Introducing The BAT System: Bot Algo Trading Made Easy.

Discussion in 'Automated Trading' started by JCTradingGroup, Jul 13, 2012.

  1. JCTradingGroup

    JCTradingGroup ET Sponsor

    Introducing The BAT System: Bot Algo Trading Made Easy.

    www.TheBatSystem.com

    The BAT System, powered by Trade-Ideas, is a highly customizable, fully automated algorithmic trading system that requires NO PROGRAMMING KNOWLEDGE!

    The BAT Scanner allows you to easily build sophisticated strategies with no programming knowledge needed. With more than 400 different alerts and filters that are highly customizable, the number of strategies that can be created is infinite: ALL WITHOUT ANY COMPUTER PROGRAMMING NECESSARY.

    After you have developed your algos, easily backtest them over the entire market in a matter of seconds with the BAT Tester. Then easily optimize your strategies for optimal performance.

    Once you have your algo perfected, let the Bat Bot autotrade it for you.

    Create your algos from scratch, or choose from the numerous strategies in the BAT Community that can easily be customized to create your own version. NO PROGRAMMING KNOWLEDGE IS REQUIRED!

    Find out more about the BAT System at:

    www.TheBatSystem.com
     
  2. Without programming, aren't the algos limited to "been there, done that"?
     
  3. on whose machine do the developed strategies reside? When I backtest strategies on which computer does such backtest run?

    Thanks


     
  4. JCTradingGroup

    JCTradingGroup ET Sponsor

    Besides traditional indicators and events, the BAT System can easily create algos based on NASDAQ Velocity and Force Data (http://www.nasdaqtrader.com/Trader.aspx?id=historicalvf), as well as StockTwits data based on individual stocks. I do not know of any other algo trading system that uses that data. Here is a complete list of the over 400 alerts and filters:

    http://www.thebatsystem.com/list-available-alerts-filters/




     
  5. JCTradingGroup

    JCTradingGroup ET Sponsor

    The developed strategies reside on your computer. All backtests are ran on our servers.


     
  6. Does that not mean your servers also have to have access to the strategies?

    (Personally, I don't care, but I have a feeling others feel they should. :))
     
  7. another "you can develop strategies even you know nothing about programming, micro market structure, but should we find your back tested results look good we make sure to have your code and we will peruse it in ways we will certainly not share with you". This is just the 2012 version of "trade with us, you can strike it rich even as a beginner and even we as broker quote you fx spreads 5-7 pips wide". As long as you appeal to the lazy and hopefuls you always stand to gain something no matter what business you are in. Sorry but what you offer and how you communicate it STINKS!


     
  8. JCTradingGroup

    JCTradingGroup ET Sponsor

    If you took the time to go to the website you would see that the BAT System does not trade forex, just US equities on the NYSE, NSDQ, and BB Market.

    Unless you own and maintain your own data, ANY backtesting would need to run on a remote server. Through our partners at Trade-Ideas, tick data, along with bid/offer size, is available for testing on over 7000 stocks over the last 30 days. Backtesting a strategy over all 7000 stocks takes only seconds because it is performed on the TI data farm. If you already have tick data and quote data on every stock that has traded on every exchange over the last 30 days, and have a computer powerful enough to run your backtests quickly....then maybe The BAT System isn't for you. However, there are many traders that don't have that data available to them and don't have the programming skills to develop advanced algorithms.

    Furthermore, the BAT System can be used at multiple firms.


     
  9. bears21

    bears21

    Is there a way you can backrest results longer than 30 days that seems too small of a sample. For instance over the last 30 days since markets been choppy mean reversion probably did well on backrest while trending strategies not so much.
     
  10. a) you did not get my point, probably because I did not make it clear enough. References to brokerages and fx trading were an example of individuals who touted the "easy way to the riches" in the same way than what your website promises "NO PROGRAMMING KNOWLEDGE NEEDED", yet "NOW INDIVIDUAL TRADERS CAN USE THE SAME ALGORITHMS AND STRATEGIES AS SOPHISTICATED HEDGE FUNDS" (I paraphrase here) but anyone who falls for such snake oil is a moron in the same way than those who believe they can make money in fx markets with 5-10 pip wide spreads. I dared to bridge the two because the advertising approach sounds all too similar.

    b) every retail focused data vendor supplies 30 days of tick data. Thus may I point you to the fact that you can access 30 days of tick data for more than 7000 stocks for way less than 100 USD per month. Someone who cannot afford that is not supposed to be in this business anyway. So, please do not make it sound like you this mere part (the historical data) is a groundbreaking service you provide.

    c) "Backtesting a strategy over all 7000 stocks takes only seconds because it is performed on the TI data farm". Let's do some simple math to verify your claim. I am willing to bet my life that you are unable to stream more than 5,000,000 ticks/second through your testing platform (even multi processor), at least I do not know any testing platform, no matter how commercial, how optimized, that can stream at such throughputs, on a single machine (maybe 5-6 high frequency houses world-wide will push markedly beyond this border with top hardware). If you claim your trading platform can do better then that then I look very much forward to a fruitful discussion of the technology involved and how you achieve such. So, lets assume now you still get to 5million ticks per second, even though as soon as you run a portfolio backtest you not only need to read multi-symbols into your platform but you need to merge the streams, thus the processing time for >1 symbol will then be slower than 5million/second:

    Let's assume 150k ticks per symbol/day, so you then process 30 days worth of 150k/day tick data in about 1 second, doing this for 7000 stocks will take you 7000 seconds (taking into account the huge credit I gave you that you manage to scale up to 7000 stocks at about the same efficiency than streaming just one symbol), thats about 2 hours. Hmm, did I make a mistake? Did you not just claim it would take seconds to perform such test? Something off between our calculations? "Some seconds" that you claim mean your testing engine can process about 720x5million ticks per second = 3.6 billion ticks/second. LOL, even with the most massively distributed and parallel computing super grid you will NEVER get even close to those numbers. One high frequency trading house that streams in batches claimed recently to be able to process 60 million ticks per second.

    -> With all due respect but you are talking shit!!! Please double check your numbers because they are off by several orders of magnitude.



     
    #10     Jul 15, 2012