Is Quantitative Analysis the only way to get an edge in the market?

Discussion in 'Trading' started by _eug_, Apr 2, 2017.

  1. Thanks the themickey,

    Good point.

    Define "x" days please?

    What happens a month or two after click starting on the algo(s), the algo(s) goes into drawdown?

    If my algo drawdown prediction is $10,0000 and in about 7 months after starting algo, I am down -$6000, what do I do?

    Do i just stop the algo or continue?

    I choose continue, cause the backtest and myself predicted drawdown near $10,000.

    Do you understand my logic on what I am stating? There is no what if algo(s) losing money, then do this. It will either make money or not. If it losing the entire $10,000 over x amount of years, then that's the way it is. Obviously, I did not have a winnable algo as I predicted.
     
    #81     Jun 12, 2017
  2. themickey

    themickey

    'X' days is for you to determine because only you know what time lines your trading system works on, ie does it also use intraday signals, how frequent trades it makes a day etc.
    After about 10 trades minimum hopefully you can see a pattern, but that also depends whether your system is in tune with the market, aka, are you running a long only system but market turns bearish or goes into a chop period.

    My suggestion is eyeball your backtesting results and advise yourself how long to run this experiment.
    If you believe that your system could easily hit $10k drawdown, then allow it to hit 10k drawdown, possibly allow another 5% to 10% on top of that if you can afford it.
    But, this too is my suggestion, when you finally do start the real exercise, you should go into immediate profit. If you think you have timing right expect a profit. Were it to immediately go into a loss I would seriously consider having in place beforehand a stoploss not too far beneath your entry. Why go into an immediate into a $10k loss, you will have gained no experience.
    One other thing, ensure when you kick off that market is not too far away from support, you want to catch a fresh run, and not entering toward the end of a tired trend.
     
    #82     Jun 12, 2017
  3. sle

    sle

    If you see that your algorithm is losing money, do you follow a process? I.e. do you try to figure out why your results are not corresponding to your model?
     
    #83     Jun 12, 2017