Books on Statistial Modeling

**brettd** · 03-07-11, 10:42 PM

Justin7's book 'Conquering Risk', it has four models in there explained step by step. 'Mathletics' by Wayne L Winston is good too.

**Curls89** · 03-07-11, 10:47 PM

Originally posted by Pushkin

In the past several weeks I have read the majority of the "gambling" books that have been recommended in the Think Tank and I have also read as many threads as I could find on the subject of statistical modeling. Recently I have been introduced to some YouTube videos on the subject and I believe I am (slowly) starting to get a better grasp of the process. Can anyone recommend any additional resources like books and videos that would help further explain statistical modeling? Specifically, are there any books, web sites or vidoes that give a step by step explanation of how to build a statistical model?

Thank you

What gambling books have you read?

**Peregrine Stoop** · 03-07-11, 10:54 PM

look for colleges' intro to stats and probability for math/engineering/stat majors. Read what they are reading. Expand from there.

**Pushkin** · 03-08-11, 09:01 AM

Originally posted by Curls89

What gambling books have you read?

I just finished "Conquering Risk" and "Weighing the odds in sports betting". I am currently reading "Sharp sports betting" and "Mathletics".

Can you recommend any others that specifically address statistical modeling?

Thank you.

**Pushkin** · 03-08-11, 09:30 AM

Originally posted by brettd

Justin7's book 'Conquering Risk', it has four models in there explained step by step. 'Mathletics' by Wayne L Winston is good too.

Thanks brettd. I just finished "Conquering Risk' and recently started 'Mathletics'. Unfortunately I found some of the modeling information in those four chapters to be somewhat confusing. I just didn't understand where some of the numbers in the formulas came from. Obviously I need to re-read those chapters. I understood the concepts, and in fact was quite surprised to discover that I actually performed some of the functions in my current NCAAF capping process. But I still do not feel like I have a good grasp of the 'modeling process' yet.

**Pushkin** · 03-08-11, 09:31 AM

Originally posted by Peregrine Stoop

look for colleges' intro to stats and probability for math/engineering/stat majors. Read what they are reading. Expand from there.

Thank you for the idea(s) and suggestion Peregrine Stoop.

**Curls89** · 03-09-11, 12:48 AM

Originally posted by Pushkin

I just finished "Conquering Risk" and "Weighing the odds in sports betting". I am currently reading "Sharp sports betting" and "Mathletics".

Can you recommend any others that specifically address statistical modeling?

Thank you.

I'm actually just getting into sports betting so no. All the books you mentioned are currently being shipped to my house though.

**NYER5680** · 03-09-11, 03:47 AM

Conquering risk ,without a tout ,basketball on paper,

**Pushkin** · 03-10-11, 06:04 PM

Originally posted by NYER5680

Conquering risk ,without a tout ,basketball on paper,

Thank you for the suggestions NYER5680.

**IrishTim** · 03-11-11, 01:58 AM

One I didn't see mentioned that is very good (and a bit advanced): Statistical Thinking in Sports.

**Ralphie88** · 03-24-11, 06:26 PM

Originally posted by Pushkin

Thanks brettd. I just finished "Conquering Risk' and recently started 'Mathletics'. Unfortunately I found some of the modeling information in those four chapters to be somewhat confusing. I just didn't understand where some of the numbers in the formulas came from. Obviously I need to re-read those chapters. I understood the concepts, and in fact was quite surprised to discover that I actually performed some of the functions in my current NCAAF capping process. But I still do not feel like I have a good grasp of the 'modeling process' yet.

I read Conquering Risk and I also have no idea where some of the numbers in the formulas came from. I was specifically interested in the college football model. For example there is a formula for Projected passing attempts = 40+0.94 x OPYA -3.15 x ORYA -0.28 x Opp DPYA -0.70 x Opp DRYA

There is no explanation of where or what 40, 0.94, 3.15, 0.28 or 0.70 are found and what they represent. Are these personal formulas, standard formulas, what?

**Maverick22** · 03-25-11, 09:21 AM

:: Sigh ::

**CanuckG** · 03-25-11, 07:33 PM

Originally posted by Ralphie88

I read Conquering Risk and I also have no idea where some of the numbers in the formulas came from. I was specifically interested in the college football model. For example there is a formula for Projected passing attempts = 40+0.94 x OPYA -3.15 x ORYA -0.28 x Opp DPYA -0.70 x Opp DRYA

There is no explanation of where or what 40, 0.94, 3.15, 0.28 or 0.70 are found and what they represent. Are these personal formulas, standard formulas, what?

Maybe just example formulas?

**Justin7** · 03-25-11, 08:43 PM

Originally posted by Ralphie88

I read Conquering Risk and I also have no idea where some of the numbers in the formulas came from. I was specifically interested in the college football model. For example there is a formula for Projected passing attempts = 40+0.94 x OPYA -3.15 x ORYA -0.28 x Opp DPYA -0.70 x Opp DRYA

There is no explanation of where or what 40, 0.94, 3.15, 0.28 or 0.70 are found and what they represent. Are these personal formulas, standard formulas, what?

The coefficients were determined by regression of past data. They represent the optimal values to minimize absolute mean error (not squared) of each projection.

**underthe total** · 03-25-11, 09:11 PM

justin you seem like one smart dude. i would be interested in see ing your plays. i would hold them if that is what it took.

**RickySteve** · 03-25-11, 10:05 PM

Modeling is a great path to riches. Look at how all these guys crush the market year after year.

**goblue12** · 03-25-11, 11:58 PM

Originally posted by RickySteve

Modeling is a great path to riches. Look at how all these guys crush the market year after year.

I hate the douches that beat me in the PK contests at other forums when all I do is pick the favorite on the updated spread.

**RickySteve** · 03-26-11, 02:38 AM

Originally posted by goblue12

I hate the douches that beat me in the PK contests at other forums when all I do is pick the favorite on the updated spread.

Game theory much?

**bulichme** · 03-26-11, 08:49 AM

Originally posted by RickySteve

Modeling is a great path to riches. Look at how all these guys crush the market year after year.

How would you use this info to crush a book???? You would have to wager against them ats to show any profit, and probably avoid large money lines on su wagers possibly incorperating a 2 team parlay system of some sort.

**Tackleberry** · 03-26-11, 10:12 AM

Originally posted by bulichme

How would you use this info to crush a book???? You would have to wager against them ats to show any profit, and probably avoid large money lines on su wagers possibly incorperating a 2 team parlay system of some sort.

I believe RickySteve was being sarcastic

**Ralphie88** · 03-26-11, 10:19 AM

Originally posted by Justin7

The coefficients were determined by regression of past data. They represent the optimal values to minimize absolute mean error (not squared) of each projection.

Thank you for the explanantion Justin.

**Peregrine Stoop** · 03-26-11, 01:03 PM

Originally posted by Ralphie88

I read Conquering Risk and I also have no idea where some of the numbers in the formulas came from. I was specifically interested in the college football model. For example there is a formula for Projected passing attempts = 40+0.94 x OPYA -3.15 x ORYA -0.28 x Opp DPYA -0.70 x Opp DRYA There is no explanation of where or what 40, 0.94, 3.15, 0.28 or 0.70 are found and what they represent. Are these personal formulas, standard formulas, what?

That model is going to have a lot of auto-correlation among supposed independent variables troubles

**Peregrine Stoop** · 03-26-11, 01:05 PM

For anyone starting from scratch, the Kahn Academy has a very good video series up on probability and statistics. Now, it won't get you to the point were you are deriving Fisher's information. But, it should teach you how to gather data and create decent models.

*disclaimer: I think the time spent doing this would be much more profitably spent pursuing other avenues in sports betting. However, for many of us, this is just good ol' fashioned fun.

**bulichme** · 03-27-11, 09:42 AM

Originally posted by Tackleberry

I believe RickySteve was being sarcastic

I was hoping someone had done research in that area.