Is it really an Edge?

Discussion in 'Trading' started by tanp21, Aug 20, 2003.

  1. Randomness in probability theory has to do with statistical independence. In your example randomness would be the case provided the outcome of the flip is statistically independent from your prior decision to go long on heads. This would be so in the case of a "fair" flip.
     
    #41     Aug 21, 2003
  2. My question to everyone out there, how do you know your method has a real edge?

    I know I have an edge when I have to pay taxes at the end of the year from my trading profits. (just one way to look at it - pretty important one though) This incorporates my method having an edge and me having an edge - the two must work together.

    My method is a trend following method.
    My method does not have random entries.

    However, (imo) a random part of the time, the market moves randomly. Therefore, some of the entries could seem to be random, but they aren't.

    My method includes money management rules. To each his own, but it's hard stops for me (from the time of the entry the stop is made as a live order). I can move my protective stops after my position turns profitable.(even so they are still hard stops because they a live orders) In addition to stop loss exits, I have a profit taking strategy.

    My method does not allow for adding to losing positions. (again, to each his own)

    My method does not allow for picking exact tops or bottoms. The price must have at least moved some in my direction of entry, before the method allows or I might say, provides a signal.

    My method does have several variables (in others words - more than one thing needs to line up - before I can consider the signal to be valid of positive expectancy) Although it has several varibles - the method is simple.

    To answer the question - Yes, my method has an edge - more often than not it does produce winning trades.

    My edge consists of a gameplan and rules as well.
    Of these rules one of the top rules deals with when not to trade.

    I have several of these when *not to trade* rules each of course based on experiences with losing money. (again to each his own)

    A method can have a great edge, but it is still left up to me to trade it. Therefore, I must exercise the discipline to trade it properly.

    That is *my* edge, therein lies either death by 1000 cuts - death by a few cuts - or the holy grail of consistent profitable trading.
     
    #42     Aug 21, 2003
  3. This problem comes up in many fields. For example in simulations, encrypting messages, etc. How can anything in a computer be random?

    In fact nothing calculated in a computer is random, it is always deterministic. The way this problem is tackled is by so called "random number generating algorithms", which generate in truth only "pseudo-random" sequences of numbers. Mathematicians can control these algorithmic processes to such a degree that insofar the statistical properties of such sequences are concerned, they can be made arbitrary close to true randomness as needed.

    Application to trading: Generate a time value based on a random number generator and enter your position when this time comes up. Voila, you did it.
     
    #43     Aug 21, 2003
  4. Fine. But AFAIK none of those people understand probability vs randomness as well as brother dbphoenix does. All feuds aside, this is more his field, anyway.
    It's not a random entry. It is biased. The fact that you drew a line and then enter above that line automatically makes it non-random already.

    Why did you draw the line there? Why not a little higher / lower? Makes it non-random. Also, there's obviously an uptrend in that chart, so you're probably biased.

    Or would you enter on the short side here? Well if you wouldn't, this is certainly not random. And I suspect you wouldn't.

    You may also have a particular long bias (as your name suggests), which automatically makes it biased, thus non-random as well.

    There are most certainly millions of circuits in your brain doing subconscious work to give you some sort of bias, making it non-random.

    Random is what I'd call a computer-generated entry decision (without a system-bias), or rolling of a die. Nothing else is random.

    You're putting on the trade yourself = Not random.


    And by the way, your jokes on dbphoenix are rotten. How about your new "Don't post crap" regime? I bet if anybody else made such a joke, you'd be the first one to delete it. Your signature below that just made me laugh. What a joke! First you take out bubba7 and then you're moving on to db. If you're in a bad mood or losing streak, don't moderate.

    Just my 5 cents,

    ~The Scientist [​IMG]
     
    #44     Aug 21, 2003
  5. The way you want to describe random, there is no way for it to occur. I have always said the same thing about the line. Put it on the chart at the close of the most recent bar. You pick the stock, I don't care. You pick the time frame... I don't care. I chose the most recent bar because that precludes using previously established support and resistance. I even blocked out the prices. I would have blocked out the text but Id idn't feel like changing the appearance settings. Just step up to any chart, draw a line thru the most recent bar and go. You don't even have to use the close. Draw it anywhere in the bar. You only have one of two ways to be in, long or short. Pick a downtrending stock in your opinion. Pick an uptrending one in your opinion. I don't care. Youset up the conditions, I don't care!

    Here is a good way to do this. Scroll backwards on any chart any time frame. Scroll back 100 bars. Draw a horizontal line somewhere on the last bar you have showing. Now trade it forward. Or scroll back to some time period that you do not remeber well. Use a time frame that you do not normally use because it doesn't matter.

    What I would really like is to read yours or db's definiton of what a random entry is, not what it is not. I have yet to see that.

    And check out the new thread I will start to demonstrate.

    Btw, I love that little blue face thing.
     
    #45     Aug 21, 2003
  6. Scientist, I just posted this to you:




    scientist,

    As "THE" scientist indicated in his prior posts, he truly goes nuts about receiving some fan mail once in a while. I am touched that a scientist of your calibre could be so sensitive a person. I thought it would perhaps be nice to send you these lines.

    Remember we are true fans of you, impatiently awaiting your book(s) that you mouthed off about. We will not forget about you. Better start doing something serious now. Don't let us wait too long or we'll fix your watch. We like a truly big scientist around here but don't care too much for half-wits.

    Lie low now and show us your first chaper soon. We all like to learn something from a biggie like you.

    So long BIG Aussie scientist. [​IMG]
     
    #46     Aug 21, 2003

  7. you better stop giving away my trademarked secrets !

    :(
     
    #47     Aug 21, 2003
  8. Darn it surf... I forgot. I am sorry bro'! :)
     
    #48     Aug 21, 2003
  9. Yeah that's pretty accurate.
    You're not getting the point here. It's not random. Why the close of the bar and not the high? Which stock / why? You have no timeframe preference? So a monthly chart is OK for you? The fact that you're drawing a line somewhere doesn't make it random either. I mean, you're not just closing your eyes, then dropping line, are you? You're obviously consciously placing it somewhere.

    You might as well say that the "holy grail" strategy is random, too. It's not random.
    My definition of "random entry" I just defined : Either a computerized random # generator or rolling a die. Or tossing a coin if you wish. At the last bar you have up. Toss a coin, head long, tails short. Then I suppose you could take that bars' OHLC/4 to enter. That's random. Nothing else is random. It's as simple as that. Even if you toss the coin for either long/short, as soon as you decide to only trade if the high or low or close is broken, etc. you're not being objective about randomness. You must take OHLC/4 or the whatever and then enter there, or enter straight away. And in order for this to be really random, you have to do it 85,000 times, so that the number of heads/tails in your statistic balances out to an exact 50/50. Good Luck.

    I'm glad you like the blue face, brother inandlong. Don't upset too many ppl here or you might get one, too... :p


    Compliments,
    ~Scientist [​IMG]
     
    #49     Aug 21, 2003
  10. "Random" can be what ever the AUTHOR chooses it to be within the context of HIS particular system. HIS only constraint is that he be CONSISTENT in the use of his definitions.
     
    #50     Aug 21, 2003