Author

Topic: Grouping lines of odds by best value, for sports betting and dice (Read 102 times)

legendary
Activity: 1568
Merit: 6660
bitcoincleanup.com / bitmixlist.org
I was trying to construct a Gaussian distribution of betting odds between 1.0 and 100, except the difference is that I'm not grouping odds by probability (which wouldn't make sense as it's dead simple to see), I am trying to group them by best value, or in lines.

After arallmuus wrote this on the UEFA thread:

been there done that  Wink rookie mistakes tho...hopefully you've learned to avoid those no-value lines, next step will be to start searching for the good value lines at 1.5x or higher. I usually look for 1.70 or higher, if i go for anything lower is prolly cos it's the best line i could find in the event i'm betting on

True, I'd stick with 1.7-2.4 odds because that is probably the best value.

I've been trying to make a distribution out of this to guess the other lines. I'll need two distributions for each side of the mean since one side is going to estimate between 1.0-2.0 and the other side is going to estimate 2.0-100.0.

So using mean=2, std=0.4 for the right side, I'll get lines between 2, 2.4, 2.8, 3.2, ...etc. And to get the standard deviation for the left side I will subtract each pair of stds reciprocals. i.e. the corresponding line for 2-2.4 will be MEAN-(1/(MEAN-1) - 1/(2.4-1)) = 1.71 to 2.

Surprisingly arallmuus' estimation was very accurate. So let's make further lines that have less value. Right std will be constant, left std will keep changing. You'll be surprised how small the left lines are compared to the right lines.

So line 2.4-2.8 has a corresponding left line of 2-(1/(2-1) - 1/(2.8-1)) = 1.56 to 1.71.

2.8-3.2 corresponds to 1.45-1.56
3.2-3.6 corresponds to 1.38-1.45

I computed 250 lines which are too long to paste here, they can be viewed at https://colab.research.google.com/drive/15hCE1N90Ie2iVEhMdiVAiIrg5i0SgmIN?usp=sharing

It can be extended from 100.0 to higher odds because it converges to but never hits 1.0, in other words the corresponding odds for 1.0, a certain probability, is infinity.

Lessons learned:

1- don't ever bet on the left lines except for the first one because they have good chances but bad value.
2- Do not chase odds over 10.00, it is not worth it.
3- Safest right lines after 2-2.4 are 2.4-2.8 and 2.8-3.2. As a corollary, 1.56-1.71 and 1.45-1.56 left lines retain the most value after 1.71-2.

I mainly wrote this for sports betting purposes but it can also be applied to a basic Hi-Lo-like dice game. My notebook computes that the numbers corresponding to the best right lines are below. Some casino games like freebitco.in dice only let you select probabilities so I also included those.

2-2.4 line edge - Over 5834 or under 4166, probability 41.67%
2.4-2.8 line edge - Over 6429 or Under 3571, probability 35.71%
2.8-3.2 line edge - Over 6875 or under 3125, probability 31.25%

So stop betting by multiples of 1000. Over 6000, 7000, 8000 etc. are poor line groups. Smiley

No spam

Self-moderated against spam. I spent a lot of time finding these results so please don't use this thread as a dumping ground to increase your sig campaign quota, thank you Smiley Discussion about these results is much welcomed.

Happy betting!
Jump to: