Where are the strategy files saved locally. If I want to back up. I’m assuming they are only stored locally?
Tickblaze - An Advanced Platform for Quants and Active Traders
- Thread starter Tickblaze
- Start date
Question to @Tickblaze and @fan27 among other frequent users:
I trade fully algorithmic strategies intraday on my own platform but now want to add another quite profitable model to my portfolio that does not immediately fit into my architecture for various reasons.I am wondering whether you can comment on the following points:
* my strategy is trading once a day (out of existing positions and into new positions once a day). I am not familiar how that works as my architecture is always on and constantly streams in data and potentially creates and submits orders, handles fills, and the like. How would one go about triggering a script at a fixed time each day to a) connect to broker interface and flatten all existing positions, b) obtain market snapshot prices, c) run my model, d) use triggers to create new orders and submit? Does tickblaze have any provisions for that?
* my existing framework is based purely on C#. But my new strategy is python based. I run predictions on a trained tensorflow model each day after gathering daily snapshot pricing data. How well does tickblaze support python? Can I run any python code and import any package of my liking and execute code inside tickblaze?
* most importantly does the IB broker and trading interface come with the standard package of tickblaze or is that an additional module that must be purchased separately? How stable is the interactive Brokers interface? Can you please specifically comment on a) how the platform handles daily resets of interactive Brokers? How are portfolios and positions reconciled with the broker? B) how are outages and disconnects handled? C) can the platform automatically reconnect after a temporary outage on IB's side or is this logic the user must all add himself?
I consider tickblaze to get this strategy up and running quickly until I find a way to run the model and predictions that can be subsequently handled by my C# based architecture. Heck, if I like the platform I might move more of my strategies over in the future but above questions are of serious concern to me. I have spent on the side probably more than 6 to 8 months improving my IB api plugin to handle all that could go wrong. How far are you guys in handing all that IB might throw at you?
Thanks for your help. I like what I am seeing so far. I have focused a lot of my time on my architecture but I see myself spending a lot of time maintaining and updating the code base when I need more of my time allocated to research and testing. A maintained platform that is highly capable is a commodity in my eyes.
I trade fully algorithmic strategies intraday on my own platform but now want to add another quite profitable model to my portfolio that does not immediately fit into my architecture for various reasons.I am wondering whether you can comment on the following points:
* my strategy is trading once a day (out of existing positions and into new positions once a day). I am not familiar how that works as my architecture is always on and constantly streams in data and potentially creates and submits orders, handles fills, and the like. How would one go about triggering a script at a fixed time each day to a) connect to broker interface and flatten all existing positions, b) obtain market snapshot prices, c) run my model, d) use triggers to create new orders and submit? Does tickblaze have any provisions for that?
* my existing framework is based purely on C#. But my new strategy is python based. I run predictions on a trained tensorflow model each day after gathering daily snapshot pricing data. How well does tickblaze support python? Can I run any python code and import any package of my liking and execute code inside tickblaze?
* most importantly does the IB broker and trading interface come with the standard package of tickblaze or is that an additional module that must be purchased separately? How stable is the interactive Brokers interface? Can you please specifically comment on a) how the platform handles daily resets of interactive Brokers? How are portfolios and positions reconciled with the broker? B) how are outages and disconnects handled? C) can the platform automatically reconnect after a temporary outage on IB's side or is this logic the user must all add himself?
I consider tickblaze to get this strategy up and running quickly until I find a way to run the model and predictions that can be subsequently handled by my C# based architecture. Heck, if I like the platform I might move more of my strategies over in the future but above questions are of serious concern to me. I have spent on the side probably more than 6 to 8 months improving my IB api plugin to handle all that could go wrong. How far are you guys in handing all that IB might throw at you?
Thanks for your help. I like what I am seeing so far. I have focused a lot of my time on my architecture but I see myself spending a lot of time maintaining and updating the code base when I need more of my time allocated to research and testing. A maintained platform that is highly capable is a commodity in my eyes.
Let me answer those questions:
1. Tickblaze and your strategy would have to be running and connected to your broker. The strategy will be processing tick data, minute data or any other data you like and then exit all calls to OnBarUpdate until the time when you want your logic to execute (you should have code to exit the call until the right time). Your logic can do whatever it wants to do, including flattening out positions and entering new orders.
2. Tickblaze is integrated with Python using an open source project called Python for .NET. You should be able to access packages the same way you'd normally access them from your Python scripts.
3. Tickblaze subscriptions include full integration with IB via their API. This is a stable connection that is in production for years now by most of our clients. You can use their Gateway or TWS. We don't control IB resets and disconnections but the platform will recognize the disconnect and it will reconnect when possible. Your strategy should take this into considerations. The platform has bi-directional sync with IB for the positions that it generated, protecting it from critical misalignments. Meaning if the platform thinks you hold 10 shares but you sold 3 shares from TWS it will recognize this and update you position to 7 shares. But if you add shares from TWS it will not affect your position.
You should test the platform out with a paper trading account and see whether it fits your needs.
If you need any help please contact us at support@tickblaze.com
1. Tickblaze and your strategy would have to be running and connected to your broker. The strategy will be processing tick data, minute data or any other data you like and then exit all calls to OnBarUpdate until the time when you want your logic to execute (you should have code to exit the call until the right time). Your logic can do whatever it wants to do, including flattening out positions and entering new orders.
2. Tickblaze is integrated with Python using an open source project called Python for .NET. You should be able to access packages the same way you'd normally access them from your Python scripts.
3. Tickblaze subscriptions include full integration with IB via their API. This is a stable connection that is in production for years now by most of our clients. You can use their Gateway or TWS. We don't control IB resets and disconnections but the platform will recognize the disconnect and it will reconnect when possible. Your strategy should take this into considerations. The platform has bi-directional sync with IB for the positions that it generated, protecting it from critical misalignments. Meaning if the platform thinks you hold 10 shares but you sold 3 shares from TWS it will recognize this and update you position to 7 shares. But if you add shares from TWS it will not affect your position.
You should test the platform out with a paper trading account and see whether it fits your needs.
If you need any help please contact us at support@tickblaze.com
Thank you for getting back so quickly, if I may ask a few follow-ups:
a) When you say Python for .Net you mean PythonNet? Are you talking about this: "http://pythonnet.github.io/"? Because if yes, then this won't most likely work for me, I need a full python interpreter that can use all installed Python packages, or at least the standard ones, including all dependencies of "tensorflow, pandas, numpy,...". -> Could you please do me a favor and ask your developers whether users can fully work with the tensorflow, pandas, and numpy packages and that there is no issue with the dependencies of those packages? Thanks
(here is a documented issue with tensorflow and Pythonnet: https://github.com/pythonnet/pythonnet/issues/473). Are you guys using Pythonnet? It seems quite dirty at first look, as it is not pure python code but the code needs to apparently be completely re-written or somehow embedded in a dll?
b) Do you offer access to the IB broker plugin and hence paper trading in your trial as well? For me IB plugin reliability is absolutely critical and a potential dealbreaker if it does not work in a stable manner and if it cannot recover from a disconnect or daily reset. There is no point for me to consider an alternative platform if I need to start from scratch handling all the issues IB confronts us with unless it is already handled by your plugin.
a) When you say Python for .Net you mean PythonNet? Are you talking about this: "http://pythonnet.github.io/"? Because if yes, then this won't most likely work for me, I need a full python interpreter that can use all installed Python packages, or at least the standard ones, including all dependencies of "tensorflow, pandas, numpy,...". -> Could you please do me a favor and ask your developers whether users can fully work with the tensorflow, pandas, and numpy packages and that there is no issue with the dependencies of those packages? Thanks
(here is a documented issue with tensorflow and Pythonnet: https://github.com/pythonnet/pythonnet/issues/473). Are you guys using Pythonnet? It seems quite dirty at first look, as it is not pure python code but the code needs to apparently be completely re-written or somehow embedded in a dll?
b) Do you offer access to the IB broker plugin and hence paper trading in your trial as well? For me IB plugin reliability is absolutely critical and a potential dealbreaker if it does not work in a stable manner and if it cannot recover from a disconnect or daily reset. There is no point for me to consider an alternative platform if I need to start from scratch handling all the issues IB confronts us with unless it is already handled by your plugin.
Last edited:
A. We use https://github.com/pythonnet/pythonnet which supports native Python development and packages. Nothing needs to be rewritten, no special embedding. Everything is taken care of behind the scenes. If you download Tickblaze you can see for yourself and test for yourself with our free community edition. This is the same technology used by QC.
B. We have a very mature IB adapter that will auto connect when needed. If IB requires you to restart TWS or Gateway daily we don't take care of that, but we continue to attempt connection until we can.
Live connectivity with IB is available to paying subscriptions. I suspect that if IB needs to be restarted daily this will still be the case with Tickblaze, but our adapter will continue to attempt reconnection.
@Tickblaze What crypto exchanges are you working on integrating? Do you have an ETA?
We recently added beta support for Crypto research and backtesting using Polygon.io crypto data feed.
We will likely add Kraken support and perhaps a few other exchanges, but there is no ETA for that.