stevehopwoodforex.com
https://www.stevehopwoodforex.com/phpBB3/
Print view

MultiBob, a multi symbol NB10.2(a) autotrader
https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=827		 Page 42 of 53
Author:  spotdespot [ Wed Oct 10, 2012 3:28 pm ]
Post subject:  Re: MultiBob, a multi symbol NB10.2(a) autotrader
Ok - I have spent the last couple of hours looking through trades that went bad and trying to spot any similarities/ideas for ways they could be avoided. It seems to me that in quite a few cases this might be possible and have come up with a few ideas for people to consider, with examples.

10.2A (pivot crosses)

1. Simply create thresholds for D1 TMA before a trade is valid. So for a pivot long D1 TMA must be greater than X (say 0.3 as default). Vice versa for pivot shorts. This in itself would have prevented quite a few failed trades. It would also prevent a few successful trades too - but I don't mind missing out on the odd one.

2. Close any trade where the daily pivot crosses the weekly pivot in the wrong direction - i.e. so close when the condition of D>W>M OR D<W<M is no longer true.

3. This one may be more challenging - do not take trades when the angle of the D1 TMA is against the trade - so if we are looking at a long pivot cross yesterdays D1 TMA should be less than today's(current) D1 TMA.

Some examples - red lines show entry/exit where appropriate. In the bottom window the black line is the D1 TMA.
ej pivot long 10.2a.gif
There is an issue with the jumping stop here for some reason. I will dig out the logs to see if I can track it down (seems ok on 10.2)
usdchf pivot short 10.2a.gif
usddkk pivot short 10.2a.gif
One we would have missed.
eurnok pivot long 10.2a.gif

A close on the d<w<m test would have saved a few pips.
audnzd pivot short 10.2a.gif
I will split 10.2 in to the next post.

EDIT: Sorry added the same chart twice under - "One we would have missed" now corrected.
Author:  spotdespot [ Wed Oct 10, 2012 3:57 pm ]
Post subject:  Re: MultiBob, a multi symbol NB10.2(a) autotrader
Ok continued from the last post - now looking at 10.2 Multibob:

It seems that the vast majority of trades that had failed that I looked at were triggered by a retracement to the MAs that simply continued in that direction. It seems to me that if we only allowed the trade when slightly stricter criteria were met then we could avoid some of the losers. What that criteria is exactly needs to be discussed but here are a few ideas for longs (obv vice versa for shorts):

1. Price must close below the 5MA 2 shifted (the middle one) (or 5MA 1 shifted - the top one) and then the trade is triggered when price crosses back above the 5MA 1 shifted (the top one in this case) OR
2. Price must move below the 5MA 3 shifted (the bottom one) and CLOSE above the 5MA 1 shifted to trigger a trade OR
3. Price must move below the 5MA 3 shifted (bottom one) and then move above the 5MA 1 shifted to trigger the trade

Another idea would be to close any trade where a bar closes on the wrong side of the MAs. So if we take a Long D1 trend trade and it triggers as per 1, 2 or 3 above then during subsequent days it closes below the lowest MA we close the trade. (EDIT: I should add if the TMA criteria etc are still met we lose little if price continues North as we would get back in as price breaks above the top MA)

Apologies for the messy bottom window - the black line in the D1 TMA on all H4 charts, the BLUE line is always the current timeframe's TMA (so on D1 the D1 TMA is the blue line). The red lines' intersection is when the trade was taken. I haven't made any notes on these as they are more self-explanatory I think (actually added on to the last screen shot).
audchf h4 trend 10.2.gif
gbpaud d1 trend 10.2.gif
audnzd h4 trend 10.2.gif
Anyway - I hope there is some food for thought there. It would be a shame for such a fantastically well written EA with huge potential to fall by the wayside.

Cheers,
Dave.
Author:  Don't Jump42 [ Wed Oct 10, 2012 10:18 pm ]
Post subject:  Re: MultiBob, a multi symbol NB10.2(a) autotrader
Ok. As I understand The rules of engagement for a manual trade... And please correct me if I'm wrong. If slope is at or above 0.4 or or at or below .4 and daily weekly and monthly pivots line up correctly. We should then verify market direction on lower time frame charts. Then and only then should a trade be taken. For the last three days D1 slope values and H4 slope values have not even been close. If you back it down to H1 the charts disagree even more. I realize this is not a scalper. No such thing as fast or easy money. Given the right trade entry this thing WILL be a juggernaut taking advantage of long term market moves and avoiding reversals. It would have fewer trades, but they would be quality trades. At that point you can use MPTM to manage exits. I am a newbie, so take it for what it is worth. Maybe just food for thought. I am running 10.2 on one $3000 demo at FXCM and 10.2(a) on another $3000 demo at FXCM. So far this system has made the most sense to me as a trading method.
Author:  cuzgeorge [ Wed Oct 10, 2012 11:09 pm ]
Post subject:  Re: MultiBob, a multi symbol NB10.2(a) autotrader
gaheitman wrote:
cuzgeorge wrote:I havent had time to go over all the results yet, will do so one the week-end, but i can say that i'm dumbstruck that some of the positions have hit the 200 pip SL. There is clearly something wrong even with 1.1.11. Sorry for the very general post but will elaborate when i study the results better. When trends change, i assumed the EA would close the position, but clearly its not doing that, even after 200 pips have passed. I'm running some with Correlation and one or two with-out correlation. Will compare and post when checked.
regards
cuz
The only code I see that is closing trades is based on TP and SL. Are there specific rules in the system for closing under other circumstances?


George
Hi George,
i only saw your post much later, sorry to only get back to you so late.

As i understand 10.2A, we trade the pivots and if a position is entered, it will need lots of breathing space, therefore allowing a 200 pips SL. i accept that a position will move up and down, and probably go into a losing for some time. Based on the method from NB, when i look at the indicators, i hardly ever or never see a 200 pip movement in the opposite direction to the method. So the way i set up with my BE really close to the entry point, (20/10), i would expect that my BE is hit after a short time. JS would be a bonus. I really just look for BE which has a little profit there too. I'd be happy with that.

With 10.2 , i set up the same scenario, using the 0.4 as buy, and 0.8 as buy and hold. But if the trend or value should fall to say 0.3, or 0.2, then i would want to get out of that position.

Code: Select all

  if ( symbolValues[i][iH4Slope] >= 0.4 )  symbolValues[i][iH4SlopeTrend] = buyonly;
      if ( symbolValues[i][iH4Slope] <= -0.4 ) symbolValues[i][iH4SlopeTrend] = sellonly;
      if ( symbolValues[i][iH4Slope] >= 0.8 )  symbolValues[i][iH4SlopeTrend] = buyhold;
      if ( symbolValues[i][iH4Slope] <= -0.8 ) symbolValues[i][iH4SlopeTrend] = sellhold;
I'm only referring to the code to describe what i'm trying to explain, but my next test set here is going to be changing the values to NB's slope to 3 and 6 that he has been testing. I see Paul has'nt got the third option coded that Steve allows in his EA's. I would set the LTF to bump me out of the trade if it gets as low as say '0'. Then i accept that the movement was not strong enough to start off with and would rather be out of the move. I know the ranging parameters are supposed to keep us out, but somehow we get open positions at the wrong times/ values.
Its a starting point for next test session, and i would play with the settings trying HTF and lowerTF, typical tinkering. I'm going to wait out this week and see what PAul comes up with before i start changing code and probably mucking things up, :lol: :lol:

I'm sure Paul's been very busy lately, so hope i'm not confusing and side-tracking everyone. HAve'nt been through the code much but i think it's time to start.

Regards and thanks for the post George, always a pleasure to read your thoughts.
Cuz.
Author:  wBear [ Thu Oct 11, 2012 2:45 am ]
Post subject:  Re: MultiBob, a multi symbol NB10.2(a) autotrader
As of yesterday, I was testing NB10.2 turned on (settings per screenshot), NB 10.2a off.

Seems like maxPipsCrossed is not working. This morning, MB1.1.11 took a GBPUSD Sell position 20+pips after DP despite default setting of 5.

Is my understanding about maxPipsCrossed not what it should be or is this a bug?
Author:  Gsasha1969 [ Thu Oct 11, 2012 6:52 am ]
Post subject:  Re: MultiBob, a multi symbol NB10.2(a) autotrader
MultiBob You can drive in the tester. To do this, install only one pair
Author:  fxozgirl [ Fri Oct 12, 2012 11:54 pm ]
Post subject:  Re: MultiBob, a multi symbol NB10.2(a) autotrader
My Multibob demo is up 13.09% for the month including a very small increase of 0.21% this week, even though conditions were not ideal. Plenty of drawdown though, but it is reducing...looking forward to next week to see if MB can recover completely :D

ps I had been running v1.09 and updated to 1.11 on Friday morning
Author:  phil_trade [ Sat Oct 13, 2012 8:54 am ]
Post subject:  Re: MultiBob, a multi symbol NB10.2(a) autotrader
fxozgirl wrote:My Multibob demo is up 13.09% for the month including a very small increase of 0.21% this week, even though conditions were not ideal. Plenty of drawdown though, but it is reducing...looking forward to next week to see if MB can recover completely :D

ps I had been running v1.09 and updated to 1.11 on Friday morning
Hi FxOzGirl

Could you share your setting please ?

Did you intervene on opened trades ?

Philippe
Author:  ahniel [ Sat Oct 13, 2012 1:46 pm ]
Post subject:  Re: MultiBob, a multi symbol NB10.2(a) autotrader
fxozgirl wrote:My Multibob demo is up 13.09% for the month including a very small increase of 0.21% this week, even though conditions were not ideal. Plenty of drawdown though, but it is reducing...looking forward to next week to see if MB can recover completely :D

ps I had been running v1.09 and updated to 1.11 on Friday morning
Wow you got some magic, cause I'm down 12% last week.
Author:  garyfritz [ Sat Oct 13, 2012 3:17 pm ]
Post subject:  Re: MultiBob, a multi symbol NB10.2(a) autotrader
Yeah, I was down about 18% when I pulled the plug on it.
All times are UTC Page 42 of 53

Powered by phpBB® Forum Software © phpBB Limited