How many bets are needed to consider long term?

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • packers51089
    SBR Hustler
    • 01-22-09
    • 53

    #1
    How many bets are needed to consider long term?
    Ive placed about 600 bets hitting at 52.5%. I am trying to figure out how many bets you need to place to determine your actuall win rate, and what are the chances that I'm a losing capper?
  • G's pks
    Restricted User
    • 01-01-09
    • 22251

    #2
    Originally posted by packers51089
    Ive placed about 600 bets hitting at 52.5%. I am trying to figure out how many bets you need to place to determine your actuall win rate, and what are the chances that I'm a losing capper?
    Quoted from an online site about.com....

    When you bet football in casino sports books you have to pay a percentage on losing bets. This is called the vig. The odds are 11-10 which means if you want to place a bet on a game to win $100. It would cost you $110. If you win you would collect $210. (Your $110 original wager and your $100 winnings.) Now to be a successful handicapper you have to make up for the vig on the losing bets so to breaks even you need to have a win percentage of 52.4 percent. If you can achieve a long time winning percentage of 55 percent it is considered very good. Anything over 60 percent is excellent however most bettors don't sustain this percentage over the long run.
    Comment
    • packers51089
      SBR Hustler
      • 01-22-09
      • 53

      #3
      Originally posted by G's pks
      Quoted from an online site about.com....

      When you bet football in casino sports books you have to pay a percentage on losing bets. This is called the vig. The odds are 11-10 which means if you want to place a bet on a game to win $100. It would cost you $110. If you win you would collect $210. (Your $110 original wager and your $100 winnings.) Now to be a successful handicapper you have to make up for the vig on the losing bets so to breaks even you need to have a win percentage of 52.4 percent. If you can achieve a long time winning percentage of 55 percent it is considered very good. Anything over 60 percent is excellent however most bettors don't sustain this percentage over the long run.
      yes i understand how juice works, just curious if anyone has any math to determine how significant the data is in the long run. If those results are close to representative, then I would make a few adjustments
      Comment
      • G's pks
        Restricted User
        • 01-01-09
        • 22251

        #4
        to break even you need to have a win percentage of 52.4 percent. If you can achieve a long time winning percentage of 55 percent it is considered very good. Anything over 60 percent is excellent however most bettors don't sustain this percentage over the long run.

        I would say this is very close! But remember I am guessing all bets are not even, so amount on each bet, parlays, teasers, pleasers, all that good stuff will have an effect. I am a sucker for the props...but when I win one I feel good!

        I would say effective money management is just as..if not more important. Knowing when to pull back. I just went thru a 3-10 slump, started betting smaller, then just went 5-2 with a nice superbowl win.

        The following copied article kind of reinforces the 50's % area for success... I am thinking 54% accuracy is a closer amount.


        Sports Betting 201 - How to Profit


        This article attempts to explain of how to profit at sports betting. This is a starting point, but far from the ending… There is a lot to learn that is not covered by this lesson. But if you stay focused and keep learning, and you could end up a winner.

        The first thing to know (and follow) is never to bet over your head (or more than you can afford to lose). Money management is cover elsewhere, but just like every form of gambling, betting your rent is too much to risk and it will come back to haunt you. By doing it once, it's easier to do it twice, and you will eventually get burned.

        Following that train of thought, the most important point for new sports bettors to know is to never chase losses. Chasing losses is the act of increasing your bet size when you are on a losing streak in order to try to break even. Doing this causes many sports bettors to go bust before they can even get going and often leads people into betting over their head. See the trend?

        Examples of chasing loses happens every weekend during football season. Every weekend, some guy bets 3 football games and loses all three. Let's say he bets $500 on each which is a decent amount of money for him. This same guy will then drop $1,500 on the money night game trying to get back to even. Half the time, this guy will lose and be out $3,000 instead of just $1,500. Avoiding this pitfall will save you a lot of stress and money.

        Now that we mentioned losing, lets talk about winning. In order to win money in sports betting (assuming that your bets are $110 to win $100), you will have to win 52.4% of your bets. For you math geniuses out there, it is calculated as 110/(110+100). This is the win rate needed when you bet on football sides and over/unders.

        These odds vary between sports and even games. If for example you bet just baseball, the percentage of games that you will need to win in order to profit will vary depending on the odds that you are betting. If you bet mostly underdogs, you will need a lesser win percentage than if you bet most favorites. Because of this, it's possible to be a long-term winner while winning less than 50% of your bets by betting underdogs.

        A great public misconception is that professional sports bettors win between 60% to 70% of their bets. This is just false. Professional sports bettors lose almost as much as they win, but because of the volume and size of their bets, they can make a nice living cashing anywhere between 54% to 57% of their tickets.

        So take it slow, protect your money and don't expect to win every game. Hot and cold streak happen to everyone, but by staying focused and perfecting your techniques, you can become a winner in the long run.
        Comment
        • Ganchrow
          SBR Hall of Famer
          • 08-28-05
          • 5011

          #5
          Originally posted by packers51089
          Ive placed about 600 bets hitting at 52.5%. I am trying to figure out how many bets you need to place to determine your actuall win rate, and what are the chances that I'm a losing capper?
          You night want to check out this thread: http://forum.sbrforum.com/players-ta...tml#post859159
          Comment
          • u21c3f6
            SBR Wise Guy
            • 01-17-09
            • 790

            #6
            There appears to be several posters to these boards that can give you very good answers. I will describe how I look at it for whatever it is worth.

            I assume these are 50/50 picks. If so, then the expected # of winners in 600 bets would be 300. At 52.5% you won 315. I look for something that would be more than 2.5 SD's (or better yet >3 SD's).

            I calculate SD on 50/50 picks as follows:
            SQRT(600*.525*.475)=12.23

            Your results are only a little more than 1 SD and therefore not conlusive IMO. It would take approx 1600 bets at 52.5% to be > 2 SD's and approx 2500 bets to be > than 2.5 SD's.
            Comment
            • smokin1
              SBR Rookie
              • 11-06-08
              • 38

              #7
              Originally posted by u21c3f6
              There appears to be several posters to these boards that can give you very good answers. I will describe how I look at it for whatever it is worth.

              I assume these are 50/50 picks. If so, then the expected # of winners in 600 bets would be 300. At 52.5% you won 315. I look for something that would be more than 2.5 SD's (or better yet >3 SD's).

              I calculate SD on 50/50 picks as follows:
              SQRT(600*.525*.475)=12.23

              Your results are only a little more than 1 SD and therefore not conlusive IMO. It would take approx 1600 bets at 52.5% to be > 2 SD's and approx 2500 bets to be > than 2.5 SD's.

              Excuse my ignorance, how does SQRT(600*.525*.475)=12.23 = little more than 1 SD

              Could you explain?
              Comment
              • u21c3f6
                SBR Wise Guy
                • 01-17-09
                • 790

                #8
                300 expected winners +/- 1 SD of 12.23 = 287.77 to 312.23 range of expected winners within 1 SD.

                He had 315 winners, a little higher than the upper limit for 1 SD.
                Comment
                • smokin1
                  SBR Rookie
                  • 11-06-08
                  • 38

                  #9
                  gotcha, thx
                  Comment
                  • Dark Horse
                    SBR Posting Legend
                    • 12-14-05
                    • 13764

                    #10
                    At least two standard deviations, but more with lower sample sizes.

                    (Wins minus losses)/square root of sample size >2


                    I would also categorize by sport and league, just because most players I know of will only excel at one or two sports.

                    Beyond numbers you probably need five years of experience. Just to make sure you can handle the emotional challenges. A lot of players start chasing when they lose, without necessarily realizing it (by lowering trigger points to place bets, because they want to beat a spell of bad luck).
                    Last edited by Dark Horse; 02-02-09, 05:20 PM.
                    Comment
                    • Ganchrow
                      SBR Hall of Famer
                      • 08-28-05
                      • 5011

                      #11
                      Originally posted by u21c3f6
                      There appears to be several posters to these boards that can give you very good answers. I will describe how I look at it for whatever it is worth.

                      I assume these are 50/50 picks. If so, then the expected # of winners in 600 bets would be 300. At 52.5% you won 315. I look for something that would be more than 2.5 SD's (or better yet >3 SD's).

                      I calculate SD on 50/50 picks as follows:
                      SQRT(600*.525*.475)=12.23

                      Your results are only a little more than 1 SD and therefore not conlusive IMO. It would take approx 1600 bets at 52.5% to be > 2 SD's and approx 2500 bets to be > than 2.5 SD's.
                      Actually, this isn't quite right. The issue isn't whether the the player's handicapping is superior to a coin flip, but rather whether or not the player were profitable.

                      If your null hypothesis were that the player were not a profitable bettor at -110 (break even rate of 11/21), then the relevant test statistic would be:
                      Z = ( 52.5% * 600 - 11/21 * 600 ) / sqrt( 11/21 * (1-11/21) * 600 )
                      ≈ 0.05839
                      for a p-value of 1-NORMSDIST(0.05839) ≈ 47.67%.

                      This means that a break even bettor would have a roughly 47.67% probability of hitting 52.5% or better over 600 wagers simply by chance. (The "true" p-value from the binomial distribution would be =1-BINOMDIST(.525*600-1,600,11/21,1) ≈ 49.33%.)

                      Even if the player were betting at odds of +100 (break even of 50%), the above test statistic you provided would still be slightly to large that one would need to calculate the standard deviation based on the parameter test value rather than the experimentally realized value.

                      With that in mind, the test statistic for a null of no profitability at +100 would be would be
                      Z = ( 52.5% * 600 - 50% * 600 ) / sqrt( 50% * 50% * 600 )
                      ≈ 1.225
                      for a p-value of 1-NORMSDIST(1.225) ≈ 11.03%. (Actual p-value of =1-BINOMDIST(.525*600-1,600,50%,1) ≈ 11.82%.)

                      I'll also point that the OP's question "what are the chances that I'm a losing capper?" can't really be answered by a hypothesis test alone, but rather would require utilizing Bayesian inference. I gave an example of this in this post.

                      Originally posted by Dark Horse
                      I would also categorize by sport and league, just because most players I know of will only excel at one or two sports.
                      It pays to mention that by proceeding categorically, p-values obtained in the traditional manner will no longer be accurate. For example, a single category p-value of 2.275% (2 standard deviations, single-tailed) would be obtained with probability 1-(1-2.275%)^4 ≈ 8.794% over 4 categories.

                      To obtain p-value of 2.275% over 4 categories, instead of 2 standard deviations one would require a single category p-value of 1-(1-2.275%)^0.25 ≈ 0.574%, or about 2.528 standard deviations.
                      Comment
                      • u21c3f6
                        SBR Wise Guy
                        • 01-17-09
                        • 790

                        #12
                        Originally posted by Ganchrow
                        Actually, this isn't quite right. The issue isn't whether the the player's handicapping is superior to a coin flip, but rather whether or not the player were profitable.

                        My focus was on this part of his question: "Ive placed about 600 bets hitting at 52.5%. I am trying to figure out how many bets you need to place to determine your actuall win rate"

                        Whether or not he would be a winning or losing capper is another question. Right now the way I look at what I think he is asking, 600 bets on 50/50 propositions and hitting 315 (52.5%) while good, is not enough trials for me to say that this result is not something that could be produced by chance. However, if he bets 1,600 times and hits 840 (52.5%), now we are getting into the range where we become confidant to say that this result is less likely to be the result of chance and probably due to some "edge" being applied by the bettor. We still wouldn't necessarily know for sure if his true win rate is 52.5% but we would be a lot surer that it was at least close to that # based on the # of trials.

                        Again, this is how I look at it and exactly how I test my results. Your results may differ.
                        Comment
                        • Dark Horse
                          SBR Posting Legend
                          • 12-14-05
                          • 13764

                          #13
                          Originally posted by Ganchrow
                          It pays to mention that by proceeding categorically, p-values obtained in the traditional manner will no longer be accurate. For example, a single category p-value of 2.275% (2 standard deviations, single-tailed) would be obtained with probability 1-(1-2.275%)^4 ≈ 8.794% over 4 categories.

                          To obtain p-value of 2.275% over 4 categories, instead of 2 standard deviations one would require a single category p-value of 1-(1-2.275%)^0.25 ≈ 0.574%, or about 2.528 standard deviations.
                          Or you could just play the categories you're good at.
                          Comment
                          • SlowRoll
                            SBR High Roller
                            • 02-04-09
                            • 203

                            #14
                            600 pretty close
                            Comment
                            SBR Contests
                            Collapse
                            Top-Rated US Sportsbooks
                            Collapse
                            Working...