Data required for analysing props

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • Duff85
    SBR MVP
    • 06-15-10
    • 2920

    #1
    Data required for analysing props
    I have begun analysing a number of cricket props, mostly Poisson stuff like total run outs.

    My question is - what is the minimum amount of data you would use to analyse different props? (I realise some would be able to use more data because of longer relevance).

    Is there anything that can help overcome limited sample sizes?
  • RickySteve
    Restricted User
    • 01-31-06
    • 3415

    #2
    Knowing nothing about cricket, I would be surprised if the event you describe is a Poisson process.
    Comment
    • GarbageMan
      SBR Sharp
      • 11-28-10
      • 484

      #3
      on ave, how many attempted run outs are there per game?
      Comment
      • Cicima6709
        SBR MVP
        • 09-12-10
        • 1023

        #4
        I play a ton of player props (MLB and NHL)...i use 2 yrs worth of data for each player and it has been extremeley profitable.

        For players you only really need 2 yrs of data, with more focus on the recent year, and more importantly recent streaks.

        For actual game props based on prior historical stats...thats tough. The answer could be 1 year. Could be 10 years. Obviously ud want to focus more on recent trends, but depending on the prop, you might need years and years of stats.
        Comment
        • FourLengthsClear
          SBR MVP
          • 12-29-10
          • 3808

          #5
          Originally posted by Duff85
          I have begun analysing a number of cricket props, mostly Poisson stuff like total run outs.

          My question is - what is the minimum amount of data you would use to analyse different props? (I realise some would be able to use more data because of longer relevance).

          Is there anything that can help overcome limited sample sizes?
          Your example of Run outs are not Poisson. Statistically they occur at a much higher rate (when talking about one-day or 20/20 matches) when a team is batting second and is behind according to Duckworth-Lewis.

          Limited sample size would be a real problem but there is lots of data out there.
          Comment
          • Duff85
            SBR MVP
            • 06-15-10
            • 2920

            #6
            Originally posted by FourLengthsClear

            Your example of Run outs are not Poisson. Statistically they occur at a much higher rate (when talking about one-day or 20/20 matches) when a team is batting second and is behind according to Duckworth-Lewis.

            Limited sample size would be a real problem but there is lots of data out there.
            Thanks for the input guys. Fourlengths thanks for your replies to a number of my threads over the last couple of days - your advice is always most welcome. Not sure if you value points, but can only send two because i'm non-pro.
            Comment
            SBR Contests
            Collapse
            Top-Rated US Sportsbooks
            Collapse
            Working...