Bollinger Band Position Trading Example

Discussion in 'Strategy Building' started by Hook N. Sinker, Sep 2, 2006.

  1. But the back test here has not been conducted in a proper manner. Why?

    a) Selection bias (as already pointed out - the fact that the security is still traded and still considered "blue chip" suggests a long only strategy will be successful
    b) Market bias - uses a long only strategy in a market (ie equities)that everyone knows has only gone up on average for the last 40 yrs
    c) Only 12 trades - we cannot have any statistical confidence in these results with such a limited sample. Need 30-50 trades MINIMUM

    etc, etc

    Naturally, if we can test a portfoilio of DOW stocks from 1960s until now (without changing the components of the portfolio) with similar results for the overall portfolio, then it would be more interesting as an investment strategy.
     
    #11     Dec 12, 2006
  2. You're right about all of that although I would point out that long only is not necessarily a bad strategy just because the last 40 years have seen stock increases. It is the inherent nature of stocks to go up in price as earnings go up. This is not the case for currencies, commodities or futures contracts, but for stocks, it is (until something changes, which it could, I acknowledge). Still, your other points are good ones.
     
    #12     Dec 12, 2006
  3. basis

    basis

    Here's a system that makes about as much sense:

    Buy when the date is January 1, 1900
    Sell never

    1 trade, open profit +1,000,000% roughly, or 10% per annum
     
    #13     Dec 12, 2006
  4. asap

    asap

    Yes, this is based on basis band strategy instead.
     
    #14     Dec 12, 2006
  5. You mean 15.33% per year, not 307%.
     
    #15     Dec 12, 2006
  6. kut2k2

    kut2k2

    Yep.

    from the backtest:

    Total profit $ 13615410
    Profit after subtracting $ 10.00 commission, slippage per transaction: $ 13615170
    ...
    Initial capital is $ 100000

    One share of BA bought 02Jan1962 would grow to 190.89 shares on 14Jul2006.

    So that initial $100,000 would have turned into $19.1 million from buy-&-hold.

    This strategy has cost over $5 million in profit, not counting the aggrievation of executing it.

    I'll say it again: no matter how "profitable" a strategy is, if it's not beating the buy-and-hold, it's a waste of time and money.
     
    #16     Dec 12, 2006
    dartmus likes this.
  7. dartmus

    dartmus

    To the naysayers. To those who insist TA doesn't work.

    Buy and hold isn't a solution unless you pick the right symbol. If you can pick the right symbols then there's a logical error in kut2k2's analysis. Less than buy and hold actually can be a v profitable strategy. Tho of those who develop strategies for picking the right symbols ...survivorship bias gets many of them.

    Bollinger band strategies are crude compared to for instance Keltner bands and I don't recommend either of them but everyone should investigate them thoroughly along with all the other well know strategies.

    Investigate them so thoroughly you eventually come to see the essence of the problem from a philosophical point of view because then you will see the seed of a gem within kut2k2's brilliant analysis.
     
    #17     Oct 3, 2015