Home made indicators anyone?

Discussion in 'Technical Analysis' started by Remiraz, Nov 1, 2003.

  1. dbphoenix

    dbphoenix

    Actually, it depends on how you conduct the backtest. I always do mine manually for just this reason. That way you can apply as much discretion as you like.
     
    #11     Nov 1, 2003
  2. It will be difficult to just trade by an indicator. THe reason might be that there are many better ways to trade. Most people just create indicators to make a "name" in the industry. Some are promoting theirs here. It is not very difficult to create a new indicator and if you try long enough to make a hype out of it you might get popular. Here is an indicator that I developed just with drag and drop (no programming). Just to show you the difference. The lower indicator on the chart is the PositiveVolumeIndex, an indicator created by a member from the wealth-lab site, with a 31 moving average.
    The upper indicator is the 10 bar (day) RSI of the PositivVolumeIndex and a 31 moving average. Note how both indicators (RSI and PVI) look a like but also see how much more profitable the 31 SMA would have been on the upper indicator. All done by drag and drop. Am I famous now? :)


    Volker

    PS: I would never trade it.
     
    #12     Nov 6, 2003
  3. Someone's rubbish can be other's treasure! :confused:
     
    #13     Nov 6, 2003
  4. Happy to read your PS !!
    Yes, I agree, avoid trade it. In general, a system with a -50% over the whole N100 market, from Jan2000 till now, is a way to loose some money, but there is no reason to create indicators for this.[A typical MACD crossover would give the same result]
    Drag and drop is a method, but what is the result ?
    If "home made" indicators have this -50% fortune, then they are not necessary.
    If you haven´t already backtest this drag&drop idea, here you are !!

    p1=PVI();
    p2=RSIa(p1,10);
    Plot(p2,"p2",4,1);
    p3=MA(RSIa(p1,10),31);
    Plot(p3,"ma",7,1);
    Buy=Cross(p2,p3);Sell=Cross(p3,p2);
    Short=Sell;Cover=Buy;

    Settings:trade at +1 Open, commission 0.25%
    Results : [Jan2000 till now, over the N100 database]
    Winners:14
    Losers: the rest
    Top winnerS:BRCM, +133%, CTXS +116%
    Rest winners:below +100%
    Total results: A net loss -51.30%

    Creating indicators may be easy, creating useful [read profitable] indicators is not as easy as drag&drop !!
     
    #14     Nov 6, 2003
  5. Hello All.
    I sometimes wonder when I visit the Stockcentral MetaStock forum if members spend all their time programming indicators and 'experts' (trading systems). When I see posts asking about indicators and 'What is your best indicator' etc, I have to wonder if people are asking for a magic indicator that actually does the trading – or makes them do the trading.

    Constance Brown's 'Technical Analysis for the Trading Professional' has a lot to say about indicators, but I think that Lobster has best stated the usefulness of indicators, that ultimately, it is NOT the indicator that does the trading, it is the trader. Wallace.
     
    #15     Nov 6, 2003
  6. Thanks for testing it, but who in his clear mind is paying 0.25% commission? I would be interested in the results again, but use 1 cent per share for each trade (IB Commission) and use 5% of the overall equity for each trade. Hey, who knows, maybe I developed a good system? Thanks for your help and the quick responce.
     
    #16     Nov 6, 2003
  7. It depends on the ...system. When the system looses money, 0.01$ is OK. When the system makes money, then give some 0.25% to your [personal] broker, he will be more than happy [I want smiling faces all around]
    On the subject again, yes, you may set commission per trade or commission per share in your amibroker settings and run the backtesting.
    I dont see any difference between a -50% or -40% average.
     
    #17     Nov 6, 2003
  8. Sorry, Mr Amibroker, here is a miscommunication. I am not out here to show the ability of the commission settings of any software, that is too trivial. I wanted to safe some time and asked you to test the system again with the conditions that I thought are more realistic then the once you came up with. I don't care which software you use, even thou I, or actually we "all" know. :) Since you had the system programmed, I am sure it would take you a few mouse clicks to make the changes (like in any other software) and post the results, simple. Thank you.
     
    #18     Nov 6, 2003
  9. I am not Mr Amibroker.
    My name is Dimitris Tsokakis, born and live in Athens, Greece, 1.82m, 80-82klgr, brown-green eyes, brown-white hairs, married [once], two kids, one dog, what else...
    My [only] relation to amibroker.com : I am a registered user, nothing more but also nothing less, no matter what you, or "all" of you may know or ignore....
    I am not here to prepare your morning coffee, dear, please try somebody else for the service...
     
    #19     Nov 6, 2003
  10. Kap

    Kap

    I have my own home made indicators for over bought/sold conditions, they differ for different markets (indexes), and some for arbitrage models, dare I say to date they work 85% of the time.

    The old indicators that did work are now so freely available that its better to use them as a contrairian indicators ! :)
     
    #20     Nov 6, 2003