looks cool, and relatively cheap, but as now my understanding is that I have to spend few bucks I am going to use tc2000, tradingview, or IB for specific alert, and then use separately python. I am not a day trader, just need real time to maximize returns
Thank you for the long explanation and the effort in replying. At the moment I am a mobile developer in career, so I do quant as an hobby, just because I love trading even more than the money, so if I gotta do run subroutines repeating operations, handling errors to get real data I have to decide to do it as a living(because yes as best scenario want all the data in real time and query them after they are persistent without giving my (hopefully) winning algo to a third part company. I am moving my first baby steps as quant, although I have both experience of trading and development, this is why I am thinking now MVP(minimum viable product) just fetching in a simple SQL db the historical data, and from there doing some simple backtesting and execute trades on yearly chart, if should be successful I will then think about real data. If they come with a fairly cost, as is acceptable that is, I want to obtain before fairly returns, during my part time trading
yes I guess the solution is to have different cloud machines and making scalable and linked each other, but before to do that want to have enough returns to do it as a living. after all a good part time trader needs just alerts, a scanner, a good broker, and some time in the week ends
You are correct: you can send a maximum of 50 messages per second from your client to IB's servers. So you would need to pace the request for the first batch of 100 tickers. The data returned from IB will be asynchronous, so with some clever software it should be possible to switch from one ticker to the next when you receive data. I don't think that the upload rate limitation will then be much of an issues any more.
So i guess in python I do a request for 90/95 per tickers, then set a small delay before every GET call if is a REST
No you can't use a fixed delay time, because you don't know how long it takes for IB to return data to you. In case you request historical data IB will send you a message once it has sent you all data you requested. You can use this message to submit a new request for another ticker symbol. If you are subscribing to live market data then you'll have to wait until you receive a price. Once you've received a price you can close this subscription and open a subscription for another ticker symbol.
Also check out ActiveTick, they have different subscription tiers for 500, 2000, or unlimited _streaming_ tickers: https://www.activetick.com/activetick/contents/MarketDataServicesPricing.aspx This makes it possible to get your app working with fewer tickers first. Btw, you never do requests for a gazillion tickers at a time - stream to a local aggregator and query that with python if you must.
well last week I paid for tc2000 and get alert and can write my business logic to the scanner.Cannot backtest and have some limitations that if I were using python, but still has tremendously changed my game. Said that I understand what you say, I m an android developer and chaining REST requests is what I do on weekly basis, great recommendation anyway, I am keeping experimenting with python but with less stress at the moment. I really like the feeling to get an alert and decide my self if want to exit or maintain the position, although taking losses is the basis of a succeful trader without any doubt thank you happy coding
hello cstone949 - I am a long time user of activetick streaming data - would like to talk if you can spare a few minutes