3:10am Mar 9, 2009 3:10am
Dear Knoxy,

Thanks very much for the graphs. Some questions, if you don't mind:

1) Did you optimize by not just ultimate SLs but also pips distances from the MA line as well as the various amounts of martingales?
2) Can you do the graphs for EURCHF and EURCAD and say what the optimized settings were?
3) The steep equity drops seem to be a result of not just the poor win-loss ratios but also aggravated by the martingales. Do you think the results would have been improved if one were to average the lots and not martingale them, like on the Spieler thread?
4) I'm sorry to see the results are so poor. Anyone out there want to comment on this? ---Especially disappointing because the theory seems sound.

Lenoxer
  • #1,031
  • Mar 9, 2009 3:13am Mar 9, 2009 3:13am
Quoting Knoxy
Hi all,

For anyone interested, attached the Tradestation code to implement anywhere from 1 to 5 levels and any combination of settings. Although I use MT4 I find the testing ability inadequate and timeconsuming. TS very much better.

Also attached performance curves for 3L 90x60 with TF as per the first post for the main 7 currencies against USD for 1 year of trading each. I've optimised the stoploss for each pair which turns out between 300 for AUD to 490 for JPY. Of course you only know the optimum after the event so these curves represent the...


Nanningbob has said hundreds of times that backtesting is useless on this system. Seriously, read the whole thread before posting something like this.
  • #1,032
  • Mar 9, 2009 4:32am Mar 9, 2009 4:32am
Thanks for comments Lenoxer and Toyoproxo,

I need to get up on my soapbox. Best to use ones mind and not 'play the man'. None of us is ever completely right and that's why we have forums to learn from each other. I HAVE read the complete thread also the AUDNZD thread. I've demo traded this system with GoMarkets and FXCM and I'm keen to see it work, why wouldn't I? Please don't presume Toyo and there's no need to insult anyone.

Neither Bob nor I can control what is a fact and that is that the markets in the last year behaved exactly the way that the historical bars show, at least with my broker. If you had have started in Jan last year with the EA and traded the market in automated fashion the result, subject to slippage, would be the performance curves given. Maths is deterministic, there's no voodoo involved, as long as the rules can be written, backtesting and forward testing will work. I believe the old saying 'those who learn nothing from the past are destined to repeat it', so I test.

It's the discretion of knowing when to stand aside and avoid the massive losses of a few trades that would make the difference to this method. That's what I'm interested in discussing. If others aren't then fair enough, but I won't trade a method until I fully understand and test it.

Lenoxer, late at night here. I'll look at the pairs suggested and at optimising. Anything can be made to look ok with curve fitting, but I'll do some work on universalisation across a few pairs. The answer is usually to apply similar numbers to different pairs then walk forward optimise.

cheers for now
Knoxy
Aspice, officio fungeris sine spe honoris amplioris
  • #1,033
  • Mar 9, 2009 4:48am Mar 9, 2009 4:48am
as i was discussing with dmjservices earlier tonight, the problem is on such timeframes there may be a 1hour or 30min bar that goes up a heap, then retraces to tp then goes straight back up but isnt really reflected in the bar...

thats why i keep every currency on the 5min chart so i can see a 'truer' indication.. when i visually backtest
Grabbing the bull by the horns!
  • #1,034
  • Mar 9, 2009 6:04am Mar 9, 2009 6:04am
Agreed Whitegoodman, got to think about the test method. Advantage of TS is can easily have the 30m chart and 1m chart on the same screen, setup the limit and stops on 30m and take the entries and exits on the 1m. Haven't done so for this as it takes a bit more coding, but will do when get serious.

Lenoxer, graphs attached for EURCHF and EURCAD. Both optimise best on 30m. EURCHF seems to have a tendency to retrace deeper, so optimal is to increase TakeProfit level to 90 rather than 60. Found that only the two trends in Dec upset EURCHF. When it gets 3 levels out has a strong tendency to retrace so the L3 trade can be strong. The attached is for 1,3,9 the default and 470 SL. I found changing to 1,3,27 same shape but increases the 1y profit to 39.5k points with the drawdown 12.8k. So the L3 size seems dependent on max drawdown you can handle. Promising if one could find an effective way of filtering trending behaviour in Dec.

For EURCAD, couldn't believe the curves so tested for 2 years as attached. Looks very good. Settings here are standard 90x60 with SL 400 and standard 1,3,9 lots. Increasing the L3 lots again seems the way to go to increase profit but depends on drawdown acceptable. Note the 700 trades in 2y, I haven't included the spread which would subtract but not dent it too much. I'll look at trading this.

cheers for now
Knoxy
Attached File(s)
File Type: zip EURCrosses.zip   280 KB | 427 downloads
Aspice, officio fungeris sine spe honoris amplioris
  • #1,035
  • Mar 9, 2009 6:31am Mar 9, 2009 6:31am
Quoting whatchee
I am not sure reading the logic correctly

for AllTF_NanningBob 60x90_HiddenTP.ex4

the take profit seems hard coded, taking 100 point profit not taken because it is hard coded to take profit at 60 points, live testing sees order closed automatically at 60 points profit

case OP_BUY : if( OrderType()==OP_BUY && (Bid-OrderOpenPrice())/Point>=60)

there is another place has similar coding
I'm using the original ea from post #1, where the tp is adjustable, as are all the other variables. I haven't gotten into this whole hidden tp thing, but I guess if I did I would need to change the tp somehow. Does anyone know if the hidden tp in the new ea can be changed?
  • #1,036
  • Mar 9, 2009 7:13am Mar 9, 2009 7:13am
Quoting llambert3
JT, let me get this straight.. you are getting in the 1st level 90 pips away from 80ma and going for 100pips with a 400pip stop. Then the 2nd and only other level is 200 pips away. So if you start at 1 lot (for simplicity), you will gain $100. If it hits the 2nd level you are getting in 3 lots for 100 and this level would have a 200 pip stop. So for the stops we have

1st level -400
2nd level -600 (-200 * 3)

Total -1000

If you win:

1st level +100

going 2 levels -100 + 300 = +200

So basically if your trades only go 1 level,...
I use whatever original tf and ma were contained in the profile from post #1. I don't presume to know enough about technical analysis to change timeframes and ma's.

The thing I hate about the strategy tester is you can't see what your entire basket of pairs would have done together as a group in a single week. You would have had 21 other pairs trading those weeks as well. I wonder how they did but I'm too lazy to sit here and strategy test 21 pairs when I can just look at my statement every week in real time. I did backtest several pairs at 100x200 400 sl at random periods in the past and they always seemed to do pretty well overall. I'm not as wise on the activity of certain pairs as some others on this thread, but I do know the eur/jpy had 5 1level tp's last week and one trade open on Friday. I'm sure there will be 2 level trades in the future and my gut tells me over the course of a year even a heavily trending pair will at least recover some or most losses and maybe even make a profit. I prefer to look at the whole basket of pairs as a single investment rather than sweat on one of them. None of us knows what any pair will do next week so cherrypicking seems like a waste of time. So maybe January and February were a really bad month for that pair. Remember this wasn't entirely random, the Japan PM came out and made a big announcement on a Wednesday and all the Japan pairs trended heavily. This happens to the euro and the usd several times a year as well. We can't predict the future, so my stategy is the more ongoing trades with as many pairs as possible the better. As I mentioned before, last week netted me $650 (demo, too bad) and drawdown was never over $150. Of course that's an unusually good week but I'm sure you have great weeks, good weeks, and bad ones. The occasional sl's are inevitable, but they are quick, decisive, unemotional, and relatively small.

I would love to hear from anyone demoing this alteration of Bob's fantastic system (thanks again, Bob) with different timeframes and entry points off the ma. Also comparisons with different brokers to see how spreads affect the p/l.
  • #1,037
  • Mar 9, 2009 8:16am Mar 9, 2009 8:16am
Quoting Knoxy
Hi all,

For anyone interested, attached the Tradestation code to implement anywhere from 1 to 5 levels and any combination of settings. Although I use MT4 I find the testing ability inadequate and timeconsuming. TS very much better.

Also attached performance curves for 3L 90x60 with TF as per the first post for the main 7 currencies against USD for 1 year of trading each. I've optimised the stoploss for each pair which turns out between 300 for AUD to 490 for JPY. Of course you only know the optimum after the event so these curves represent the...
Knoxy -Was this testing done manually or with a tester? Did it test with a shut down over each weekend? Thanks much
  • #1,038
  • Mar 9, 2009 8:35am Mar 9, 2009 8:35am
Cable is giving this method a beating right now.
  • #1,039
  • Edited 10:01am Mar 9, 2009 8:54am | Edited 10:01am
knoxy, thanks for the info, as I think reports on possible or even probable disaster scenarios are far more important than only the good news. Nanningbob has said that bact testing is unreliable for a number of reasons. One reason is turning the EA off for the weekend and avoiding news days, particularly NFP. If you take October out of the mix, you get some very different results. I presume your test included October. I also know personally that mid December wasn't the best for this system. Also, if you look at January, there were some rough patches depending on the pair you traded.
If I gave you my settings for distances away on the various, could you run a test? I believe one way to avoid these disastrous drawdowns, which are as a result of being completely stopped out on a pair(s), is to enter at further distances away than 90 on every pair. In some cases, 120, 130 or 140. If you could, I'd appreciate it, and might be of help to others here.
  • #1,040
  • Mar 9, 2009 8:58am Mar 9, 2009 8:58am
nutsford, I agree. That's why I exclude GBP/USD and GBP/JPY. However, I get an alert when cable is 250 from center, as it was awhile ago, and I make a decision as to whether or not to make a single trade or an EA trade.
Quoting nutsford
Cable is giving this method a beating right now.
3 traders viewing now