MNB 10.4 Streamlined by Snailbeard

[snailbeard]

Update on CSS

After more testing over a longer timescale the results from CSS were disappointing, forward testing was also a disaster. It appears to be due to the common issue of not detecting when a pair leaves the main trend and starts to range. An interesting diversion, and not totally wasted as I had to create some useful utility code for manipulating arrays.

Also, CSS crossovers occur at in sync with mid-Bollinger crossovers so one can simply refer to the Bollinger without the extra overhead of CSS.

[snailbeard]

The Holy Grail - could this be it...

The 4TF MA indicator looks very promising.

Now that I have spent some time on visual back-testing: it seems to work quite well with RSI2.

In the attached screenshot I have labelled the interesting bars:
MNB uses pending orders but I also added some logic for placing instant orders, so we could place either type of order.

Lets say that we are still using the NB H1 960 MA as our major trend, so here we are under it and looking only for a SELL (At some point we could say that the 4MAs allow us to dispense with the H1 960 MA - but for now lets take one step at a time)

At Line 1
RSI2 is pinned at the top
4MAs are all red
Option A - enter a pending order at the next level below the price (1.5185)
Option B - instant order - jump in with both feet now

Move forward a few candles and the fastest MA reverses into an uptrend:
How should we respond?
Is your style to use wide stop losses - then just ride it out - 3 lights are still red so this is probably a minor pullback.
Is your style to minimise drawdown - for basket trading it is preferable not to lock out potential winners because of drawdown - so my preference is to
A) delete a pending order
B) exit an instant order with a small loss
C) Do more testing to see if it is really beneficial to exit with small losses or better to ride the pull-back.

Moving on a few more bars: line 3: green switches back to red: RSI 2 is still pinned at the top so
Option A - reenter the pending order
Option B - a second instant order or are we still riding the first one?

Note:
If we had left in the first pending order in: it looks like it would have been triggered by a spike, followed by a reversal of about 65 pips - adding to drawdown and possibly blocking a better trade or worse hitting a stop-loss.

Moving on a few more bars: Again one red goes to green at 4, lets say we kill the orders
and then Holy $%^&! at 5 it goes red again, but now we have lost RSI2 - so no new pending or instant order!

Now in an alternative universe we don't kill the orders -
which means the earlier order is still active or hit a stop-loss.

In the case of a stop hit we got a new order at 3
In the case of a pending order: it gets triggered at 6

Now we have finished with entry logic and we are thinking about exit signals:
The first target is at green WPS1but the price bounces off strong support twice and at 7 one red switching to green - should we exit?

Moving on we see there is a sudden strong reversal going past WPR1.

Summary
There is a lot to take on board with all these different permutations
However, perhaps we can distil it into few sentences:

Say we use raw NB
A) with big stop losses of 150 pips then we are quids in with say 90 pips
B) with modest stop loss of 50 pips then we are a double loser: - 100 pips
However, with raw (automated) NB the long term gain is quite modest - according to BT

Say we use bail out on green lights
Pending orders - no wins, no losses
Instant orders - (-4 pips) then (-14 pips) so about -20pips

In the next post we look at some positive results.

[snailbeard]

Something I have mentioned in the past is that NB stop losses sometimes make great entry points:

Even better if you didn't hit the stop loss in the first place!

The purpose of RSI2 is to indicate that the price is pulling back or has pulled back and this is the crux of ideal trade - correctly entering when the current pull back has finished. So this new screenshot follows on from the previous one

In the raw NB universe with a big stop loss you are currently in significant drawdown.
In the alternative universe we abandoned those orders and find ourselves at Line 1 with RSI at the top and 4 red lights:
Option 1: using pending orders - we get one at around 1.5185
Option 2: instant order at around 1.5255

We get a perfect Captain Jack move all the way down to 'a' - almost 400 pips!
Then we get the bounce and several things could happen depending on your trailing and profit logic:
Firstly, I have been using the profit lock and there is 50 pip retrace to 'b' which might trigger an exit giving 350 pips.
On the other hand if bounce didn't exit the trade then there is a 2nd touch of strong support good for a manual exit.
Then there is '3' : red MA changing to green which could be used to signal an exit at around 1.4890.

In summary

Instant Order Profit 1.5255 - 1.4890 = 365 pips
Pending Order Profit 1.5185 - 1.4890 = 295 pips

Accounting for earlier Instant Order losses we reduce profit to 345 pips

Conclusion

This is worth back-testing with more pairs and over long periods to find out if this would effective in the long term.

[rosst]

I have this indicator but I am not sure it is the same as the one you are using - I will look at it on some charts too...

[theforexedge]

Brian

Did you manage to have a look at Baluda's Global Market Trend indi?

I've attached a image to show how i've got it setup to give me a visual indication over D1,H4 & H1 TF.

It seems that anything under 2.5 is showing that the strength of the trend is reducing and a good indication to stay of of the market..

My coding is very basic right now, but I was wondering if this indi could be called using iCustom in a EA?

I've made a simple mod the the original indi that sends a SendNotification when the 2.5 Threshold is broken

if ( currencyValues[j] > Threshold)

Any help or guidance much appreciated..

The exhaustion candle idea is also a good direction, as I trade live using VSA / Wykoff principles and the Upthrust or Test After Shake Out is always a sure winning trade..

Jason

[MrLong]

Brian

Did you manage to have a look at Baluda's Global Market Trend indi?

I've attached a image to show how i've got it setup to give me a visual indication over D1,H4 & H1 TF.

It seems that anything under 2.5 is showing that the strength of the trend is reducing and a good indication to stay of of the market..

My coding is very basic right now, but I was wondering if this indi could be called using iCustom in a EA?

I've made a simple mod the the original indi that sends a SendNotification when the 2.5 Threshold is broken

if ( currencyValues[j] > Threshold)

Any help or guidance much appreciated..

The exhaustion candle idea is also a good direction, as I trade live using VSA / Wykoff principles and the Upthrust or Test After Shake Out is always a sure winning trade..

Jason

Hi Jason,

What timeframe would 2.5 represent?

I've hooked the GMT function into my latest EA and have been monitoring the reading on a H4 timeframe.

Code: Select all

//+------------------------------------------------------------------+
//| getGlobalMarketTrend()                                           |
//+------------------------------------------------------------------+
double getGlobalMarketTrend(int timeframe)
{
   double currencyValues[8];
   int i;
   // Get Slope for all symbols and totalize for all currencies   
   for ( i = 0; i < PairIndex; i++)
   {
      double slope = GetSlope(Pairs[i], timeframe, 0);
      currencyValues[libCSS.getCurrencyIndex(StringSubstr(Pairs[i], 0, 3))] += slope;
      currencyValues[libCSS.getCurrencyIndex(StringSubstr(Pairs[i], 3, 3))] -= slope;
   }
   double globalMarketTrend = 0.00;
   for ( i = 0; i < 8; i++ )
   {
      // average
      currencyValues[i] /= libCSS.currencyOccurrences[i];
      globalMarketTrend += MathPow(currencyValues[i], 2);
   }
   return ( globalMarketTrend );
}

Andy

[theforexedge]

Andy

Thanks for the code snippet

I wanted to experiment using both H4 & D1, but not sure if you have found that D1 is too lagging and prevents too many trades?

As you know my coding skills are quite limited, but what would i just need to change to get the D1 TF?

Thanks, Jason

[snailbeard]

I have this indicator but I am not sure it is the same as the one you are using - I will look at it on some charts too...

The one in my screenshots is buggy, I'll upload a comparison with my variant (which seems to be correct - except after the daily trend should be the weekly trend which is not much use when looking at a H1 chart and too long term to be useful in the range of pips that we are interested in, so I have just made TF4 also show days - not a bug unless you really wanted to see the weekly trend).

[snailbeard]

Andy

Thanks for the code snippet

I wanted to experiment using both H4 & D1, but not sure if you have found that D1 is too lagging and prevents too many trades?

As you know my coding skills are quite limited, but would i just need to change the int timeframe for D1 etc?

Thanks, Jason

Yes, Thanks Andy, I'd like to follow up on this when my current experiments are finished.

[snailbeard]

Here is a better version of the 4TF MA in comparison to the previous one.

Note the staggered change which occurs when the a trend is building for a longer stretch.
Perhaps this pattern can be used to gauge the strength of a move: extend targets and relax exit strategy, let winners run longer etc.