Dashboard TZ

some time, some moments
automated ?

Yes. I let run some weeks without modify anything for check mistakes and yesterday I applied some fixes. I only got a partial bad trade before CAD news ( GC ). It was 70p spike then it's inside atypical cases.

My main task is system be able to follow long trends as last weeks. As swing strategy suffers in some pullbacks

This is kind of entries I try force. Normally you get buy as swing strategy system. But it's third sell I got since yesterday. Previous were winners and this we will see ( it have a gap area around 75% level a few days ago )

How a simple conversation question ends in a coloful drawing analisys

Hi Brain
a good setup and patience. Price line alert and trade button, it's your creation?

No. It's FF member capture. Not idea if it's free stuff or private

Improving MAs framework...

Blue is original SMA 12


Blue is SMA 12 with 2 pre/post filters. Black is same with a improved pre/post filter

I'm downing trades in every iteration. Profit per trade growth much in current too ( +20p average ). Still not sure how deal with bad trades. Question is where set the limit. Anyway if day profit csn be constant and quite big lke now I can accept some -100p without drain profits

A way for get reentries or hedge is use a different setup. As setup have different behavior then new entries are less predictable or mechanical that using reentries with same setup

Making some week analysis. This show effect of close trades early. System make mistake at third trade. Not closing any of previous trades profit it will be big.

Other interesting thing are trades DD. Question is always same. When a trade is really a bad trade. -100p is a big DD but if you keep it enough finally can be a big winner as show capture. Evidently we have same in opposite side. Last trade didn't recover making very big DD. Then is important find a way for guess which trade will not be recoverable and close it fast while you keep some spike trades enough time for go in profit

While i find a reliable way for manage bad trades i was making some numbers. Im making in manual around 200/250p. That moves around 15/20% month. Last setup made 5000p in 12 days. Is not max profit because many closes are very early ( more 100p missed sometimes ). Anyway at this moment and if can maintain similar performance i got:

1) 20 times more profit in half time.
2) if 250p are around 20% then its 400%

Then performance is good. Now i need avoid bad trades kills account...

Last 2 setups are working better. High profit per trade and low bad trade ratio. Recovering is going fast and profit per day had a big boost. Number trades is still stable. I had 30 yesterday but now down to 23 again. Below 20 is optimal

If continue like now is candidate for use with stack MM

Constant profit covering DD. This means mistakes can be covered over time while DD reach a peak and not growth more

One of setup running have multiple entries combinations. Last weeks only one made entries unbalancing system because the others almost didn't have weight in trading. I made a small adjustments and now all have similar weight.

Checking MA framework implementation power.

Red is original ZeroLag Tema ( 48 ). Black is ZeroLag MA ( 12 ) + Tema ( 24 ). MAs are mixed with some special processing between them. Its clear the effect. Similar reaction but smoother.

This implementation allow make multiple mixes of different MAs getting the MA properties in every pass. Then you can have faster MAs with more smooth and more adaptative behaviour depending pair needs

Here difference between ZeloLag MA ( blue ) and ZeroLag + Tema ( black )

SMA 12 with 1000 bars. Difference between iCustom ( 100ms ) and direct process ( 2ms ). 50x slower...

A few numbers about what means this difference in performance:

1) 4 server instances
2) 28 pairs per instance
3) 2 MAs per pair
4) 2000 bars

With iCustom this is:

1) 2 MAs per pair: 2 * 200ms = 400ms
2) 28 pairs * 400ms = 11200ms ( 11 seconds )
3) 4 instances: 11s * 4 = 44s

With direct process this is:

1) 2 MAs per pair: 2 * 4ms = 8ms
2) 28 pairs * 8ms = 224ms ( 0.22 seconds )
3) 4 instances: 0.22s * 4 = 0.88s

We pass from an eternity to almost lightning load...

Normally we complicate things so much trying have perfect system. A one rule system with big delayed signals sometimes is more profitable and safer

Sometimes I dont have enough space with default dashPIN groups ( 5 ). Now I can add until 25 groups