Holy Grail Strategy (knowledge vs. opinion)

Discussion in 'Strategy Development' started by RockRudy, Nov 9, 2006.

  1. RockRudy

    RockRudy

    I would like to ask our experience traders to answer my question below. Unless you are a trader that consistently makes money do not answer please. I am looking for a knowledge and not an opinion. The former are scarce, the latter are all over the place.

    I have been in a lot of stages of trading. Starting with emotional gut feeling trading, through indicator trading, up to strategy trading. I am still working on the last one and here I need your help.

    Back in the past I realized no one can predict a market (what a huge step forward in my trading), then I realized that there is no holy grail indicator (again a step up).

    Now I have back tested tens and tens of strategies, most of them my own. Some of them are pretty profitable and others are not. What I find myself always doing is that once I find a profitable strategy I back test it on a multiple stocks and I always find some that the strategy does not work well with. Then I try to tweak/optimize so that it will always work well. But then, after spending countless hours on this, I realize that I am again chasing a holy grail strategy, that I already know does not exist.

    So I am not asking you to solve this for me, but a good kick in a good directions would be appreciated. How do you get to the point that you are comfortable trading a strategy? Do you back test it on a single stock only? Strategy may work in year 1 but not year 2 or for stock 1 but not stock 2 . How do you make the call for a good strategy? What are the criteria?

    Thank you.
     
  2. vikana

    vikana Moderator

    The "Holy Grail" in my opinion is the following: trade many strategies with small positions over many stocks.

    Each strategy should be simple with few "variables" and be universal across all stocks. Must also (not surprisingly) have positive expectancy.
     
  3. RockRudy

    RockRudy

    "be universal across all stocks" that is exactly my problem. Is it really reasonable to expect a strategy to be profitable across all stocks? Across all time periods?
     
  4. Vikana is correct...adding stock select and bar creation criteria (volume,tick etc.):D
     

  5. There is your problem. You are just barely getting started. You need more patience. Ive tested tens of thousands of strategies before I found a few I would trade.

    They work across baskets of 2000+ stocks, for all years, in all market conditions. I wouldnt trade anything less.

    Ive executed millions of backtests to find these areas of non-efficient pricing.
     
  6. RockRudy

    RockRudy

    There is my hero: traderdragon. So there is a holy grail strategy! Anyway, I thought one should focus on several strategies and master them. I can spend hours/days/weeks on tuning and optimazing couple strategies before I call it inefficient. How would I know that a strategy is inefficient if I did not spend that much time playing with it. You cannot do that if you are testing thousands and thousands of strategies. If you do, then your luck is just like a lottery win whether you happen to stumble accross one that actually works.

    I am fine with patience. I do not require results now or even soon. I would however like to know that I am walking in the right direction. Just seeking the knowledge.

    Shouldn't one have a strategy for finding a strategy? What do you think?
     
  7. for pete's sake

    there ... is... no... holy... grail...

    i think what is more important is that the search for a holy grail is counterproductive to trading.

    develop a strategy, whether automated or discretionary that has an edge.

    use strict money management and rules and let the edge work itself out over time

    choose a methodology and market that matches your personality and goals.

    traders start making money when they stop looking for the holy grail, or the magic indicator, and learn how to manage risk
     
  8. vikana

    vikana Moderator

    More specifically (imo): You do not need for every individual security to make money. As long as the test across ALL securities makes money with good statistics you have something tradable.

    Re time periods: I would not expect the same rules to work EOD as on for instance 1 min bars. If you have a 5 minute system it should also hold up on 3 min and 10 min bars or I'd find it suspect.

    Also, make sure you read ButterMilks reply.
     
  9. No, there is no holy grail. Yes, you need a strategy for finding strategies. Mine consisted of very advanced backtesting.

    Im able to test thousands of strategies thru automation, and do millions of backtests via slick custom software.

    You know have found something inefficient thru statistical analysis. I can absolutely prove my strategies are statistically valid and not data fitted, or I wouldnt trade them.

    There is no luck in finding my strategies. Data minining billions of of sets of data and finding any inefficiency is not luck, its methodical scientific snooping.





     
  10. jrlvnv

    jrlvnv

    Grats to you on finding a strategy that works for you and makes money. I have yet to be able to do that. The part of your statement that I question is the part where u say you need more patience......

    Say the strategy that you currently use, you came across it as maybe the first dozen you tested instead of going over tens of thousands...... So here you have a strategy that works but with very little research as far as testing is concerned. Do you feel you would of traded differently because you didn't go through the years of pain trying to find something that really works?
     
    #10     Nov 10, 2006