Ways to assess Average Predictive Performance

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • betso
    SBR Rookie
    • 03-18-07
    • 18

    #1
    Ways to assess Average Predictive Performance
    While doing some research to (hopefully) arrive at a personal line-making methodology I realized that I need to know what I am up against, i.e. how well bookmakers can predict the outcome of a game. I would like to know good ways to assess the predictive performance of a wagering strategy. Below I explain one approach which I think makes sense for a higher (season) level analysis, please suggest alternatives that you might think is more appropriate.

    This was explained to me by a Phd friend and I am pretty sure it has a fancy statistical name (like log-likelihood test or something like that). As a performance measure it would be interesting to analyze how far off one is in its prediction of the outcome in terms of probabilities.

    For example, if a bookmaker posts a Money Line odds at +120 for a certain game, it implies that the bookmaker thinks the fair value is a 45% probability that the certain outcome will occur. If the outcome does occur, the amount the bookmaker would be off in its prediction regarding the outcome of the game would then be |0 - 0.45| 0.45. On the other hand if the outcome did not occur the amount would be |1 - 0.45| 0.55. The "|" here represents is an instruction to evaluate the absolute value.

    To assess the predictive performance of multiple games would then a matter of getting the product of the distances which the prediction is off. E.g. if we are looking at two games with MLs posted at +120 (45%) and -110 (52%) respectively and the result of the games is that bets on both odds pay off. The overall prediction of the bookmaker would then be evaluated to
    |0-0.45|*|1-0.52| = 0.216

    Since evaluating many such games would result in a very very tiny number, a trick one could use is to take the logarithm of the expression, giving a more manageable number. In the example above we would then be looking at:

    ln(|0-0.45|*|1-0.52|) = ln(|0-0.45|) + ln(|1-0.52|) = -1.5325

    This is a nice number indeed, but what does it tell us about predictive performance? Not much really. So we need to benchmark the number against the predictive performance of an more "basic" strategy. This could be the equi-probable strategy, i.e. to say that in any given game each team has a 50% probability of winning. It could also be something slightly more sophisticated such as, "if the team is playing home field then its 55% probability to win, otherwise its 45%". Assuming we use the equi-probable approach to benchmark our results above, the resulting number would then be

    ln(|0-0.45|*|1-0.52|)/ln(0.5*0.5) = [ln(|0-0.45|) + ln(|1-0.52|)]/[ln(0.5)+ln(0.5)] = 1.1054.

    This measure will then be portable to any game and any scenario in which one wishes to assess the predictive ability of its method against an opponent, just like what I would like to do against bookmakers.

    As I do not have the opening lines, which are available from sources like www.donbest.com I am interested to know the predictive performance of bookmakers (ball park measure) over the basic equi-probable strategy and possibly also when accounting the home-field advantage. Anyone that has these numbers at hand for a MLB season I would very much appreciate if you could share them here.

    Note that the approach to evaluate the performance above does not say anything about the profitability of your wagering strategy as it only analyzes your performance on an aggregate level. Averages can certainly be misleading and it is really how your prediction against the bookmaker in each game falls out that really counts.
  • RickySteve
    Restricted User
    • 01-31-06
    • 3415

    #2
    mean square error
    Comment
    • Ganchrow
      SBR Hall of Famer
      • 08-28-05
      • 5011

      #3
      I'd have to agree with RickySteve on this one.

      Typically you do a maximum likelihood estimation after you've formulated a model. You do this in order to estimate parameters. You can then use the mean square error (for instance) to judge the goodness-of-fit of your estimators. If you're just looking at opening lines versus actual resultsthen you already have a model (i.e., the opening lines themselves) and there are no parameters to estimate.

      But you're probably getting ahead of yourself, however. I think your first step would probably be to get some data.
      Comment
      • betso
        SBR Rookie
        • 03-18-07
        • 18

        #4
        Originally posted by RickySteve
        mean square error
        Cool, thanks. This makes sense.

        Do you have a ball park measure of the average MSE for bookies over an MLB season?
        Comment
        • betso
          SBR Rookie
          • 03-18-07
          • 18

          #5
          Originally posted by Ganchrow
          ...If you're just looking at opening lines versus actual resultsthen you already have a model (i.e., the opening lines themselves) and there are no parameters to estimate.

          But you're probably getting ahead of yourself, however. I think your first step would probably be to get some data.
          Maybe I could have made myself a bit clearer. I already have a model, but before I go to Don Best and get lines I would like to gauge how "good" my model needs to be in order to be comparable with the predictive ability of the lines.

          What I am thinking is that if the average MSE of the opening lines is x and my model MSE is y (<x) then I'm in good shape.

          Got game data now, so this is where the fun starts.
          Comment
          • Ganchrow
            SBR Hall of Famer
            • 08-28-05
            • 5011

            #6
            Originally posted by betso
            Maybe I could have made myself a bit clearer. I already have a model, but before I go to Don Best and get lines I would like to gauge how "good" my model needs to be in order to be comparable with the predictive ability of the lines.

            What I am thinking is that if the average MSE of the opening lines is x and my model MSE is y (<x) then I'm in good shape.

            Got game data now, so this is where the fun starts.
            You may want to determine if your model is profitable versus opening numbers.
            Comment
            • Arilou
              SBR Sharp
              • 07-16-06
              • 475

              #7
              If you're not looking to bet too big and you are a strong handicapper, by far the best way to bet MLB is to bet into Pinnacle's openers. MSE is what I would try first, but it definitely 'has issues.'
              Comment
              • RickySteve
                Restricted User
                • 01-31-06
                • 3415

                #8
                Originally posted by betso
                What I am thinking is that if the average MSE of the opening lines is x and my model MSE is y (<x) then I'm in good shape.
                Given this, combined with competent risk management, you might as well start shopping for a yacht.
                Comment
                • betso
                  SBR Rookie
                  • 03-18-07
                  • 18

                  #9
                  Originally posted by Ganchrow
                  You may want to determine if your model is profitable versus opening numbers.
                  Yea, thats the way I am planning to go eventually. I just thought comparing MSEs with my model and the opening lines would give me a rough idea on how good my predictions are. But given the comments on this thread I think I need to re-consider this approach .

                  As I guess no one is happy to give me Don Best opening lines straight off I need to go out and scrape them, and I tend to try to avoid boring work like that but maybe thats the only way to go. Its a shame though, since every serious bettor have done the same exact work, scraping lines and game stats to compile to their own data source. I guess thats part of the game.
                  Comment
                  • betso
                    SBR Rookie
                    • 03-18-07
                    • 18

                    #10
                    Originally posted by RickySteve
                    Given this, combined with competent risk management, you might as well start shopping for a yacht.
                    That would be the day. Or a beach villa by the Mediterranean.
                    Comment
                    • Arilou
                      SBR Sharp
                      • 07-16-06
                      • 475

                      #11
                      You don't actually NEED to be better than the openers to beat the openers. Consider how much harder it is to be the one who posts than the one who bets: If the game is, say, -140, then anything I post outside -135 to -145 is going to get me into trouble and anything far away gets me in BIG trouble. Whereas if I nail it, you probably just pass on the game. Actually being better than the openers and yeah, go ahead and shop for that yacht. But all you really need to do is get to a similar level to the openers, and then attack the big differences. When they say -120 and you say -160, you get -125... and if it's really -140, you're golden.
                      Comment
                      • RickySteve
                        Restricted User
                        • 01-31-06
                        • 3415

                        #12
                        Originally posted by Arilou
                        You don't actually NEED to be better than the openers to beat the openers.
                        No.
                        Comment
                        Search
                        Collapse
                        SBR Contests
                        Collapse
                        Top-Rated US Sportsbooks
                        Collapse
                        Working...