looking for trends

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • rfr3sh
    SBR Posting Legend
    • 11-07-09
    • 10229

    #1
    looking for trends
    ok lets say you find a trend with lets say around 1000 samples
    the trend shows that the home team ends up winning the game 60% of the time SU
    the last 10 games the home team has won the game SU
    is it ever profitable to take the away team SU ie fade the trend

    i know this is a vague example but just a starter question as I am exploring databases

    also is there such thing a a trend that is to vague or too specific
  • roasthawg
    SBR MVP
    • 11-09-07
    • 2990

    #2
    There is what's known as "regression to the mean" which basically says that if you have a bunch of 50% coin flips land on heads then a bunch more will eventually land on tails in the future bringing the total back close to 50% where it should be. The thing is the games individually are still 50% coin flips and there's no way of knowing when the "regression to the mean" will occur... so fading an angle that's been successful in the past is not a valid form of handicapping.
    Comment
    • rfr3sh
      SBR Posting Legend
      • 11-07-09
      • 10229

      #3
      so is finding trends really that useful?
      Comment
      • bztips
        SBR Sharp
        • 06-03-10
        • 283

        #4
        Originally posted by rfr3sh
        ok lets say you find a trend with lets say around 1000 samples
        the trend shows that the home team ends up winning the game 60% of the time SU
        the last 10 games the home team has won the game SU
        is it ever profitable to take the away team SU ie fade the trend

        i know this is a vague example but just a starter question as I am exploring databases

        also is there such thing a a trend that is to vague or too specific
        Complete waste of time.
        The outcomes of upcoming games are not influenced by what may have happened recently with the trend you have found.
        Comment
        • roasthawg
          SBR MVP
          • 11-09-07
          • 2990

          #5
          Originally posted by rfr3sh
          so is finding trends really that useful?
          Worthless imo.
          Comment
          • rfr3sh
            SBR Posting Legend
            • 11-07-09
            • 10229

            #6
            thanks guys
            Comment
            • brewers7
              SBR Sharp
              • 03-11-06
              • 298

              #7
              Originally posted by rfr3sh
              so is finding trends really that useful?
              I would tend to disagree that they are a "waste of time"...

              If we are talking about one trend, a trend standing alone by itself as your only piece of info to pick a game, then yeah, no matter how the general trend may be, it means next to nothing without "cross-referencing" the trend with several other trends and situational spots...

              To me, if you have 3 or more situational spots (trends) coming together in one game and the track record is strong for this combination of situational spots, then you have something worth looking at...
              Comment
              • brewers7
                SBR Sharp
                • 03-11-06
                • 298

                #8
                Of course, you need to have the data to evaluate these situations...

                Here is a small sample of my database that goes back 19 seasons:

                Sample1
                Comment
                • rfr3sh
                  SBR Posting Legend
                  • 11-07-09
                  • 10229

                  #9
                  i was just playing around with sportsdatabase.com and i found some of the stuff interesting

                  like i know MIZ is 14-7 ATS (+3.9 ppg) since October 11, 1980 as a road favorite after a loss.

                  is irrelevant, but i was thinking maybe on a larger sample size as a whole the trends might be relevant
                  Comment
                  • Flight
                    Restricted User
                    • 01-28-09
                    • 1979

                    #10
                    Some good handicappers use trends.

                    Others don't.

                    I think Dr Bob uses them (as an indicator along with many other things) as well as Big Al and, well, now that I think about it tons of handicappers use them.

                    I think they are rubbish. More specifically I find them to be superstitious, and I reject all superstition in life. I attribute nearly all trends identified by gamblers to be randomness (or rather, underestimating randomness).

                    Read this book, and get back to me on the whole trend thing.



                    (it's not just about markets... I found it more applicable to life in general, philosophy, outlook, etc)
                    Comment
                    • Wrecktangle
                      SBR MVP
                      • 03-01-09
                      • 1524

                      #11
                      What are we defining as a trend here? The long term "trend" Home Dogs works slightly in sides in many sports because the public tends to overweight Away Favs but it may or may not be enough to overcome your vig.
                      Comment
                      • bztips
                        SBR Sharp
                        • 06-03-10
                        • 283

                        #12
                        Originally posted by brewers7
                        I would tend to disagree that they are a "waste of time"...

                        If we are talking about one trend, a trend standing alone by itself as your only piece of info to pick a game, then yeah, no matter how the general trend may be, it means next to nothing without "cross-referencing" the trend with several other trends and situational spots...

                        To me, if you have 3 or more situational spots (trends) coming together in one game and the track record is strong for this combination of situational spots, then you have something worth looking at...
                        Couldn't disagree more.
                        In most cases, all you are doing is data-mining (or what some call "data-dredging"), meaning that if you look long and hard enough you will find lots of different "trends" just by statistical chance. Kicker is that if you were to continue to look further, you could find just as many other trends that support the other side of the bet.

                        But I don't discount Wreck's observation about some very simplistic trends such as home dogs, IF there's a reasonable explanation for it -- and the more important point that the effect is probably quite small, not big enough to actually be profitable net of vig.
                        Comment
                        SBR Contests
                        Collapse
                        Top-Rated US Sportsbooks
                        Collapse
                        Working...