I am brand new to trading, but I am interested in building an automated trading system in Java, and I've seen that certain brokers have APIs. I am looking for the simplest, but mostly the cheapest system that has a Java API, including broker + API. A manual interface is not needed, although I guess it could be useful occasionally. Are brokers strictly necessary for easy trading? I've looked at several brokers and they seem to be full of non-transparent fees, and I need almost none of what they provide (I don't need their data, advisors, etc.; a visual summary of gains, losses, etc. might be nice but not necessary). Thank you for your help!
Thank you for your replies, and I'm sorry for the delay. IB seems far and above the best option. Unfortunately, they require a $10000 account minimum, one year of trading experience, and 100 trades. I don't know if they actually verify any of this; but it's kind of annoying that I have to use an inferior service and waste a lot more of my money in order to meet the requirements to use a better one. The worst part is that I'm only interested in writing software to automate the trades, so I'd have to build the entire platform using the wrong API, assuming I ultimately wanted to switch to IB. Any advice?
Here is my advice: 1. Buy at least 10 years of quality intraday data for the markets you intent to trade 2. Use an off-the-shelf platform to try find 1 (several) edges. If/when you succeed on 2., with at least a sample set of 2000 trades averaging $100 or more per contract after comms & slippage, then you can consider your options to turn this into an automated trading system. If you have no money to spend on historical data, and no money to fund your account, there is really no point thinking about trading.
Kafe, IB is the best choice really, especially for Java. Commissions are low (for retail) , and their requirement is lower than most other brokers' to use API, unless you are planning to do FX trading. If you don't have 10,000 to trade - it'll be tough to do get into automated trading...
Thanks for the advice. Yeah, IB seems far and away the best option. If I do it before my 25th birthday, I can have an account with a $3000 minimum instead. Sounds like the way to go, but I'm trying to find out if I can practice with the API before creating an account just like the other trial versions. It would really help solidify my decision. Also, I'm wondering if I maintain that $3000 minimum after turning 26...
You can only get a paper trading account once you open AND fund a real money trading account. You can always pull out all but $500 to continue receiving real time quotes, but if you do not generate at least $30 a month in commissions, they will charge the account.
Kafe: If you are going to trade in an automated manner there is a very good chance that you will become a "pattern day trader"(PDT). In order to trade as a PDT you must have (and maintain) an account balance at or above $25,000. For all practical purposes this means you'll need about $30,000 to allow for market fluctuations. FYI - FINRA and the NYSE define a Pattern Day Trader as one who effects four or more day trades (same day opening and closing of a given equity security ("stock") or equity option) within a five business day period. If you cannot come up with the $30K you'd have to build a system that rarely, if ever, day-traded. Futures contracts and futures options are not included in the rule. Also: It has been awhile, but you used to be able to run an API against the IB Demo account. It is available to non-customers. It is nowhere as good as the Paper Trading Account as the Demo account data is not real and has some other limitations that have probably changed since I last used the Demo. The Demo account used to run 24/7 so you could test against it. That wasn't its purpose, but it was good enough for me at the time. I used to use it for simple interface testing before IB started allowing Saturday logins to regular and paper accounts. Jack
Hey guys, thanks for the info, especially the bit about the day-trading. I think I can avoid issues, as long as I avoid day-trading. The automated method is about prediction, rather than reaction, so maybe I will just trade to the best of my ability for the following day, rather than same-day. I'll just limit the frequency of my trades until I have a decent proof-of-concept (which I should do anyway). Will this let me avoid the issues you mentioned? I'd love to run the API against the IB demo account. It wouldn't be ideal, but as long as it has frequent changes in the data (that aren't completely random) it might still work as intended. Otherwise, I'll just bite the bullet and fund the account, so I can use the API and paper trade. Again, thanks so much for taking the time to write this out; it helps a lot.