Votes for future features

Here is my vote for the priority of the next Zorro features:
1. Interactive Brokers API plugin
2. Historical price tick data for backtests
2. FIX API plugin
3. Plot functions for MT4 chart windows

I'd like to second the vote for the Interactive Brokers API plugin for the next Zorro, please!

Oanda fxtrade rest api plugin

1. FIX API plugin
2. Interactive Brokers API plugin

need to get off the FXCM broker

Hi jcl. Further to the posts above and your reply to JobTick in the Starting... section:

The Broker Interface page in the manual for a long time has had this:



This statement was written a (relatively, given how long Zorro has existed) long time ago, 2015 (so far) has brought changes to FXCM, other brokers have no doubt changed as well, etc. Given all that and the obvious interest here in other brokers, a couple of questions:

1) Does the Zorro team still feel the same way about FXCM vs. other brokers?

2) Given the flexibility currently provided by the MT4 plugin to go to pretty much any broker, does the Zorro team have "good feelings" about any other broker(s)? (Actually, I suppose there's a U.S. vs. non-U.S. aspect to this also...)

Thanks.

Both Interactive Brokers and Oanda support FIX API, so with this feature I guess most requirements concerning "access to another broker but FXCM and no MT4" should be satisfied. And there are still many other brokers to which you can connect via the FIX standard, so this would surely add a lot of flexibility to Zorro.

I vote for the IB API as number 1 priority also.

Ok, I understand the arguments. We'll make other brokers a higher priority and will order some broker plugins for the next updates, like 1.36 or 1.38.

Thanks jcl!

Based on Sphin's comments the FIX API seems the way to go to add a number of other brokers with 1 effort? Assuming each broker's "support" of FIX API is what it should be...

Maybe you can even get some of these brokers to throw in "new client" bonuses for us?!

Thanks.

Hi

my future request goes to Interactive Brokers API, due to its broad number of instruments including stocks, futures, cfds,...my interest is in stocks, as they're not included in small forex brokers,

Cheers

First I want to thank you repeatedly for bringing Binary Options to Zorro!

I would appreciate if I could re-enable Spread for Binaries, because

a) many Brokers "protect" (=cheat) themselves against too successful traders with a spread
b) there are hybrid binary options (I've seen theese at dukascopy) that have spread (and a higher payout in return).

With longer term expiry times, there shouldn't be much difference with the performance, but with 60s Options, a 1Pip spread will probably make a significant difference. It would be interesting to see, how much room for broker cheating your strategy has until you fall below break even.

Ok, I'll put re-activating Spread for binary options on my list.

I have been looking forward to implement some API, and I have been recently reading the documentation of the Interactive Broker API and apparently their 'API' is an API to their trading client, the TWS (trading workstation). That is you have to execute TWS and then you can automate it with their API. Alternatively, you can run the IB gateway but still it is not the same kind of API as the API from FXCM.

jcl, what are your thoughts on this? For reference:

https://www.interactivebrokers.com/en/software/api/apiguide/c/ctutorial_connecttotws.htm
https://www.interactivebrokers.com/download/newMark/PDFs/gateway.pdf

PS: On the other hand Oanda -was the one as dukascopy that didn't charge traders for negative equity after eurchf peg removal wasn't?- has a rest API that for what I read, would allow to make a true API to use with Zorro. I will delve into it meanwhile.

PS2: I found this documentation from ninja trader that shows that their development team implemented the interactive brokers 'api' using the TWS, that is, a third party...Is this acceptable for Zorro?

I think Economic News Awareness could be a powerfull addition. It could even add some fundamental Trading to the 100% technical approach as it is right now, or at least avoid some losses (don't open new Trades just right before NFP report or potential Rate Changes).

One great thing to insert in Zorro, could be Option trading (not binary but plain vanilla)!!

My priority vote:

1)OpenMAMA bridge for market data. This is a lightweight vendor-neutral integration layer for systems built on top of a variety of message orientated middlewares. In this way Zorro could be part of a modular modern platform and expand the visibility of Zorro.

2)FIX API for order routing and trades.

3)Drawdown alert by comparison with the backtest equity curve through the Cold Blood Index. It is very important to get a prompt feeling of when our strategies are not performing fine enought. This is probably the trader's most difficult task: when to retreat from battle.

4)Dukascopy api bridge. This broker has a free good price history, fair commisions, high speed execution and java API which could expand Zorro capabilities with the huge java universe.

Thank you. Really great Job!

I go for parrondo suggestions.
I would add
5- Verifying script correctness by comparing live results with a backtest of the same period. It is a quick way to know that the strategy is well developed.

+1 for the FIX API!

I vote for a chance to save the trading statistics over a restart of a strategy, so that its results can be continued.

Good point Sphin

Thanks for the suggestions! All are noted.

I'd like to suggest a script wizard. I mentioned this under the Z Systems forum header earlier.

I suggest this for the following reasons:

1 - It's very important for the future of Zorro as a whole to attract new users.

2 - The majority of them will not be programmers, but will have many good ideas.

Therefore 2 feeds back into 1, with one proviso; new users of mixed programming ability can get started quickly with their ideas in Zorro.

Further, I'd like to suggest that you might consider asking Radoslav, the creator of fxDreema for MetaTrader, if he'd like to develop or migrate a tool? He's based in Bulgaria. My experience of this product has been very positive.

New users are your life-blood, and making access to Zorro easy with a wizard, amongst others, is a vital ingredient.

Thanks for an excellent product, which I hope continues to grow and improve.

I'm a bit sceptical about script wizards, but I've contacted Radoslav. If he wants to do a Zorro version of fxDreema, we'll certainly support that.

I'd like to suggest a "Zorro Protocol" broker plugin. What I envision is:

- Zorro will support a basic TCP protocol and act as a client.

- Potential broker plugin developers who do not wish to write C++ plugins (there are a lot of developers who don't write C++ on Windows) can instead write Zorro bridge servers to brokers. Brokers can potentially deploy these servers or implement Zorro's protocol themselves if they wish.

- Zorro gets wider broker support and developer attention.

- Zorro can be listed as a partner gaining a little publicity.

I came up with this after trying to implement the Spotware Connect API. While it's relatively easy to communicate with the Spotware API (I have near zero C++ experience but managed to hack together a client for their TCP API with Boost::ASIO in a weekend), doing so while subjected to Zorro's constraints is extremely difficult.

It'd make my life much easier if I could just write a small server in Go/Python/Java/Javascript/C#/F#/non-Windows C++ and point Zorro at it.

On a related note, I'd also like to suggest a Spotware Connect broker plugin. Their APIs (there's a HTTP/JSON API for account data and historical ticks and a TCP API for live prices and trading) are well designed and easy to work with and implementing the one plugin gives you access to every broker that supports cTrader (there appear to be ~15 of them). I can share the code I hacked together to work with it if it's handy.

My priority vote:

1. FIX API - because there's no other way to go for many professional traders.
2. Dukascopy connectivity - because of their tick data history. I tried many different data sources and Dukascopy is one of the better ones.
3. Real (variable) spread simulation - for me it's very important as we make thousands of trades per week, each with very small edge. Thus we need to test on our broker's data and with absolute precision. Sometimes the trades tend to cluster during certain hours when the spread is bigger and this alone can make all the difference between profitable and unprofitable system.

But maybe the variable spread simulation is possible even now with your own data source? I imagine you could create one asset from ask prices, one artificial asset from bid prices and subtract them during backtest to get the real spread. Is it the way to go?

+1 for FIX API

pcz, you can hack a variable spread by doing something like:

Hey boatman!

Thank you for the tip. To have the maximum possible precision I think it's better to have spread for each tick. This is my current solution: Backtesting with T1 data and variable spread

(so I take back the 3rd point from my priority vote as it is apperently already possible with currently available tools )

Cheers!

Support and resistance indicator. It can be done, I'm using a commercial S&R indicator for Metatrader when I'm trading manually, and it totally rocks: http://www.pointzero-trading.com/Products/view/PZSupportResistance
The asirikuy guy spent some time on it and developed some ideas in 2011 as well:
http://mechanicalforex.com/2011/06/intro...sistance-o.html

I think this would be an extremely valuable extension in zorro, because S&R / price action is what the big market movers like banks and funds trade. It's extremely reliable compared to technical indicators.

I do not know a reliable S&R algorithm, and am not even 100% sure that such a thing as S&R exists at all. But there are lots of different ways to define S&R indicators. So if you can describe a method that works, just let me know - I'm certainly interested. The blog post by Daniel in 2011 did not lead to a profitable system.

Hi!
Could you provide an ability to change plot's color depending on its value (like Tradestation does)?
It would be very helpful, e.g. for Swamicharts drawing.

As far as I know a logout and trade again of a strategy means the beginning of the statistical results of the performance report. I think someone posted already that it the performance report of htm file remains the same. Or, is there a way to keep the old htm performance report after stop and trade a strategy?

I would like to vote for a color coding option on the main window, too! That means to allow different background colors dependent on a logical value or something like this.

A current problem for me is to draw dots and triangles on the main window when other windows (using the 'NEW' option of the plot command) are opened. The dots and triangles are always drawn on the last window, not the main window. Any suggestions how to solve that?
regards,
Opentrades

I would like to vote for support of MetaTrader 5. For some reason, all brokers in my country works only with MetaTrader 5.

Where are you from? MT5 support is already on the todo for a future version -> http://zorro-project.com/manual/en/new.htm.

Hi,

I am from Brazil.
I had already seen the todo list for a future version. Just thought that my vote could speed up things a little bit. ;-)
I´ve seen somewhere that MT5 was planned for the version after 1.46, but it was not included in the current beta version.

Looking forward for that future release.

Thanks,

Yes, it's on the To Do list, but won't make it into the next update. It's now scheduled for 1.52. Depending on contract work the priority of features can change from time to time.

I remember up to 1.47 one could read on 'What's new' under 'Planned for future Zorro versions':

latest in addiotn with 'in development' before it suddenly disappeared. Why? IMO this would be a very interesting and important feature.

An implemented feature is naturally not planned for a future Zorro version anymore.

Ok, that this feature links to a 'Retest' possibility was not obvious. Very well hidden!

I have made an MT4 to etoro bridge to use with etoro (www.etoro.com). I never got around to use it, but I still have the code.
Would there be any interest in having an etoro API for Zorro?
It is not a "mouse clicker" it uses JSON calls to access the etoro platform.

As a new feature I would like to see a WFO for factors too.
At the moment there is only for parameters.

Thanks
Vincenzo

Would it be possible to increase the precision of price representation in trade spreadsheets? It would be much easier to find potential bugs. I've checked the new beta and it's still only 4 decimal places for lets say EUR/USD.

Ok, 5 digits should be easy to implement.

Hey Veratyr,
Did you read this?
http://www.spotware.com/about/news-updat...der-account/789

I read that and got excited, then posted the question here:
http://www.opserver.de/ubb7/ubbthreads.php?ubb=showflat&Main=55832&Number=463538#Post463538

I'd be really interested to see what you have coded up.

JCL:

I would strongly suggest adding the ability to live populate a grid/excel type sheet with historical and live data for portfolio testing. TS has this ability in RadarScreen and MultiCharts does as well.

As soon as you have a portfolio of symbols trading it becomes very important to know certain info in real time. All are easy to program but are not easy to view in real time without real time grid output of some sort. Outputting to CSV is ok to some degree but with that method you cannot view live data and trading action to debug/improve the strategy which is designed to run over multiple symbols at the same time.

This also helps with more complicated strategies (multiple methods of entry/exit/stop) on a single symbol to see live which entry/exit/stop is being used and other stats about each entry/exit/stop in real time.

The following is only an example of a few datapoints that become very important in development and actual live trading:

1. Position status of each symbol, entry price, target, stop
2. Current profit/loss in real time for each symbol for open trades
3. Profit/Loss for the day for each symbol
4. Number of trades each symbol has done in last x hours, x days, x weeks
5. Percentage win for each symbol for the last x trades
6. Sharpe ratio for each symbol
7, 8, 9 ... Other risk metrics etc.

Did you read the section about Strategy Control Panels? Isn't this what you are looking for?

I saw that - that appears to be a static panel derived from a .csv file with some additional control aspects. It appears to be only for static data (unless I am mistaken).

I am talking about the ability to push live data (last price, entry price, exit price, trail stop etc.) to a panel so one can easily view the data in real time.

Additionally populating historical data for each symbol/system which is live trading and/or currently turned off based upon rules.

As soon as strategies become more complex and run on multiple symbols or a large portfolio of symbols, creating switching rules, debugging and specific improvements become more difficult without seeing live data as the strategy calculates in real time. Odd things can occur which are undetectable in backtesting/optimizing against a portfolio but are clearly evident when watching in real time.

Perhaps adding the ability in Zorro to use a DDE link/server to export into excel would probably do the trick.

From the manual:

With this command you can update the content of any cell of the panel with any live data at any time you like while the command 'panel' initiates a panel optional with loading its initial structure from a .csv.
I guess the question is more how useful is watching the 'live parameters' with your eyes instead of let them watch by a machine. A quick & dirty solution (I use them often, indeed) might be to save the parameters of your interest to a text file by Zorro (each minute or each intervall you like), set a DSN to it and use the DSN in Excel as the data source.

Thanks - will give it a shot.

Curious, how does zorro connect to brokers at the moment, if not through the MT4 bridge? I thought zorro connected already via fix api?

Hi Z community,

I vote for Artificial assets as very high priority.

Meanwhile, it would be nice if any guru can give a script example, (based on current Zoro version) of e.g USDX (dollar index) to be calculated on the fly

( example of weighted formula from MQ4:

ExtMapBuffer1[i]=50.14348112*MathPow(iClose("EURUSD",0,i),-0.576)*MathPow(iClose("USDJPY",0,i),0.136)*MathPow(iClose("GBPUSD",0,i),-0.119)*MathPow(iClose("USDCAD",0,i),0.091)*MathPow(iClose("USDSEK",0,i),0.042)*MathPow(iClose("USDCHF",0,i),0.036);

)

and use some series transformation functions or an example of a peak/ valey or crossover example.

As this would be an artificial value on some brokers, there will be no bars or price series, but can be constructed on the fly.

The time stamp in message log should include the seconds of each bar creation, which currently shows only the hour and the minute.. Me thinks .

Better resume trading support. As far as I can see, Zorro still doesn't exactly pick up trading where it stopped. When I restart and I click result, the chart only shows 1 day of history and the portfolio.txt doesn't show the actual portfolio, even when the strategy ran for weeks. I don't know if this is fixed already, but in former Zorro versions restarting Zorro also messed with the Z12 portfolio. Just after restarting, like 20% - 30% of all open trades were closed almost immediately.

And I want to report a bug (didn't find a better place for this): I'm using Windows 10 on a Core2Duo Notebook (slightly above 2GHz). When I hit start in Zorro (I only tested this with Z12), Desktop Window Manager jumps from 8% CPU to 55% - 58%, causing the fan to run at all times.

Support for economic indicators / news releases (For example Forexfactory calendar).

Relying 100% on technicals is like you are driving a car by only looking in the rear mirror!

http://topshelftraders.com/articles.php?id=335

Hi,
Since the release of 1.54.1, you can trade index futures via the IB Bridge, too.
One important feature the IB Bridge is missing though, is the ability to get live traded volume data send when doing live trading via IB.
For example, this prevents the use of filters which depend on the total bar volume to either validate entry signals or to be used as early exit parameters if the volume falls under a certain threshold.
On intra-day strategies with a short BarPeriod such filters make a lot of sense on momentum or trend strategies to limit the amount of “false signals”….

So pls vote for this enhancement of the IB Bridge in the next release.

Thanks

Fully agree! That would indeed be great to have.

As a cTrader User, I would love to see native cTrader (a.k.a. FIX) support.
Spowtware seems to have started something about a year ago: https://github.com/spotware/connect-zorro-plugin

I vote for this feature too. It is one that I would consider critical for futures trading.
Thanks!

Live volume was implemented some time ago, and historical volume theoretically also, but the latter does not work yet for all assets.

A dynamic plotting function. In MT4, I can zoom and move graphs in backtests, I can see Entries, Exits, StopLosses and most importantly: Plot the indicators that generate a trading signal next to the simulated trades. This would help me while debugging.

+1 for dynamic plotting.

+1 for dynamic plotting - also from me

Nobody up to Python Bridge? As more I learn about machine learning as more I encounter python more suited for advance models

Would be surprised if there is something you can do in Python that you cannot do in R?

Thats probably true.... I just see more examples around of python machine learning libraries than with the R ones.

+1 dynamic plotting too

implementation of the Leverage-Space-Model

+1 dynamic plotting too

Looks like a lot interest for dynamic plotting.

Since we have MAE already, a function for Maximum Favorable Excursion would be nice :-)

+1 dynamic plotting too

It would be great to use the functionality of STEPWISE for "visual debugging" for both trades and indicators.
The user could define for such "Market Replay" both:
1. step ("bar by bar" or "trade by trade")
2. speed of replay.

There are not many platforms with the Playback functionality and each of them address it in a slightly different way. So it may be worth checking first:
1. Multicharts: https://www.multicharts.com/features/data-playback/
2. NinjaTrader: https://ninjatrader.com/support/helpGuides/nt8/en-us/?playback.htm
3. Sierra Charts: https://www.sierrachart.com/index.php?page=doc/ReplayChart.html

I would like Zorro to safe the Strategy Script directory. I don't use the default directory and every time I start Zorro, I have to change the directory.

+1 on Safe Script Directory

A Python3 bridge would be awesome as Python is my weapon of choice now. And I am absolutely sure I am not alone with this. Please support my vote if you agree!

I agree with it. You maybe can do everything as well with R, but you may need to reinvent the wheel for that

Agree - I've always loved R and the R bridge was one of the reasons I bought Zorro but I switched to Python as it's better suited for many things I need.

Pythonic is easier and there is not much if anything Python cannot do but R can do. Even academia uses Python more now than R. Just check out google search term trends......

While I am a big fan of R, I find myself also doing more in Python and less in R as time goes by.

capability to Zorro for downloading data from https://stooq.com/

Another data source would be nice, so definitely +1.

stooq is awesome

I've just looked into it and Stooq makes indeed a good impression.

I hope you will write a script to download stooq data or include it in zorro later.

Sorry for OT, but when I go to stooq, everything is in polish. Can I change the language?

yes, you can change the language into english, juste look on your screen. try this page https://stooq.com/q/d/?s=eurusd

and Ctrl+F english

One more thing I'd really like is to be able to produce more than one plot at a time. I often find myself running the same backtest multiple times just to produce different plots, like MAE, seasonal plots, price profiles etc which is really time consuming.

It would be handy to be able to expand the main window
In fact, when the name script's is too long, it is cut in the menulist "script" and it is impossible to read all the name.
It is a problem if you're usually renamed different versions of the script with the suffix V1,V2 , etc...
For example I have "Trend Trading with ADX_V1.c" and "Trend Trading with ADX_V2.c" and the two script are indistinguishable

Thanks

You'd like to run backtest and at the same time have multiple plots ?

Yes, and by that I mean many main plots, not one main and several smaller underneath.
As an example, if I plot MAE and price profile during the same run, they are both plotted on top of each other.

+1

Hi,
in the test result page, I'd like to have a statistic showing the correlation between the the different equity lines produced by the Strategies/Algo.

Thank you

Hi,

I am rather new to Zorro, and i am really impressed with the power and speed of the platform. But i miss some sort of graphical interface, something similar to Amibroker where i can easily plot a chart for any asset, zoom it, see my trade signals, etc.

A lot of times is easier to find out a strategy when you can view it directly on a chart and play with it.

¿Any plans for achieving a more integrated user interface?

Thanks.

Replacing the static chart with an interactive chart was already voted for by many users. It's on our list, although not for the immediate future.

ok, i understand.

+1 for expanding the main window, not least because the window is too narrow to allow me to read any longish error message.

How about the addition of a Compile button. Or an option to auto-compile on save. Here's why... I have had a few WTF moments when experimenting with sliders. I code my slider, save the script, go to Zorro's launcher, and the sliders have the default names. It took me a while to figure out that I need to compile the script in order for the slider names to change, and then it took me a while longer to figure out how to compile the script. So now I click on Test, then Stop, then I set the sliders as I need them, and finally I click on Test again and wonder how the process could be made easier.

Another gripe I have is that when I click on Results, the script editor window gets unmaximised every time.

All that being said, I'm still way more productive with Zorro than with MetaTrader.

Thanks!

- Python bridge

- +2 for expanding window

- Zorro for Ubuntu Linux

- Plot the plots included on profile.c at the end of a backtest for each asset-algo combination

zorro in ubuntu, it's like a dream

Please. I think it would be a good idea to allow people to read T1 data using the free Zorro version. Thanks.

For those impatiently waiting for a dynamic charts capability, I should note that there is an R package designed exactly to do that: dygraphs for R .

You can zoom in and out. It has OHLC candlesticks. Using the Synchronization feature, you can plot an indicator/filter directly beneath. The list goes on.

If possible I'd want to calculate Optimal F factors for each period when using WFO, rather than from the whole backtest

+1 for:
Interactive chart window
Notepad++ integration with auto-compilation and help lookup.

NumSampleWFO option.

Currently there is a NumSampleCycles which reshapes the price curve and allows to see how the strategy perform in different conditions ( then if the PF is similar between the cycles, then it still has an edge ).

The strategies are built using a WFO fix window. It means that eventhough this is OOS, in reallty the oos periods are completly biased. If the strategy could be tested over different wfo cyles, then we would really know how it really behaves. Then for exmaple, something like NumMaxWfo=10; and NumCyclesWfo = 5; would do 5 backtest of the strategy with wfo as 10,9,8,7,6 each.

I believe you can do that already. Use NumTotalCycles, and change NumWFOCycles at any cycle. I haven't used this for WFO cycles yet, but in theory it should work.

You get error032. Not sure how to go around it.

+1 for python bridge because I am much more familiar with pandas, sklearn, etc. than I am with the equivalent methods in R.

I would also like advise(NEURAL) to accept way more than 20 signal values. I know that I risk overfitting when using huge numbers of features, but I do need to export the features in order to run feature selection algorithms on them.

+1 also for Notepad++ integration with auto-compilation.

Error 032 means that the number of WFO or sample cycles changed not from cycle to cycle, but inside a cycle. The number must remain constant after the INITRUN. If you can't fix it, post the code in another thread on the forum and I'll check it.

Minor: This setting reads like it can be automated. If two Zorro instances run strategies for the same API, then it could simply auto enable it, otherwise turn it off. Thus the user would not have to worry.

http://zorro-trader.com/manual/en/mode.htm

+1 for Notepad++ integration with auto-compilation and help lookup.

I think it would be nice with a mode flag that would enable brute force optimization, which is useful when data mining. I know it's possible to do this now as well, but it´s VERY easy to make a mistake when you have many parameters to optimize. Already with three parameters you get something like this

+1

It is good to get more samples for white really check

I vote +1 for interactive display charts with zoom option!

The strategies we write are only as good as our understanding of market behavior. Ease of investigation is the foremost driver of success!

+++

+1 for OptimalF factors for each WFO period...And make the function itself customizable by the user!

+1 for "Simultaneous connection to two brokers, or to a broker and a price source (Zorro S)"

+++

We should be able to retrieve and apply OptimalF factors on the go with each bar, just as you do with any other indicator. Now you only get the factor of the final bar, you can´t do any kind of OptimalF backtesting.

One suggestion:
Maybe the function verify could just be var probability = verify();. Because the dbl file should be the default name of the strategy and the depth and long of the drawndown can be calculate from Equity

Suggestion for Multi time frame.

that will enable experts to be programmed with any combination of data feed from two different time frames, or three or even more, and enable multi currency expert trading.

And enable compiling of code for stand alone application which will give even higher performance to work with, like an IDE compiler.

I suppose that's one of the rare examples of wishes fulfilled before they even came up.

+1 for python bridge

+1 artificial assets from a linear combination of real asset for basket trading
+1 interactive chart window

+1 for python bridge

I'm thinking this program is something that Zorro would qualify for, given the integration to IB for example.

CME Non-Professional Globex Data Package
http://www.iqfeed.net/index.cfm?displayaction=data&section=feewaiver

Would be great to be able to get a feed from IQFeed at a discount :-)

If myfunc() is called 100 times per tick from multiple different places, then this error message is totally useless to me. Besides according to the manual various different errors could cause an Error 111.
Can Zorro not be a little more specific as to the cause of the error?

Could Zorro not provide a call trace? Just knowing which line of code called the function that crashed would be nice. Knowing the function call hierarchy up to the most recent run() or tick() call would be even better.

Check this page: http://zorro-trader.com/manual/en/trouble.htm
Specifically check the pos function, which seems to address your concern. (Also note that pos function is new in the v1.70, currently in beta)

Imo if you want better code and debugging then you should consider DLL strategies. See http://www.zorro-trader.com/manual/en/dlls.htm

Zorro comes bundled with the DLL API. It's easier to use than Lite-C.

For a better c++ API you can clone https://github.com/xezon/zorro-dll (was made for ZorroBeta_1.70.7 as proof of concept).

Feedback on above implementation would be much appreciated.

Hi,
it would be nice to have the same title of the Zorro's window displayed in the Trading status bar. At this time the version and the name of the account is missing.

Thank You

Unfortunately, I have no experience coding in C or C++. I am much more used to the ease of more modern programming languages that tend to have much more informative stack traces.

1. Nadex API Plugin
2. Cantor Exchange API Plugin

Here goes my vote for next releases:

1.- Python bridge
2.- Advise() method for training and executing a TensorFlow graph.

i know it has nothing to do with Zorro itself, but can you tell me why the manual never opens with the "Site map" if you folow a specific link? Thats very annoying

better GUI (dropdown, dynamic content, ...)

News indicator (historical for backtest + live) from any of the websites providing all kind of FX news.

generic optimization

rework profile.c - poor performance

allow negative values for optimization e.g. range from -10 to 10, step: 2 - at least this gives me an error atm

non-linear options for Data slope

Hi easyX. For

allow negative values for optimization e.g. range from -10 to 10, step: 2 - at least this gives me an error atm

use 0 to 20 in the optimize and subtract 10 outside.

HTH.

I vote for "Interactive chart window."

Thank you

I find the use of a panel very helpful. Once you start using a panel, the desire arises to be able to create more than one. So I vote for enhanced panel functionality enabling the use of more than a single panel.

Artificial assets from a linear combination of real asset for basket trading has been on the todo list for a long time. For what version is its implementation planned?

+1 for Python bridge
+1 support for configurable neural nets via TensorFlow
+1 interactive dynamic chart window

Add Name as one more parameter to the function Optimize, so it would be possible to create your own parameter name for the WFO Zorro Report.

The report would be more intelligible and easier to understand at the first sight.

MT4 to Oanda API bridge so we can trade EA's. I can't find a good solution for this that doesn't cost $1k. HELP!

1) Interactive chart window.

2) Python bridge.

3) Artificial assets from a linear combination of real assets, for basket trading or multi-asset arbitrage.

Any chance of seeing a numerical erf(x) function in Zorro some time soon?
https://en.wikipedia.org/wiki/Error_function

var erf(var x) { return 2*cdf(x*1.414)-1.; }

Perfect! Overlooked that one, thank you Petra.

+1 for Python bridge

+1 Python bridge

Python bridge is already in the scope
http://manual.zorro-project.com/new.htm

+ 1 for linux

that would allow the use of gpu deeplearning ami from amazon. https://aws.amazon.com/marketplace/pp/B06VSPXKDX

High on my wish list is a zorro connection for the DeGiro.

DeGiro has for the european stock and ETF market better conditions than IB.
The main issue with trading the US market is the exchange rate risk which is currently indeed, at least temporarily, chewing up any income you generate with Z9 or other conservative strategies one may run. Therefore the addition of a DeGiro API would be excellent to the tap european market and benefit from the lower costs. THis seems feasible:
https://github.com/pladaria/degiro

I know python bridge is already in the plan.

However i would add here my +1 vote for it.

Python bridge pretty please!!

is it somehow possible to get an offset by seconds? So that e.g 1 min bars are shifted just by 1-2 seconds?

+1 python bridge!

+1 for "Artificial assets from a linear combination of real assets, for basket trading or multi-asset arbitrage." - very important for option strategies in particular

Taking into account the TimeZone for dataParse would be great. So that it would be easy to read CET based history data (including DST).

Also consindering DST before the year 2000 would be nice (1996 there was a change in germany).

An automatic time adjustment of any data history based on NFP news (always at 8:30 EST) would be the very best option. Data with inconsistency could be fixed.

+1 python bridge!

Use NumCores for testing too

I'd like a new indicator: Super bandpass filter by John Ehlers. Thx

Hi

VWAP and TWAP as indicators with a possibility of fixing the anchor (monthly, quarterly, daily, annualy) would be a great addition to the indicator list.

Hi,

Interactive chart windows is already long on our Christmas list for Zorro.
I would like to draw the op group to the possibility of using the free charting library of tradingview which has an API to use ones "own" (i.e. zorro's) data.
https://de.tradingview.com/HTML5-stock-forex-bitcoin-charting-library/
Integration with zorro would greatly enhance our trading system
developement and mining experience.

Please add multi core for standard optimizing (non-WFO).

A function returning the number of trading days between 2 DATE's would be nice. Or is there an easy way to do this i missed?

- Some sort of interactive chart window
- Being able to download previous versions of Zorro
- Specify a different location for History folder in Setup, add a new HistoryFolder variable.

+1 Degiro API

Hi Team:
Thanks for this wonderful application and your efforts.

1) Could we have an overlay application for generating the basic script template? The user would simply click on a number of checkboxes that fit their intended script and then the overlay app would generate the basic script in Notepad.

Hi Team:

Can we expand the genetic programming capability by expanding the decision tree feature to incorporate operators such a multiplication, division, etc?

There is a trading program I used in the past that uses the RSI, CCI, Max, Min, Modulus, Comparison (><=), Stochastics, Moving Average, OHLC, and a similiar set of features to auto-generate the decision-tree Trading Rules. It creates a population of artificial traders, each with unique computer-generated trading rules for Long, Short, and Exits.and maintains them in an auction system to model real world market participants. Parameters are automatically selected by the genetic algorithm as it creates the tree.

Zorro's DTREE outputs looks just like it, except it is not automatically utilizing indicator sets and mathematical operators. Or, I am not familiar enough with Zorro to see that we can do the same thing here.

My wish is to expand Zorro's current DTREE feature to also incorporate more mathematical operators and a set of indicators (Fisher, MA, Stochastics, CCI, etc..) to create trading decision tree rules.

Hi Team:

For your consideration:

I believe the Manual is using the term "Signal" instead of the more accurate term "Input" in some spots. In both text and code, we refer to an indicator we just created as a "Signal" instead of just calling it an Indicator. At the same time, the manual also refers to the final trade decision as a "Trade Signal". This should be streamlined.

We use signal in both of these two senses in this statement from Workshop 5,

"For generating trade signals, it is normalized and Fisher transformed, and the resulting Signal series is finally compared with an upper and lower Threshold for generating trade signals."

Can we reserve "Signal" for the context of a Trade Signal.. ie. a Buy ,Sell, Exit, or Hold signal ? An indicator crossing a threshold is an event. It is that crossing event that we would call a signal. Ex, it is commonly stated that 'Neural Networks accept inputs in order to output trade signals'.

Likewise, it is easier to say that we feed inputs to adviseLong() or adviseShort() in order to generate a Buy or Sell "Signal".

Here are some restatements of parts of the Manual:

"To create our final indicator, it is normalized and Fisher transformed, and the resulting Indicator series is finally compared with an upper and lower Threshold for generating trade signals."

"In training mode, Zorro learns rules that predict the success of a trade from the Inputs". "Inputs should be normalized roughly to the 100.. -100 range. "

vars Fisher_BandPass_Indicator = series(FisherN(Filtered,500)); // instead of
//vars Signal =

Native support for Kraken crypto exchange

I would love a Pepperstone API

Currently when switching between multiple brokers (as for instance in the broker arbitrage script) variables relevant for money management such as Balance, Equity, MarginVal etc. are not downloaded for the broker of the selected asset, but only for the initial broker selected in the zorro window. It would be a valauble addition to also updated these variables when switching to a symbol from a different broker using the asset command.

Have Zorro Trader work with Mathematica, Matlab.

Enable Zorro Trader to Compile Stand alone strategy executables with Visual Studio.

Zorro Trader should also have an application like : www.strategyquant.com

Hi,

my suggestions is Zorro with real time chart..

This feature is available in the beta.

Native support for bitMEX crypto futures exchange

Sierra Chart supports BitMEX, so you can use the Sierra Chart plugin (now in beta, testers wanted).

I have 2 ideas.

1. We need a bug report system like github. Because it's inefficient with mail.
2. We need a public market to buy or sell strategy with members. And provide reliable backtrace and walkforward report.

Python bridge +1

Working with Python(or Jupyter respectively) would be an immense dooropener.
Python has trillions of very sophisticated libraries for working with financial data.

+1 to both ideas, especially to bug report system like github.

Ideally it would be awesome for Zorro to be open sourced, maybe not entirely but it definitely would be a deal-breaker and would expand Zorro market share a lot.

+1 to Python bridge too.

I agree with the bug reporting system, but I strongly disagree with the idea of a marketplace to buy or sell strategies. I can see loads of potential issues with such a market place, and I'm sure we get better value for OP groups time being put into developing the Zorro platform rather than administrating a market place.

Someone mentioned adding a cross hair to the chart with details at that point. This is an example from Pro real time that does this using a "cursor details" window.

That was me, Dalla.. Thanks for the example.

Please add a metric for "Max %DD from the preceding equity peak".
%DD of total profit is nice, but it doesn't tell you what kind of DD you could expect in actual live trading.

*** Variant of the Switch Case *** Optimizing Order of Logic

Please add a feature to enclose lines or segments of code in "cases" such that Zorro tries re-ordering the designated lines of code during optimization and then tells me the optimal arrangement.

By manually re-arranging my logic, I can sometimes improve results. It would be nice if I could tag specific areas of code and ask Zorro to optimize the order of the logic for me.

A python bridge would be very powerful. +1

+1 for a suggestion on how to optimally integrate Python scripts

+1 for C# dll support (broker dll, strategy dll)
+1 for telegram api (https://core.telegram.org/)
+1 for ChartDirector (https://www.advsofteng.com/finance_demo.html)
+1 for Upbit api (https://sg.upbit.com/home)

Add Tiingo as a data provider: https://api.tiingo.com/

If technically possible: 64bit support! If easier, drop 32bit support. Nobody is using 32bit OSes anymore.

Benefit: No more memory related error messages! No more restrictions on how many assets to use, how long to choose the simulation period or what timeframe to use.

+1. But it could be a expensive feature. Also +1 to "Stopping and resuming sessions without interrupting trade statistics"

I think two version may be ok.
32bit and 64bit each.
Because 64bit zorro can not load 32bit dll(broker, strategy aleady developed).

+1 for 64 bit support.
Linux support? :-)

A special 64 bit version for backtests is possible, but would then not support lite-C. lite-C is strictly 32 bit. You would need VC++ for scripts. If that is acceptable, I can put it on the list.

I think VC++ may be ok. Thank you very much jcl.

I would definitely use this feature! And I would gladly recompile my plugins in 64-bit.

Hi jcl,

what about a vertical zoom possibility for the Chart Viewer?
For long x axis with major value variations in the y axis, the chart becomes flat when you zoom in.

Or do i missed a command to zoom vertically ?


Edit 20.03. I just had a prblem while plotting from the script with:
if(!Train) plotGraph("A1",1,ZL[1]*1.02,DOT+MAIN,RED);
Without this code all graphs are zoomed in the y axis perfectly. So forget my proposal...

Regarding 64bit:

VC++ means that this will be Zorro-S exclusive (fair - well at least not unfair!), right? Would that break the syntax as well?

Yes, you'll need Zorro S. The syntax is almost the same - details in the manual under "VC++".

- The vertical zoom is automatic so that all details are visible. If the curve gets flat, there's probably another curve with very different values in the same chart - that should better go in a new chart.

Well, that would put a lot of pressure on me to finally put all the pieces I developed over the last 2 years together and send my first strategy live... Zorro is awesome, I just hesitated to buy/rent the Zorro-S license because it has become very expensive over time. Too expensive if it's not already earning money. Just for you to consider, if it was cheaper, you would have turned me into a paying customer two years ago...!

Maybe a Zorro-S educational license without the ability to trade live...? I'd love to pay 10 EUR/month for this.

I can accept there is some limitation number of assets during backtest.
But in live trading, no reason to do so.

Hi,

I like to see Control Charts(CUSUM) in Zorro Trader for monitoring trading systems and indicators : http://momentum.technicalanalysis.org.uk/LaYa97.pdf http://www.verticalsolutions.com/notes/control_charts.html http://www.verticalsolutions.com/notes/broken_systems_and_control_charts.html

Just adding my vote for the 64-bit C++ version

1. Drag and Drop support.
When .c zorro script file dropped to zorro, Script combo box changed to appropriate folder and selected to dropped file.

2. Strategy folder selected from ini file.
When zorro excuted, initial strategy folder is selected from zorro.ini file.
Or like notepad++ workspace docking pane, folder navigation support.

If this were implemented, would we then be able to use 64bit Python with Tensorflow?

64 bit Python yes, Tensorflow depends on when numpy fully supports embedded Python.

Hi Team:

I appreciate the NEURAL and such features in Zorro. My request is to expand the functionality as follows:

- Provide a default Levenburg Marquadt neural network file with deep learning topology (e.g 2 Hidden Layers) and parameters *preconfigured * as a generic file in the Zorro folder. ( Basically a preconfigured Neural Network file or dll like some apps use today. )

- Allow the User to easily load this configured neural network file in the script, then specify inputs as usual, and train it, with Zorro saving the determined weights in a copy of the Net (the trained Net under some provided filena,e) and then easily use it in the future by specifying the trained Net name in a script. So long as it is in the root folder, Zorro will load it ( the trained weights are already in the file.)

- This will give users an easier way to work with a defined Topology without having to deal with the nitty gritty code configuration. Maybe even the ability to train and save multiple networks under different filenames for later retrieval.

- This would eliminate the painpoint of having to install any packages or frameworks, etc like I have to do today.

- In time, maybe even expand the list of available pre-configured Topologies in the root folder (TDNN, etc) so that all we would be doing in our script essentially is selecting which NN type we want and letting the training determine the weights.. all within the existing framework for selecting DTREE or NEURAL today.

The net benefit is ease of utilizing advanced and known Topologies as well as training, saving, and retrieving your Net. For instance, if I ran some solid training sessions and saved my work, I could retain the best 10 nets as 10 files (delete the others) in my root folder so that I can load up anyone I want in the script just by typing the corresponding file name.

With thanks,
OptimusPrime

Confirming my request for 64bit support. I just found out, that this is about TensorFlow support as well. Is there any schedule for this feature known yet?

AFAIK the Numyp guys are working on it for the next update. Has nothing to do with 64 bit support though.

What I don't get is that Zorro 32bit can talk to R 64bit. Why does Python 64bit need Zorro 64bit? Maybe I shouldn't wait and start working on a Zorro -> R -> Python bridge...

It would be nice if Tai-Pan EOD could be used as a data provider.

The DLL for R is always 32-bit, while the DLL for 64-bit Python is also 64-bit. That's why Zorro can use 64-bit R, but not 64-bit Python.

One alternative approach might be for 32-bit Zorro to launch a child 64-bit python process and communicate with it using IPC (such as networking) and a protocol.

It would be a lot of work to implement this, though. It might be better to wait for Zorro 64-bit to be released, but I'm not sure what kind of a timeline we are looking at.

I'd like to cast a vote for a solution in the MT4 bridge, to prevent mix up of comments and magic #'s when trading multiple strategies.

Thank you.

.t6 file system is good, but there is thousands of stock tickers in the markets.
If someone try to save stock history price datas to .t6 files, maybe thousands of files needed in one folder.
This cause extra efforts to manage data.
To resolve this problem, I kindly suggest HDF5 database support to save historical price data.
Using this database format, we can write all stocks historical data to a one file.

And also there is not enough room to save historical datas to a only one folder,
I recommand that zorro can recognize sub-directories of history folder.

I believe HDF5 is not the best choice for time series. A possible better solution would be having historical files in zip archives.

I think zip archives is very good solution.
Please add that feature.
Thank you.

Would it be possible to allow us have a historyDirectory() call when running a strategy, similar to how we set assetList?
I like to keep daily and 1 min asset files in different directories, and currently that requires me to update the ZorroFix.ini and restart Zorro.

Yes, this has in fact already been requested and is on our list.

I have an idea how to speed up backtesting of strategies using deep learning dramatically. Reason for the slow backtesting is the R-Bridge being incredibly slow.

The workaround with the neural_train() function is really elegant, where you circumvent the R bridge and feed the signals and the objective as a *.csv file. Wouldn't it be possible, to use a similar trick for backtesting?

When a machine learning algo is trained, predictions will be calculated for each row of signals anyways. Why not just put these predictions in a cache, where they can be read from Zorro?

By the way, I wrote two indicators in R that were next to unusable because backtesting was so slow. By implementing a cache on the Lite-C side, I was able to get huge performance gains, because I save a lot of requests to the R session! (The indicators keep most of their values for several bars and the other values can easily be updated within the cache).

The problem with predictions in a cache is that your signals can have any value. Even when you store signals in discrete 1% steps, with 10 signals a cache hit had a probability of 1/100^10.

You misunderstood my approach: I wouldn't cache the signals (which is of course impossible because of the virtually unlimited number of combinations), but the predictions. The predicted value only depends on time, because the signals themselves only depend on time, unless you try to optimize parameters on which they depend. Is that even possible? At least for me, this isn't desirable. If it is possible, you could implement a flag DEEPCACHE or something, to switch caching on.


So if I was to implement that cache, I would pass a timestamp - a bar number or wdate() - along with the signals into the R session. So one of the columns of my signals.csv would be that timestamp. I would put that aside before my neural net gets trained. When the training is finished, I would first add a new column with my predictions to the signals table, then add the timestamp column. Then I would export only the timestamp and the predictions column to a file. So I would get a timeseries with all the predictions which I would read out from zorro.

Maybe the term "cache" doesn't describe the approach. It's more some pre-calculated time series.

I would love to have that "range inefficiency Z-strategy", also I would love to have the Z-strategies combined and be able to run them at once in one single instance activating by .ini file e.g. which are relvant for the instance running.

E.g. Z1, Z2, Z12 within one instance, for poor people, to make most efficient use of those badly scaled VPSs the brokers provide.

I would like Zorro to use multiple CPU cores during a non-WFO optimization.

I would like Zorro to use multiple CPU cores during a non-WFO optimization.

Is this feature available in the beta?
I still do not see the real time chart when starting the trade in v2.30 or 2.33 beta.

Hey, everyone! Any plans on being able to run production strategies on Linux?

I am about to embark on the devops side of the execution and I would not like to invest time and effort on running it on WIndows if Linux is on the roadmap.

Best!
Juan

Is there a way starting zorro don't push any button but only select script using command line?

plotBar Display Improvement Suggestions

I appreciate that there are probably lots of higher priority new features in development and that one can export plotBar values to a .csv file via LOGFILE to make nicer plots.

It's mentioned in the "Future Zorro Features" that there will be Interactive Chart improvements coming. With that in mind, I have included below some small, suggested changes to the way the plotBar is displayed. In my view they would make the existing Histogram plots easier to read (and would remove the need export values make them nicer in most cases). Some of these suggestions may already be in the works or be in place already (but if they are, I can't find the documentation about how to use them). The attached plot shows each of the items that are being talked about below.

1 - plotBar Labels - Consistent amount of decimal places
The Plots being produced by plotBar currently use different decimal places to display the bar label. e.g., No decimal places if it's a whole number, two and three decimal places depending on some internal rule - all on the same graph.
Similar comment for consistent padding the initial part of the number. 00.00, 01.23, 10.22 etc.
If there was an optional bar "type" option like lbl1, lbl2, lbl3 to optionally specify number of decimal places to use. Default behaviour could still be the current.
If not bar type, then something like PlotLabels (e.g., PlotLabelPadding & PlotLabelDecimals).

2 - plotBar Name - Option to not display Value
It seems like there is always a value next to the plotBar name string. Sometimes it would be beneficial to optionally not display the name values and just display the name string (e.g., PlotNameValueDisplay = 0).

3 - plotBar Labels - Consistent x-axis low and high labels
When I use limits to ensure that the plot is symmetrical in terms of x-axis lower and higher values, the left lowest label displays the lower value, but the right highest label never does.
This is a very cosmetic change

Hi,

There is:

https://zorro-project.com/manual/

Regards,

Neb

The -edit command line option would do the trick.

Thank you.

But I use visual studio. and it shows another notepad windows.

Below is what I found for this problem, somewhat weired.

I know it's already been taking care of, but improved and dynamic graphs would be awesome!

One of my Data Preprocessing R-Scripts takes 60 minutes on my MacBook Air, whereas my Core-i7 2600 Desktop needs 200 minutes for the very same task. I'm just saying... Maybe it's not difficult to compile Zorro for MacOS. Just for development, not necessarily for trading.

No, this would be a monumental task, for the same reason that we still don't have a 64-bit Zorro.

Consider re-writing your pre-processor in C/C++ DLL or Lite-C. This can reduce processing time to mere seconds. See this blog post:
https://financial-hacker.com/hackers-tools-zorro-and-r/

Or optimize your R script using the rcpp package.

I guess I will just run processing intensive R-tasks on the M1. I doubt that Zorro itself would run 3x faster than on x86, but if so, it would be nice to have.

Hi,

Any chance to add an FTX Exchange (With derivatives / PERPs) supported plugin? They have a well documented API https://docs.ftx.com/#overview and also C++ examples: https://github.com/ftexchange/ftx/tree/master/cpp

Plugins are usually added on client order. So it is hard to tell when a particular plugin will become available.