BtB Workspace | Page 7 | Forex Factory

Options

Go to
Permalinks
Print Thread
Attachments: BtB Workspace
Exit Attachments

BtB Workspace

  • #121
  • Jul 21, 2017 11:47am Jul 21, 2017 11:47am
Quoting braintheboss
BtB SDK ( v0.10 ) - See post #1 ( BtB Develop ) + added PIN protocol + added Symbols manager {image}
Actually B. we were talking with husky other thread about Mt5 and reminded me to ask you
if i am loading the history the right way as now it is loading Css and Snr but they do not display on chart
due to lack of history_ now history i see it is in .hcc rather than .hst
I know that on Mt5 the history is stored in M1 but it is loading history on UsdCad M1 and it has not completed yet after 15 mins_
I got a script that i am using for Mt5 and it is not as it used to be for Mt4 .. Loading all pairs ? here s pair by pair ?
I went into tools/options/charts and have enabled the Bars to unlimited too ,,,
Any tips for this ? Thank you
Attached Image(s) (click to enlarge)
Click to Enlarge Name: history css snr for mt5.jpg Size: 231 KB
Click to Enlarge Name: history loading for mt5.png Size: 1.0 MB
The arriving better than the arrival, the journey has just started
  • #122
  • Jul 21, 2017 12:45pm Jul 21, 2017 12:45pm
Quoting jenbols
{quote} Actually B. we were talking with husky other thread about Mt5 and reminded me to ask you if i am loading the history the right way as now it is loading Css and Snr but they do not display on chart due to lack of history_ now history i see it is in .hcc rather than .hst I know that on Mt5 the history is stored in M1 but it is loading history on UsdCad M1 and it has not completed yet after 15 mins_ I got a script that i am using for Mt5 and it is not as it used to be for Mt4 .. Loading all pairs ? here s pair by pair ? I went into tools/options/charts...
Its working well. The problem is download 1M bars x 28 pairs. Even in capture only have +300K bars is savage. You can need long time for download all and will use a lot CPU ( data is compressed ). Put less 50000 bars in MT5. I'm using 5000

PS: In capture you only got 4 pairs. You can guess how long can be download. You dont forget restart MT5 after change history bars
Try don't lose pants never...
  • #123
  • Jul 21, 2017 1:02pm Jul 21, 2017 1:02pm
Thanx, so instead of simplifying things to save time they make it longer ?
I found this and thought this script was to load history fast .. but mhmh not really tried already
i dont attach here i just send link , but that is for the use of .csv
https://www.mql5.com/en/code/232

I tried to put chart on 1 month then press the up keys and make the chart go to the left without the scrolling
and did all the rest of TFs W1/D1/H4 up to H1 until i saw no joy and will postpone it to other days to do
maybe until that day someone will invent a tool to do the same as load historical data same with Mt4 other script...
but i have to say good to have started with Mt5 so can discover new things ,,, and be ready for the future

Quoting braintheboss
{quote} Its working well. The problem is download 1M bars x 28 pairs. Even in capture only have +300K bars is savage. You can need long time for download all and will use a lot CPU ( data is compressed ). Put less 50000 bars in MT5. I'm using 5000 PS: In capture you only got 4 pairs. You can guess how long can be download. You dont forget restart MT5 after change history bars
The arriving better than the arrival, the journey has just started
  • #124
  • Jul 21, 2017 1:06pm Jul 21, 2017 1:06pm
Quoting jenbols
Thanx, so instead of simplifying things to save time they make it longer ? I found this and thought this script was to load history fast .. but mhmh not really tried already i dont attach here i just send link , but that is for the use of .csv https://www.mql5.com/en/code/232 I tried to put chart on 1 month then press the up keys and make the chart go to the left without the scrolling and did all the rest of TFs W1/D1/H4 up to H1 until i saw no joy and will postpone it to other days to do maybe until that day someone will...
My tools download automatically all data needed. You dont need do anything. Only you dont put big history. I had a history with more 7Gb in that folder only using small bars history
Try don't lose pants never...
  • #125
  • Jul 21, 2017 1:09pm Jul 21, 2017 1:09pm
Quoting braintheboss
{quote} My tools download automatically all data needed. You dont need do anything. Only you dont put big history. I had a history with more 7Gb in that folder only using small bars history
Oh oh .. you always got a solution Mr BtB ..
Errhmm we might have a private chat then
The arriving better than the arrival, the journey has just started
  • #126
  • Jul 21, 2017 2:06pm Jul 21, 2017 2:06pm
first shot
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Screenshot_3.jpg Size: 646 KB
Click to Enlarge Name: Screenshot_4.jpg Size: 552 KB
Take what you can, give nothing back
  • #127
  • Jul 21, 2017 2:25pm Jul 21, 2017 2:25pm
Quoting Husky77
first shot {image} {image}
looks good order and colors
The arriving better than the arrival, the journey has just started
  • #128
  • Jul 21, 2017 2:33pm Jul 21, 2017 2:33pm
Quoting jenbols
{quote} looks good order and colors
Now only left the "toy"
Try don't lose pants never...
  • #129
  • Jul 22, 2017 3:31am Jul 22, 2017 3:31am
I will explain how will work the new trader ecosystem will begin with Trader PIN. This new system is a professional set of tools for retail traders. It's not a toy and it don't works with a single magic button. If you search this then this system will not be for you. And now I small explanation about stages involved.

You think you are working with some PIN tools. You have an idea about how mix that tools. What tools will work as entries and what will work as exits. Evidently ideas don't work out of box then you want test that idea.

Now test an idea is a high consuming time taks. You have test in demo account along the dsy/week for see how works. And normally report its limited about economic area. Check in chart trades don't helps because when you use some tools for make an strategy you can't see how worked together in every moment. Then finally you can't tune fine the system because you don't have enough data for check the problematic areas and fix it. And here is where my new system solved all this problems

Will continue in next post...
Try don't lose pants never...
  • #130
  • Jul 22, 2017 3:51am Jul 22, 2017 3:51am
Now we will start to work with new system. As I said in last post we have a good idea but not the tools for develop it well. But now we will use the new flow.

First is make the mix. We decided use Builder PIN for build the PIN tools strategies. We will use separate entries and exits. When we finish the setup we will send every PIN ( entry and exit) to Trader PIN.

Trader PIN is the EA for create trading sessions. A trading session is the common tests traders make in real time. A session have a range date like backtester. And how MT5 backtester allow you check that range in fast wsy. The difference is data involved in that process. MT5 backtester only shows economic info plus trades taking. But you continue without can know how mix worked.

Trader PIN and Builder PIN register all events involved in strategy. After run the test on the fly ( normally a few seconds in long date ranges) will be able to check all events in chart. Trades with all info ( DD, max profit...), histogram with signals of every PIN and group PIN and their info ( signal type, time inside candle...) and session info ( DD, individual info of every pair, total...).

Now we are ready for improve our idea.

Will continue in next post...
Try don't lose pants never...
  • #131
  • Jul 22, 2017 4:07am Jul 22, 2017 4:07am
We have the first report of our strategy. Really wasn't be like expected. It's not horrible but holy grail is so far. That report looks more a big hole in your pocket. Ok. We are disappointed but we can analyze what failed. Then not all is lost.

First thing is see global report about session. We see total is green but we suffered big DD. Checking pairs report we see only worked well pairs in trendy market. This history is not new in auto systems. It's evident our strategy don't works out of box in all pairs. It's a problem. Or maybe not at all.

We decided fix the strategy for steps. First doing simple steps for check if this solve the problem enough for don't need advanced fixes.

One thing in trades analysis is know the optimal entries. This is easy because we have included a tool for detect the most optimal swings. With that reference you can analyze better how good or bad are trades. And can see signals made every trade is a valuable info for understand what tools are failing.

After a few minutes analyzing some bad trades we see one PIN tool is not working well with rest of tools. We decide try modify parameters of that PIN. We can apply changes to all pairs or only problematic pairs. And this modification can be done manually ( you decide parameters values) or you an optimization approach like MT5 backtester. We prepare the new parameters and make a snapshot of that new parameters. Everything in a few seconds.

Now we have new snapshot only applied to that bad pairs. We used optimization then snapshot will show the best parameters for every pair inside range we selected. This looks better. Now some pairs avoid red results and global basket shows better profits. It's enough for try jump to next step.

Will continue in next post...
Try don't lose pants never...
  • #132
  • Jul 22, 2017 4:21am Jul 22, 2017 4:21am
Now we have a candidate setup. This setup use different parameters for some pairs. Is not a problem because tools can manage different setups at same time. Now it's time to try this in MT5 backtester.

We will try the closer test to real. We put this strategy with real ticks and we will use our farm for make backtest fast. When this backtest finished we saw results are similar to preliminar tests made with Trader PIN. Evidently are slightly different because backtest in my system is a fast approach for take a reference without spend much time.

Now we decide its time to try first long run using demo account. All tools can use previous setup in real time then in a few hours or days we will see how worked with real market. After this long test we saw some small issues in results. This is bad news. But not all it's lost

All demo session was registered then we can analyze that real time data like backtest sessions used for create the strategy. And here is the most big difference with normal tools. What will happens if we use same range date used in demo account with virtual session? We will try...

We create new session with same date that demo account on long tests. We load setup used and we wait report generated. Now we have 2 sessions ( demo account and virtual using same setup). We can compare both. It's a great idea

After some time checking difference we see problem it's inside candle. Some tools make it fake signals. This happens in MT5 backtester but wasn't affected much because date chosen was more friendly with setup. It's evident we need make some adjustments

Will continue in next post...
Try don't lose pants never...
  • #133
  • Jul 22, 2017 4:28am Jul 22, 2017 4:28am
This time the fix is more complex. Before use deep modifications we will try a easy way. We try tools work only in open candle. This will avoid the inside candle effect but will delay some trades in fast and big movements. We can use a TF optimization for check the best TF for this behavior

After optimization with Trader PIN we got the best TF for use open candlelight mode with setup is M30. We can compare this new session with previous sessions for see entries difference but we liked te economical report then we will try demo account directly.

After make this long run again with new behavior we saw results are quite well. Then its time to try real account with the security we have a solid strategy. Evidently all real account sessions will be recorded for can recover it later and analyze areas we want.

And that is basically how it will work my new system. We can call it a holy grail hunter...
Try don't lose pants never...
  • #134
  • Jul 23, 2017 10:23am Jul 23, 2017 10:23am
This start to take color. In capture I'm using virtual backtest.

Trades were simulated opened at 6 bars ago and porfit is using current candle
Attached Image (click to enlarge)
Click to Enlarge Name: Captura de pantalla (252).png Size: 93 KB

Or using current candle
Attached Image (click to enlarge)
Click to Enlarge Name: Captura de pantalla (254).png Size: 45 KB

Or previous candle
Attached Image (click to enlarge)
Click to Enlarge Name: Captura de pantalla (255).png Size: 45 KB
Try don't lose pants never...
  • #135
  • Jul 23, 2017 11:13am Jul 23, 2017 11:13am
This is Real Art of Beauty Signal and Colors
BtB you surprise us all the time with your Creative MIND !
The arriving better than the arrival, the journey has just started
  • #136
  • Edited 9:44am Jul 26, 2017 8:52am | Edited 9:44am
Testing EA trades manager...
Attached Image (click to enlarge)
Click to Enlarge Name: Captura de pantalla (271).png Size: 142 KB
Try don't lose pants never...
  • #137
  • Jul 27, 2017 7:27am Jul 27, 2017 7:27am
Finally metatrader have API for custom charts The possibilities are infinite
Attached Image (click to enlarge)
Click to Enlarge Name: Captura de pantalla (272).png Size: 123 KB
Try don't lose pants never...
  • #138
  • Jul 30, 2017 12:40pm Jul 30, 2017 12:40pm
Metrics of Trader PIN session file:

1) Session length is 5 days ( one trading week )
2) Pair bars are 7200 in one full session
3) Pair full session size is 900Kb. This contains all info needed for restore trading session ( signals, trades, ... )
4) A typical 28 pairs full session is around 25Mb

And this only in EA. Builder PIN session can be giant if you use all PINs. Taking EA metrics something like 625Mb ( 25 PINs x 25 Mbs a 28 pairs week )

I dont know how can be shared that over net...
Attached Image
Try don't lose pants never...
  • #139
  • Aug 2, 2017 8:19pm Aug 2, 2017 8:19pm
Quoting braintheboss
This start to take color. In capture I'm using virtual backtest. Trades were simulated opened at 6 bars ago and porfit is using current candle {image} Or using current candle {image} Or previous candle {image}
Looking good Brain I can't wait to try it someday- probably 2 years from now.
  • #140
  • Aug 3, 2017 2:04am Aug 3, 2017 2:04am
Quoting PipFlak
{quote} Looking good Brain I can't wait to try it someday- probably 2 years from now.
I think when Brexit is done. Many tools to make for can have minimal usable system
Try don't lose pants never...
0 traders viewing now