TradersStudio vs TradingBlox

Discussion in 'Trading Software' started by dealer, Mar 31, 2006.

  1. Murray Ruggiero

    Murray Ruggiero Sponsor

    The real time product has not been released yet and currently we only offer an off line backtesting program , which can backtest down to 1 minute bars. In our real time product we do process tick data, for processing orders, but do not chart it.

    From polls taken on elite trader about 90% of traders thought allowing using 1 minute bars and not supporting tick by tick data for charting and indicators was good enough.
     
    #11     Apr 1, 2006
  2. "The real time product has not been released yet and currently we only offer an off line backtesting program , which can backtest down to 1 minute bars. In our real time product we do process tick data, for processing orders, but do not chart it. "

    I find this kind of funny - so if I understand you correctly - you process tick data for your orders but not for backtesting ? LOL that is funny - can't wait to see the great product to waste my time on pretending to backtest strategies that will have nothing to do with reality - lol
    :D :D :D
     
    #12     Apr 1, 2006
  3. Murray Ruggiero

    Murray Ruggiero Sponsor

    If the smallest timeframe I analyzed was 10 minute bars I might agree with you, but it is 1 minute bars. The main reason that it is not important to backtest using tick data is that you really should backtest over 5 years or more of data. In this case even 5 minute bars are good enough to see if a strategy should works. You can optimize your results and test your systems. On the S&P500 1 minute data for five years is 514350 bars so that millions of ticks. I guess you have a 100 million dollars and a bank of super computers.

    Most of our customers do not so 1 minute data is good enough.
    If you are trading on such a short timeframe that 1 minute bars are not accurate enough , I think your kidding yourself with your trading methodology.

    I know several intraday traders who have traded for 15 years the S&P500 , using 45 minute bars. They have made real money in their account and develop their systems backtesting on 45 minute bars , using TradeStation 3.0 back then, which did not support tick level access to the bars. This is a new feature since TradeStation went on line.

    The concept of backtesting is not perfect, so your comment that if it not perfect your wasting your time , does not make any sense. Here is another example , TradeStation and other backtesting platforms will report a limit order filled if you touch the price. In the real world you most often will not get filled unless you trade though the price. TradersStudio has a option to let you backtest with and without this requirement and when you see the difference in markets like the S&P500. Once you see this you will see that you are wasting your time with these other platforms. In addition how much curve fitting your system has is much more important than testing on a 3 tick chart.
     
    #13     Apr 1, 2006
  4. Its the year 2006 and people are still trying to sell products that don't address the "bouncing tick problem"
     
    #14     Apr 2, 2006
  5. 1
     
    #15     Apr 2, 2006
  6. Murray - you said
    The main reason that it is not important to backtest using tick data is that you really should backtest over 5 years or more of data. In this case even 5 minute bars are good enough to see if a strategy should works.

    so iof I understand you correctly - lets say for whatever reason I would want to test data for less than 5 years than 5 minute bars would not be ok - so why is it that for testing of data more than 5 years 5 minute bars are ok - but for only testing data of the last couple months 5 minute bars would NOT be ok ?

    (I personally think that 5 minute bars are not enough)
     
    #16     Apr 2, 2006
  7. Murray Ruggiero

    Murray Ruggiero Sponsor

    My point was that if you had the computers that NOAA uses to to create the weather models and 20 years of tick data that would be perfect. Traders don't have access to that so they need to compromise. Yes 1-2 years of five minute data is still ok. My problem is that I believe , that a model needs to be tested over a long enough time period and 1 year is not enough. This will lead to a system that might work for a few months and then blow up.

    That is why I would like to test a model ideally on 20 years of intra-day data if I could. It might still trade it even if it performed badly during the 1980's to mid 1990's because markets change.

    What I gain by using such a long backtest is that if the system starts to fail and I see that the same low volatility conditions which caused the problems during the earlier periods ,I can either stop trading it or set a point where I will stop trading it.
     
    #17     Apr 2, 2006
  8. Murray - you wrote about me -
    The concept of backtesting is not perfect, so your comment that if it not perfect your wasting your time , does not make any sense.

    I never said that - I said - quoting you first -
    "The real time product has not been released yet and currently we only offer an off line backtesting program , which can backtest down to 1 minute bars. In our real time product we do process tick data, for processing orders, but do not chart it. "

    and then I said
    "I find this kind of funny - so if I understand you correctly - you process tick data for your orders but not for backtesting ? LOL that is funny - can't wait to see the great product to waste my time on pretending to backtest strategies that will have nothing to do with reality - lol"

    I never spoke anything about perfection - so if you are going to quote us - the least you can do is quote us accurately
     
    #18     Apr 2, 2006
  9. Murray - Previously you wrote -
    "The concept of backtesting is not perfect, so your comment that if it not perfect your wasting your time , does not make any sense. Here is another example , TradeStation and other backtesting platforms will report a limit order filled if you touch the price. In the real world you most often will not get filled unless you trade though the price."

    here is where I agree with you - and that is why one should not use tradestation
     
    #19     Apr 2, 2006
  10. Murray - previously you attacked me by saying -
    "Most of our customers do not so 1 minute data is good enough.
    If you are trading on such a short timeframe that 1 minute bars are not accurate enough , I think your kidding yourself with your trading methodology. "

    I do not know why you are answering in an attacking way – I am not the one who is trying to sell anyone anything !

    My trading methodology works ! I have been profitable for the last 18 months. (the 5 months before that were not profitable). The drawdown I had on a monthly basis is less than 10% - and on a intra month basis 15% which happens to be 25% of the annual rate of return that I have been achieving.

    And there are a couple of answers to your up top quoted statement. I will try one of them and hopefully you will be able to comprehend.

    There are many of use who always like to go home flat overnight – As a result we day trade but are not daytraders. We might only do a couple of trades a day - and in most situations 20 minute bars would be good enough. But there are times when things might get down to 10, 5, 1 and even the ticks within a one minute bar - where the trading may be so volatile that one has to check at the ticks within the one minute bar to see if a profit or stop out happened first - these times are rare but do happen - and any back testing program that does NOT include for these situations is not a complete backtesting program – and if you think differently, you are kidding yourself !

    Murray - If you are not able to comprehend the above aforementioned, I will be glad to dummy it down for you – and/or give you another answer
     
    #20     Apr 2, 2006