Re: 969 SEA +136
When people throw stats around it is nice to know how to use them.
My database has 17995 games in it (current up to yesterdays games) Dogs (the closing lines I used) lost by exactly 1 run 2872 times so 2872/17995 is 16%. Which is right in line with the yearly expectations as well. So it is a pretty consistent probability.
Now my database also give average lines. Which isnt perfect but gives an idea. Average dog line in those 18K games is +136. Ave dog line of the dogs that lost by exactly one run is +138. 2 cents doesnt sound like much but it is a nice jump. But also factring in you would have to lay about -150 to 155 on those dogs to get them on the +1.5 RL would mean quite a different set of returns. Also the favs in those game were only -151. So anything over that you had to pay on the dog (per average) would make taking the dog mathematically 'wrong'.
Looking at it that way really doesnt prove much since it is in a vacuum. But to get theoretical "value" you have to look at all examples and go from there. But as I always say past results dont predict future successes or failures. But if people really want to try and use math in sports betting (other than figuring out their balance) they have to look at all the angles available.
I am not a big believer in worth or value of points or half points or any of that, be it baseball, football or hoops. I just look at the teams, the prices available and go from there. Baseball simply differs because most of the times no points are involved. But when they are, either the -1.5 or the +1.5 RL you have to know how they equate out. Now obviously these numbers also work for favs losing to the -1.5 RL. So you can work backwards (or forwards) depending on what you want to do.
As far as MLs in football (NFL). I say take the ML on any dog of 6 or less that you think will 'cover'. But on games with a -3 line shop around, see where the MLs are skewed, how the vig is on the 3. (-3 -110 and -3 -130 for example) and how that equates to the ML. Because even though the ML is a separate pool it is skewed, and sometimes point spreads are set to have a more accurate ML rather than point spread. Dogs of 6 or less points since 1992 are 1075-1590-3 SU, and 1309-1269-90 ATS. Obviously ATS results can vary, I have one DB where the dogs are 1367-1235-66 So a huge difference. But that one basically shows extremes and getting the best available number. But Su numbers never change.
On those games the average ML was +147 on the dog. So mickey mouse math would be 1075 x 1.47 = 1580. 1580 -1590 = -10. So as close to proofed as you can get. With the ATS you get 1367 x .909 (will use -110) = 1242. 1242 -1235 = 7. So a pretty good proof out there also. But that is also the best case scenario set of numbers.
Either way dogs that cover games in the NFL end up winning SU 77-83% of the time (depending on your ATS results). Once you dip to 2.5 point dogs and less the numbers improve even more, as you go to 376-419-1 SU and 397-392-7 ATS. in those cases only 19 more dogs 'covered' than won SU. Which is easilly outwieghed by the ML you would have gotten betting them on the ML.
Arent a large portion of MLB games won by exactly 1 run?
Doing the math, it gives him an extra 30% chance to push or something like that.?
Seems like a bad bet going from +130 to -150 for just one run.
When people throw stats around it is nice to know how to use them.
My database has 17995 games in it (current up to yesterdays games) Dogs (the closing lines I used) lost by exactly 1 run 2872 times so 2872/17995 is 16%. Which is right in line with the yearly expectations as well. So it is a pretty consistent probability.
Now my database also give average lines. Which isnt perfect but gives an idea. Average dog line in those 18K games is +136. Ave dog line of the dogs that lost by exactly one run is +138. 2 cents doesnt sound like much but it is a nice jump. But also factring in you would have to lay about -150 to 155 on those dogs to get them on the +1.5 RL would mean quite a different set of returns. Also the favs in those game were only -151. So anything over that you had to pay on the dog (per average) would make taking the dog mathematically 'wrong'.
Looking at it that way really doesnt prove much since it is in a vacuum. But to get theoretical "value" you have to look at all examples and go from there. But as I always say past results dont predict future successes or failures. But if people really want to try and use math in sports betting (other than figuring out their balance) they have to look at all the angles available.
I am not a big believer in worth or value of points or half points or any of that, be it baseball, football or hoops. I just look at the teams, the prices available and go from there. Baseball simply differs because most of the times no points are involved. But when they are, either the -1.5 or the +1.5 RL you have to know how they equate out. Now obviously these numbers also work for favs losing to the -1.5 RL. So you can work backwards (or forwards) depending on what you want to do.
As far as MLs in football (NFL). I say take the ML on any dog of 6 or less that you think will 'cover'. But on games with a -3 line shop around, see where the MLs are skewed, how the vig is on the 3. (-3 -110 and -3 -130 for example) and how that equates to the ML. Because even though the ML is a separate pool it is skewed, and sometimes point spreads are set to have a more accurate ML rather than point spread. Dogs of 6 or less points since 1992 are 1075-1590-3 SU, and 1309-1269-90 ATS. Obviously ATS results can vary, I have one DB where the dogs are 1367-1235-66 So a huge difference. But that one basically shows extremes and getting the best available number. But Su numbers never change.
On those games the average ML was +147 on the dog. So mickey mouse math would be 1075 x 1.47 = 1580. 1580 -1590 = -10. So as close to proofed as you can get. With the ATS you get 1367 x .909 (will use -110) = 1242. 1242 -1235 = 7. So a pretty good proof out there also. But that is also the best case scenario set of numbers.
Either way dogs that cover games in the NFL end up winning SU 77-83% of the time (depending on your ATS results). Once you dip to 2.5 point dogs and less the numbers improve even more, as you go to 376-419-1 SU and 397-392-7 ATS. in those cases only 19 more dogs 'covered' than won SU. Which is easilly outwieghed by the ML you would have gotten betting them on the ML.