Does anyone use linear regression to day trade the markets?

Discussion in 'Technical Analysis' started by kevinmclark, Sep 5, 2005.

  1. I have been searching the internet (including various forums) for generic strategies to use linear regression lines and/or channels (standard deviation and standard error) to day trade the eminis to no avail. Does anyone use these tools? If yes, please describe how to best use them. Thanks!

    Kevin
     
  2. Linear regression is a powerful technique of Applied Mathematics applicable to many problems. As with many such tools it also enables one to produce immense heaps of garbage.
    Frankly, if you want to make money in the market, Internet searches on regression ain't going to help much. As long as you don't become convinced that people who know aren't going to tell you, you will remain far from your aims.

    This is the best I can help you,
    nono
     
  3. It is the basic big choice for making money.

    The choice that uses envelopes and periodicity

    OR

    The choice that uses the centered path of the market and its rubberband, the tendency to revert to the mean.

    This forum is for:

    "TA junkies, this is your forum for talking about technical indicators and chart patterns."

    So what you are talking about may be misjudged as not being "technical indicators and chart patterns."


    As usual it is best to be informed about both thoroughly. I always ask: "How does this help to make money?"

    The path of the market is a fine thing to know about. But mostly it comes down to handling turning points of the market.

    These turning points occur at extemes on the envelop it turns out. By focusing there at those extremes of price, is where the monitoring, analysis, decision making and timely action really pays off.

    It like asking: "Do 'technical indicators and chart patterns.' help to make money. May be they do but not as well as envelopes and periodicity.

    I am not welcome in the TA forum because I and others like me cause the moderator to raise the question "Where's the TA?".

    It will be better if I don't post here anymore as a way of keeping the thread open for you.

    I vote for the envelopes and periodicity stuff which is determined by the P, V relation. Math comes second; the market's ooperation comes first.

    It is not a debate on right and wrong.
     
  4. 0008

    0008

    If you can use it to make profits or even hugh profits, would you share it with others? :D :D :D :D
     
  5. I’m seeking tools to better trade at extreme levels; therefore, I plan to study if linear regression lines work better than moving averages; and, if linear regression channels are useful to forecast major support and resistance levels.
     
  6. Is this the reason why there is so little information about how to trade with these tools?
     

  7. No 0008 is in la la land.
     
  8. maxpi

    maxpi

    I know a guy that was an intern with Long Term Capital Management and his job was doing some regression analyis. He said it "did not work".

    I have looked at channels built around linear regression lines for use in detection of turning points, I can do as well [or as badly] with it as I can with MACD/Stochastic. Some people really know how to read the action and have statistical things worked out as to how many breakouts to the low side result in a reversal, etc. I am not that far into the game. Find some books on the subject written by traders, you will probably learn something good.
     
  9. I played around with 100 period regressions channels years ago, could not find any consistency. There was a guy that wrote a system around them...can't remember his handle...Trader Bamboo...that was it...
     
  10. Lot of truth in Jack's first post, although I must admit I use neither regression lines nor volume to daytrade. I could never find anything useful from either of them in a lot of backtesting, but others may have been more successful.

    Personally, I think daytrading is about getting with the dominant trend, S and R and reversals. I think a simple trend line is more useful than a linear regression line.
     
    #10     Sep 5, 2005