Luck vs Research... Luck is winning (for now)..

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • Shoomakie
    SBR Hustler
    • 11-03-12
    • 61

    #1
    Luck vs Research... Luck is winning (for now)..
    Good evening all,

    So after a long time away from the craft, i decided to dip my toe back in the water. I've never been one for research. I've always went with the gut plus "common sense" tactic (ie team X should really win even if they don't physically show up). Iv'e done well...not great. This year I decided that I will do as much research as possible on NBA over/under and go with the statistical/numerical favorite. In order to be considered a selection the over/under has to be greater then 3 based on a few different criteria.

    I'm not sure if I am allowed to post my criteria/formulas so for now I will not. The issue I am having is that using my gut/common sense I am 11-5. Using statistics I am 4-12. I don't want to play long term on gut/common sense. I'd prefer to have some numbers to blame when I don't win ?

    Can someone point me towards how to research NBA over/under? I'd like to compare with what I'm doing.

    Thanks!
  • u21c3f6
    SBR Wise Guy
    • 01-17-09
    • 790

    #2
    I hope I word this correctly so that it is understood.

    Your statistics may actually be correct, you are just looking at it incorrectly. Most players I believe are trying to pick the winner. I do not try to pick the winner, I am more concerned with finding selections that may be over or under bet and then wager accordingly. Your statistics may be signaling the selections that are over bet (probably because you are using the most common statistics) and therefore the value is wagering on the other side.

    Joe.
    Comment
    • Shoomakie
      SBR Hustler
      • 11-03-12
      • 61

      #3
      Joe,

      Thanks for the reply. Let me clarify.. I am looking at the totals for over/under picks. I think (and I could be wrong) that you are talking more about which side has more action. Here is an example of how I derive my selection using numbers.

      Example #1: Toronto/Philly O/U Line 182.5

      I use 3 sets of numbers:

      Group One consists of the average Points For (PF) & Points Against (PA) for both teams: Toronto 96.10/99.80 Philly 98/100.89. These average out to 197.39 (+14.39 Over)

      Group Two consists of the total PF for both teams: Toronto 96.10 Philly 98. These total out to 194.10 (+11.60 Over)

      Group Three consists of the total PA for both teams: Toronto 99.80 Philly 100.89. These total out to 200.69 (+18.19 Over)

      In this case all three groups return a number higher then the over which means I would play the over.

      In the case of the other two games being played tonight, one is 2 Over & 1 Under and the other 1 Over & 2 Under, I wouldn't select either.

      What else needs to be taken into consideration to firm up the analysis?

      I was thinking about possibly adding PF/PA Home & Away and perhaps number of days on the road or at home..

      Thoughts?
      Comment
      • CHUBNUT
        SBR Sharp
        • 06-30-09
        • 321

        #4
        Without getting long winded you need to fit that any numbers with the league average to be very basic. Also your figures will not take into account recent information, 2 star players may have got food poisoning at lunch which would wipe 10 off the total no problem. Bottom line is no matter how good you think your statistical analysis is, if you dont take into account latest real world information your betting is based on sheer luck.
        Comment
        • u21c3f6
          SBR Wise Guy
          • 01-17-09
          • 790

          #5
          Shoomakie, I wasn't referring to how much was wagered on each side. Let me try to clarify. This can be long-winded but I am going to try to be as brief as possible. I assume the criteria (stats) that you posted above is what went 4-12 which is exactly my point.

          You are using basic stats available to everyone yet the line is quite far from your average. Therefore there must be something(s) very significant missing from the average. But because the average is high, the missing something(s) may not be factored in enough. Only based on what you wrote above, my inclination would be to play the under and not the over. This does not mean that I believe the under will win, it only means that I believe the value to be on the under and not the over in this scenario. Obviously additional research and forward testing will need to be done.

          Joe.
          Comment
          • Shoomakie
            SBR Hustler
            • 11-03-12
            • 61

            #6
            Originally posted by CHUBNUT
            Without getting long winded you need to fit that any numbers with the league average to be very basic. Also your figures will not take into account recent information, 2 star players may have got food poisoning at lunch which would wipe 10 off the total no problem. Bottom line is no matter how good you think your statistical analysis is, if you don't take into account latest real world information your betting is based on sheer luck.
            I'm assuming that the line already takes into consideration the real world info such as Dirk being out for example. It doesnt take into example Harden getting dehydrated mid game since no one can predict what will happen to a player mid game. My basis for this assumption is the book is trying to be as close to 50/50 as they can get and live off the vig. Would you agree?
            Comment
            • Dark Horse
              SBR Posting Legend
              • 12-14-05
              • 13764

              #7
              Originally posted by Shoomakie
              Joe,

              Thanks for the reply. Let me clarify.. I am looking at the totals for over/under picks. I think (and I could be wrong) that you are talking more about which side has more action. Here is an example of how I derive my selection using numbers.

              Example #1: Toronto/Philly O/U Line 182.5

              I use 3 sets of numbers:

              Group One consists of the average Points For (PF) & Points Against (PA) for both teams: Toronto 96.10/99.80 Philly 98/100.89. These average out to 197.39 (+14.39 Over)

              Group Two consists of the total PF for both teams: Toronto 96.10 Philly 98. These total out to 194.10 (+11.60 Over)

              Group Three consists of the total PA for both teams: Toronto 99.80 Philly 100.89. These total out to 200.69 (+18.19 Over)

              In this case all three groups return a number higher then the over which means I would play the over.

              In the case of the other two games being played tonight, one is 2 Over & 1 Under and the other 1 Over & 2 Under, I wouldn't select either.

              What else needs to be taken into consideration to firm up the analysis?

              I was thinking about possibly adding PF/PA Home & Away and perhaps number of days on the road or at home..

              Thoughts?

              You're the guy that wants to build a house, carries one brick, and then asks others to help him out.

              Stick with the lazy chair of the old days, aka going with your guts.
              Comment
              • Shoomakie
                SBR Hustler
                • 11-03-12
                • 61

                #8
                Originally posted by Dark Horse
                You're the guy that wants to build a house, carries one brick, and then asks others to help him out.

                Stick with the lazy chair of the old days, aka going with your guts.
                Dark Horse,

                I'm not sure whats up with the negative attitude. I thought this was called "Think Tank"?? I came to the table with something I am working with and trying to improve upon it. Did I come in here with nothing and expect everything? Did I disrespect someone, are you just having a bad day?? If its the former, please point it out, if its the latter and you don't have anything constructive to add then just move along.
                Comment
                • Dark Horse
                  SBR Posting Legend
                  • 12-14-05
                  • 13764

                  #9
                  You took it as negative. I would have taken it as positive, and encouragement to go get my second brick.
                  Comment
                  • Shoomakie
                    SBR Hustler
                    • 11-03-12
                    • 61

                    #10
                    Oh, is that how you add help people? Who are you to come onto a thread and assume that I've done nothing or is that how you became such a big shot in your own mind? Have you been keeping track of the hours I spend researching or the spreadsheets I build? Guys like you are a dime a dozen....troll boards, post comments to build up their post count and add absolute nothing to society. Look, its obvious you are too good for my little thread so please try not to post again. Save the bandwidth for someone who might have something positive to add.
                    Comment
                    • Dark Horse
                      SBR Posting Legend
                      • 12-14-05
                      • 13764

                      #11
                      Everything you say confirms my initial evaluation. You don't know what it takes, and aren't willing to do the work. Period.
                      Comment
                      • Inspirited
                        SBR MVP
                        • 06-26-10
                        • 1788

                        #12
                        maybe injuries, how the teams do on back to backs, home/road, travel distance etc...

                        i guess your approach is a "stats approach," but you don't have any weights on your variables.

                        you're gonna find short term trends and probably nothing more doing things like this.

                        i kind of did something like this a couple seasons ago. i got a sample size and then found a trend and then that trend vanished cause it was only just a trend, and i didn't get off the trend and that was no good.

                        i actually kicked @$$ afterwards by going by "gut". i just watched the scores every night. i could see the ebbs and flows of the games and certain teams came to focus for short periods of time. i was at like 58% during that stretch.

                        your gut is more flexible then this "stats" approach, but a genuine stats approach gives you more ease because you trust in its architecture. even if it's going bad you trust that in the end it's gonna hit 54% or something. with your gut you never know if your abilities are good enough.

                        in the end you have to learn how to do regressions as a starting place. keep in mind i don't even know what i'm talking about. just my experience with the two approaches you are using.
                        Comment
                        • CHUBNUT
                          SBR Sharp
                          • 06-30-09
                          • 321

                          #13
                          Originally posted by Shoomakie
                          I'm assuming that the line already takes into consideration the real world info such as Dirk being out for example. It doesnt take into example Harden getting dehydrated mid game since no one can predict what will happen to a player mid game. My basis for this assumption is the book is trying to be as close to 50/50 as they can get and live off the vig. Would you agree?
                          bookmakers lines might or might not contain certain info, thats what you have to work out. The stats you are using are very basic and by using the same ones on 3 different approaches will dilute it into a very square and insignificant number which really is meaningless. You need to dig deeper into subsets of stats especially in Soccer where each league is unique in its traits.
                          Comment
                          • hutennis
                            SBR Wise Guy
                            • 07-11-10
                            • 847

                            #14
                            Originally posted by Shoomakie
                            Oh, is that how you add help people? Who are you to come onto a thread and assume that I've done nothing or is that how you became such a big shot in your own mind? Have you been keeping track of the hours I spend researching or the spreadsheets I build? Guys like you are a dime a dozen....troll boards, post comments to build up their post count and add absolute nothing to society. Look, its obvious you are too good for my little thread so please try not to post again. Save the bandwidth for someone who might have something positive to add.
                            Look.

                            When you come to a public forum to get some information there are two basic possibilities on what you are going to hear back.
                            Either you will hear something you like or you will hear something you don't like.

                            Chances are overwhelming, that information you are going to like is going to be useless and will get you nowhere, costing you some money in a process. On another hand, information you don't like that makes you post some angry replies will be an information you actually should be welcoming, believe it or not. Why? Because if you put your feelings aside and think about it you'll understand that this unpleasant truth is actually there to save you a lot time, effort and money.

                            Here are my unpleasant 2c.

                            Your understanding of what a research is is medieval.
                            In fact it has nothing to do with a research at all.
                            It is basically a form of coin flipping and inevitably will produce the same results.

                            True, expected to produce an edge research is an extremely sophisticated process that involves cutting edge science, takes years and
                            costs millions.
                            To expect this kind of methodology to be revealed in a few posts on a public forum is unimaginably naive.
                            Please, use some common sense, will you?
                            Last edited by hutennis; 11-21-12, 02:54 PM.
                            Comment
                            • Sawyer
                              SBR Hall of Famer
                              • 06-01-09
                              • 7720

                              #15
                              Shoomakie,

                              You need a bigger sample size in order to make an analysis/eveluation. Say..like 250-300 NBA games. Maybe more..
                              Comment
                              • Shoomakie
                                SBR Hustler
                                • 11-03-12
                                • 61

                                #16
                                Originally posted by Inspirited
                                maybe injuries, how the teams do on back to backs, home/road, travel distance etc...

                                i guess your approach is a "stats approach," but you don't have any weights on your variables.

                                you're gonna find short term trends and probably nothing more doing things like this.

                                i kind of did something like this a couple seasons ago. i got a sample size and then found a trend and then that trend vanished cause it was only just a trend, and i didn't get off the trend and that was no good.

                                i actually kicked @$$ afterwards by going by "gut". i just watched the scores every night. i could see the ebbs and flows of the games and certain teams came to focus for short periods of time. i was at like 58% during that stretch.

                                your gut is more flexible then this "stats" approach, but a genuine stats approach gives you more ease because you trust in its architecture. even if it's going bad you trust that in the end it's gonna hit 54% or something. with your gut you never know if your abilities are good enough.

                                in the end you have to learn how to do regressions as a starting place. keep in mind i don't even know what i'm talking about. just my experience with the two approaches you are using.
                                Thanks for the response...

                                To start it was purely a stats approach. I tend to go with the "numbers don't lie" train of thought. I also assumed that the lines makers are trying to get as close to the total as possible. Factoring in injuries is simple enough.

                                How would I apply travel distance and X games in X days to a stats approach?
                                Comment
                                • Shoomakie
                                  SBR Hustler
                                  • 11-03-12
                                  • 61

                                  #17
                                  Originally posted by hutennis
                                  Look.

                                  When you come to a public forum to get some information there are two basic possibilities on what you are going to hear back.
                                  Either you will hear something you like or you will hear something you don't like.

                                  Chances are overwhelming, that information you are going to like is going to be useless and will get you nowhere, costing you some money in a process. On another hand, information you don't like that makes you post some angry replies will be an information you actually should be welcoming, believe it or not. Why? Because if you put your feelings aside and think about it you'll understand that this unpleasant truth is actually there to save you a lot time, effort and money.
                                  hutennis,

                                  Firstly, I've read many of your posts in the think tank and find them to be loaded with information so thank you for that.

                                  I totally agree with EVERYTHING you are saying. My issue with his post was that he was "accusing" me of not wanting to do any work. If someone sees a thread that interests them and they come across a post stating something to the fact that the poster is looking for a handout, the person who might have some good info might just pass on posting. I'm not asking for people to give me picks or share there own data. I'm putting out what I am doing, knowing that it's not good, and asking for feedback. If I was just a person who wanted to "like what I was going to hear" I would take offense and respond accordingly to your "understanding of what a research is is medieval....." comment, but it is constructive and forces me to reevaluate what data I am collecting and how I am utilizing it.

                                  Originally posted by hutennis
                                  Here are my unpleasant 2c.

                                  Your understanding of what a research is is medieval.
                                  In fact it has nothing to do with a research at all.
                                  It is basically a form of coin flipping and inevitably will produce the same results.
                                  Again I totally agree which is why I took the time to post here. I can troll around looking for the "Free Picks" and other "Locks of the Day" but I want to put something together that I understand and can eventually share with others. I am sure I am not the only one person who wants to mine their own data, I'm just the stupid one who is willing to try and improve it with the help of others.

                                  Originally posted by hutennis
                                  True, expected to produce an edge research is an extremely sophisticated process that involves cutting edge science, takes years and
                                  costs millions.
                                  To expect this kind of methodology to be revealed in a few posts on a public forum is unimaginably naive.
                                  Please, use some common sense, will you?
                                  If this was a career for me I would throw in the towel however its something I will continue to work on and see how it goes.

                                  Again I appreciate you taking the time to respond!
                                  Comment
                                  • Shoomakie
                                    SBR Hustler
                                    • 11-03-12
                                    • 61

                                    #18
                                    Originally posted by Sawyer
                                    Shoomakie,

                                    You need a bigger sample size in order to make an analysis/eveluation. Say..like 250-300 NBA games. Maybe more..
                                    Agreed... I brought in data from 09/10 onwards on a separate sheet and Im going to combine this data with with this years to see how it changes.

                                    I feel anything more then 3 years is stale data since personnel have changed/aged.
                                    Comment
                                    • hutennis
                                      SBR Wise Guy
                                      • 07-11-10
                                      • 847

                                      #19
                                      Originally posted by Shoomakie
                                      hutennis,

                                      If this was a career for me I would throw in the towel however its something I will continue to work on and see how it goes.

                                      Again I appreciate you taking the time to respond!
                                      As long as you enjoy what you are doing and are going into it with open eyes it is your call all the away.

                                      Good luck.
                                      Comment
                                      • Dark Horse
                                        SBR Posting Legend
                                        • 12-14-05
                                        • 13764

                                        #20
                                        Shoo,

                                        I was challenging you to come up with some original creative insight, instead of with something extremely basic, present that as 'research', and then lazily ask others for their input. Your quick reliance on others being the main error. You now present that as some sort of social approach to a research project. Apparently, your idea of social involvement is to come up with something that went 4-12. For real? Instead of critiquing yourself, as we all must in this field, and as would have been so easy for you in this instance, you leave that to others as well. And then you act indignant. Your whole approach is completely unrealistic. You have to be your own toughest critic to have any chance at all, and if you were, and still were socially inclined, you would perhaps ask others to participate in something that had gone 12-4 (preferably 60-20) rather than 4-12. I'm not putting you down, by the way. There's nothing personal about this. The crappy type of advice would be to tell you to keep doing what you're doing and fade yourself. I didn't go there, did I?
                                        Last edited by Dark Horse; 11-22-12, 01:23 AM.
                                        Comment
                                        • Inspirited
                                          SBR MVP
                                          • 06-26-10
                                          • 1788

                                          #21
                                          read this

                                          Comment
                                          • chunk
                                            SBR Wise Guy
                                            • 02-08-11
                                            • 808

                                            #22
                                            Originally posted by Shoomakie
                                            Agreed... I brought in data from 09/10 onwards on a separate sheet and Im going to combine this data with with this years to see how it changes.

                                            I feel anything more then 3 years is stale data since personnel have changed/aged.
                                            This may or may not be true. It depends on how you use the data. More is better in many cases.
                                            Comment
                                            • indio
                                              SBR Wise Guy
                                              • 06-03-11
                                              • 751

                                              #23
                                              Originally posted by Shoomakie
                                              Joe,

                                              Thanks for the reply. Let me clarify.. I am looking at the totals for over/under picks. I think (and I could be wrong) that you are talking more about which side has more action. Here is an example of how I derive my selection using numbers.

                                              Example #1: Toronto/Philly O/U Line 182.5

                                              I use 3 sets of numbers:

                                              Group One consists of the average Points For (PF) & Points Against (PA) for both teams: Toronto 96.10/99.80 Philly 98/100.89. These average out to 197.39 (+14.39 Over)

                                              Group Two consists of the total PF for both teams: Toronto 96.10 Philly 98. These total out to 194.10 (+11.60 Over)

                                              Group Three consists of the total PA for both teams: Toronto 99.80 Philly 100.89. These total out to 200.69 (+18.19 Over)

                                              In this case all three groups return a number higher then the over which means I would play the over.

                                              In the case of the other two games being played tonight, one is 2 Over & 1 Under and the other 1 Over & 2 Under, I wouldn't select either.

                                              What else needs to be taken into consideration to firm up the analysis?

                                              I was thinking about possibly adding PF/PA Home & Away and perhaps number of days on the road or at home..

                                              Thoughts?
                                              No disrespect intended, but you need to wake up to reality. Basketball linemakers are good at putting out a generic total, then, while the limits are low, they follow wagering indicators by certain customers they rely on to adjust the line, and then they open up max bets to 20k.

                                              Now, do you really think you're going to have some sort of long term profitable formula by simply relying on points for and points against averages of the teams? Think about it.

                                              Good NBA basketball bettors watch more basketball than you could possibly imagine. They know the players almost as well as pro scouts. They know who's been on an extended road trip and tired, who's healthy, who's missing key reserves, etc..etc.., and they understand game dynamics, flow, and variables you never thought of, and even with all this, they are usually hitting lines early and only when they have a slight edge, and all of this is to hopefully hit around 55-57% of their picks if they don't get too unlucky. And the time they spend doing this is equal to that of a lab scientist.

                                              If you like betting totals, stick with your gut instincts, play some games, have some fun, and hope you get lucky, but please don't think you're going to beat these guys using a simple formula based on points for and points against averages, because it wont happen with a large enough sample to restrict the effect of variance.
                                              Comment
                                              • allin1
                                                SBR MVP
                                                • 11-07-11
                                                • 4555

                                                #24
                                                Originally posted by indio
                                                No disrespect intended, but you need to wake up to reality. Basketball linemakers are good at putting out a generic total, then, while the limits are low, they follow wagering indicators by certain customers they rely on to adjust the line, and then they open up max bets to 20k.

                                                Now, do you really think you're going to have some sort of long term profitable formula by simply relying on points for and points against averages of the teams? Think about it.

                                                Good NBA basketball bettors watch more basketball than you could possibly imagine. They know the players almost as well as pro scouts. They know who's been on an extended road trip and tired, who's healthy, who's missing key reserves, etc..etc.., and they understand game dynamics, flow, and variables you never thought of, and even with all this, they are usually hitting lines early and only when they have a slight edge, and all of this is to hopefully hit around 55-57% of their picks if they don't get too unlucky. And the time they spend doing this is equal to that of a lab scientist.

                                                If you like betting totals, stick with your gut instincts, play some games, have some fun, and hope you get lucky, but please don't think you're going to beat these guys using a simple formula based on points for and points against averages, because it wont happen with a large enough sample to restrict the effect of variance.
                                                if 98% of sports bettors would read and understand your post... it would probably diminish the profits bookmakers make
                                                Comment
                                                • InTheRed
                                                  SBR Sharp
                                                  • 12-25-09
                                                  • 455

                                                  #25
                                                  Since no one has mentioned it yet. I'd go back to your example and think it through. You're using the same number different ways and essentially coming up with the same results explained differently.

                                                  You're using the team avg for PF and PA. You're averaging the 4 numbers together. Get a number. Then you're splitting it in to two groups and adding together. But its still the same numbers. Go one further. Take your PF total and your PA total and average them, and you get the original average number.

                                                  Your system may not be working because you're just manipulating the same result several ways.


                                                  As you said, you want to consider Home and Away. Good. Those are different numbers. It'll explain things more. Going further, you can add H2H numbers, numbers from days rest, player avg, etc.

                                                  The key is that the numbers are different and can help to explain better. Getting the numbers are easy. So is the math. The key is what to do with them and how to read it. That is where a good system comes from.

                                                  Just using PF/PA avg. different ways is overcomplicating something simple and making up indicators that aren't indicating anything relevant.
                                                  Comment
                                                  • VLR100
                                                    SBR High Roller
                                                    • 01-10-10
                                                    • 217

                                                    #26
                                                    Originally posted by allin1
                                                    if 98% of sports bettors would read and understand your post... it would probably diminish the profits bookmakers make
                                                    Let's hope not, rational investors are necessary for rational investors to profit.

                                                    Shoomakie, you may want to look at pace as a significant explanatory variable with regard to NBA totals.
                                                    Comment
                                                    • SortsofSports
                                                      SBR Rookie
                                                      • 11-14-10
                                                      • 36

                                                      #27
                                                      I can look back on my modeling attempts and take a few things away. I was extremely lucky when I began looking into math for sports betting. I built a database, and a program that has been crushing the NHL for 3 years now. That being said, I've spent those same 3 years unsuccessfully building an NBA model. Through dozens, if not hundreds of iterations, I still have a statistical pile of garbage, and have blown most of my NHL profits testing something I thought was statistically significant without actually going through the process of checking it.

                                                      Forward Looking:
                                                      Everyone can calculate averages, even some of the most unintelligent people wagering on sports. This alone will not find you an edge. This may be a good starting point but it will only be that.

                                                      NBA Totals are one of the hardest lines in professional sports to beat. Why not start with hockey, college basketball, or another market where lines are softer, and you can expect to find an advantage?

                                                      3 years, and a couple hundred hours of working, and I have nothing to show for it. I will keep throwing my head against the wall looking for what works. Don't give up, but start to expand on what variables you are looking at.

                                                      Don't simply try to justify why your numbers, model, database should work. Prove it. After you are done proving it, re-prove it. When you can do that, you will make money. Keep a close eye on performance to make sure the market hasn't adjusted. Then keep proving it.

                                                      If you can write computer programs, automate everything you possibly can in building a database, and making calculations. It frees up your time, for testing, checking, and tweaking the model.

                                                      If you aren't familiar with statistics, take some time and learn what you can. At least get to the point where Regression Analysis is part of your mathematical arsenal.
                                                      Comment
                                                      • GunShard
                                                        SBR Posting Legend
                                                        • 03-05-10
                                                        • 10031

                                                        #28
                                                        Everything is a factor. Find out what is a major factor and what is a minor factor.

                                                        1. What happens in real life, such as injuries and missing key players is a more important factor than statstics and history.

                                                        2. Betting on a sport that's more predictable compared to another sport that's less predictable.

                                                        3. Money management: You should be betting less than 5% of your total bankroll per bet.

                                                        4. Never bet because it's your favorite team or betting for action. Be unbiased and be disipline.
                                                        Comment
                                                        SBR Contests
                                                        Collapse
                                                        Top-Rated US Sportsbooks
                                                        Collapse
                                                        Working...