Math Models

F

Foom

EOG Enthusiast
Hello all. I've been reading eog for some time now & finally got around to registering. I've been casually wagering on sports for 5 or so years, but strictly thru a 'feel' process. Watching games, reading news articles on the net & then evaluating that info. Long story short, this has been -ev.

I'm interested in creating a 'model'. The problem is that I have no idea where to begin? What data is important? What data would be weighed heavier than other data? Do I need excel? Can this be done by hand? This very may well be sport specific, but I'm just looking for a starting point. I completely understand not giving out your secrets; I am not looking for a handout, I simply want to become a more knowledgeable sports bettor.

Lastly, I'm in between jobs & without a computer.(posting on my iphone) This means I won't be wagering until mlb. Thanks to eog for any help you may be able to offer & thanks for listening.

-Foom
 
John Kelly

John Kelly

Born Gambler
Staff member
Re: Math Models

Welcome to the "G," Foom.

My favorite model stands 5'11" and she wears nothing but lingerie from Victoria's Secret.

But seriously, plenty of math guys here in the community will offer advice.

Basketball handicappers should start at possessions per game and points per possession.
 
WVU

WVU

EOG Master
Re: Math Models

Play the bonus game to build up your bankroll and then work on your model. Bonus play pays better than any model, but if you learn to do both you are golden
 
F

Foom

EOG Enthusiast
Re: Math Models

Thank You John! She sounds wonderful.IT, thanks for the link. I glanced at it; seems very helpful.WVU, what do you mean by 'bonus games'? Excuse my ignorance.
 
WVU

WVU

EOG Master
Re: Math Models

Don't ever make a deposit without having a bonus attached to it. 2 good places to start are bet revolution and just bet dot com. Do research on some of my older posts and you will find ways to beat the house without picking winners. It takes intense discipline. Good luck.
 
F

Foom

EOG Enthusiast
Re: Math Models

Thanks for the clarification WVU, now I understand. I will check into your past posts.
 
Sleepy

Sleepy

EOG Master
Re: Math Models

Screw the math models.

The only thing that wins is painstaking research and lots of experience.
 
John Kelly

John Kelly

Born Gambler
Staff member
Re: Math Models

The house uses math models.

I look for angles not incorporated in the model (player matchups, coaching edges, motivation, team chemistry, injuries, etc.).

I also like "changes."

Focus on any change (personnel change, coaching change, etc.) that will invalidate a team's past performances.
 
C

Coach Taft

EOG Member
Re: Math Models

Sleepy said:
Screw the math models.

The only thing that wins is painstaking research and lots of experience.
Click to expand...

agree with this.

Also, if you have at least 20 outs, you are off to a good start.

The game is all about finding value. Its not complicated. Manage you BR, don't over-bet, and you will be fine as a second job.

prob. with full time is that, a. people are flakey and don't pay, most gamblers, bookies are scumbags, b. sportsbooks don't take big plays, and the ones that do are sharp.

good luck.
 
LQQKer

LQQKer

2
Re: Math Models

Foom, unless you are very into Excel (at least), grok db grunt work, are at least a decent math guy and not only can stand statistics, but avidly work probability problems like other folks play with crossword puzzles, don't bother.

...many aspire, but few achieve.

Do bonus whoring, as WVU says...pays better.

...@nd you then h@ve extr@ t!me to fvck ur w!fe and bre@k keybo@rd$
 
BrickShitHouse

BrickShitHouse

2
Re: Math Models

LQQKer said:
Foom, unless you are very into Excel (at least), grok db grunt work, are at least a decent math guy and not only can stand statistics, but avidly work probability problems like other folks play with crossword puzzles, don't bother.

...many aspire, but few achieve.

Do bonus whoring, as WVU says...pays better.

...@nd you then h@ve extr@ t!me to fvck ur w!fe and bre@k keybo@rd$
Click to expand...
:LMAO
 
F

Foom

EOG Enthusiast
Re: Math Models

So creating a 'model' isn't the direction I should go in after all? Well thanks for clearing the air. That's why I asked.
 
C

Coach Taft

EOG Member
Re: Math Models

Foom said:
So creating a 'model' isn't the direction I should go in after all? Well thanks for clearing the air. That's why I asked.
Click to expand...

Its part of the game, and worth it. But, it depends on what you are trying to accomplish. If you are not betting at least a dime a game, it is worthless imo.

You can have the best math model in the world, but if you don't have the outs that pay and let you get down (good outs are more important than math models), then the model is worthless.

There are easier ways to make money, like poker and stocks and working a regular job with benefits.
 
Doughy

Doughy

2
Re: Math Models

Like JK said, "the house uses math models". If your model drastically differs from the house, it's likely yours is off. I had a guy with a PhD in statistics from Rice University making several models for football games and for the NFL we rarely saw the models produce a spread/total deviate more than 3 points from the actual opening line. And when we did, it was usually a game with a significant injury involved. I saw much more discrepancy with college football, but that's obviously expected when the spreads and totals are oftentimes much larger. I think on average I was spending 5-10 hours a week gathering all of the relevant stats the model needed for the next week's lines and putting it into .xls or .csv format for him to compute.
 
F

Foom

EOG Enthusiast
Re: Math Models

Coach & Dough, great posts. Thanks! Esp the point about at least betting a dime per game. I am no where near that figure.
 
Doughy

Doughy

2
Re: Math Models

I'm not going to pretend that I knew the math involved with what my friend developed, nor would I divulge his trade secrets if I did, but I'm curious if other modelers agree that power rankings are of paramount importance for NCAAF modeling.

Team X could rush for 200 yards per game and convert 50% of the time on 3rd down, but if it was against crap competition it must be heavily discounted. In the NFL, you don't really have that problem since every team could beat every other team on any given Sunday. There are only a few sites that rank all 120 something NCAAF teams on a week-by-week basis and I know my friend was using CBS Sportsline and one other site's composite rankings to help ensure that the stats he used were as relevant as possible. One thing we always struggled with was quantifying injuries and typically laid off those games altogether when a key starter was out. I know some modelers can measure that, but how they do I don't know.
 
LQQKer

LQQKer

2
Re: Math Models

Doughboy said:
Like JK said, "the house uses math models". If your model drastically differs from the house, it's likely yours is off. I had a guy with a PhD in statistics from Rice University making several models for football games and for the NFL we rarely saw the models produce a spread/total deviate more than 3 points from the actual opening line. And when we did, it was usually a game with a significant injury involved. I saw much more discrepancy with college football, but that's obviously expected when the spreads and totals are oftentimes much larger. I think on average I was spending 5-10 hours a week gathering all of the relevant stats the model needed for the next week's lines and putting it into .xls or .csv format for him to compute.
Click to expand...

Unfortunately a poor conception. The idea is to find the true value of the teams ON THE DAY OF THE GAME.

The initial line might be set by some math model, but it is driven by the market, which is mostly right by closing. As the culmination of many smart folks, it is thusly pretty good and hard to beat in terms of overcoming the vig.

One last note, PhDs are particularly bad in building models, as they come pre-packaged with all sorts of preconceptions about "what should work" and the "theory says this". With all the storms in their heads, they can't think, nor build anything that works.
 
trytrytry

trytrytry

All I do is trytrytry
Re: Math Models

Coach Taft said:
agree with this.

Also, if you have at least 20 outs, you are off to a good start.

The game is all about finding value. Its not complicated. Manage you BR, don't over-bet, and you will be fine as a second job.

prob. with full time is that, a. people are flakey and don't pay, most gamblers, bookies are scumbags, b. sportsbooks don't take big plays, and the ones that do are sharp.

good luck.
Click to expand...

this guy needs to post more.the only thing her forgot to mention is find some reduced juice options as well..:cheers
 
Doughy

Doughy

2
Re: Math Models

LQQKer said:
Unfortunately a poor conception. The idea is to find the true value of the teams ON THE DAY OF THE GAME.

The initial line might be set by some math model, but it is driven by the market, which is mostly right by closing. As the culmination of many smart folks, it is thusly pretty good and hard to beat in terms of overcoming the vig.

One last note, PhDs are particularly bad in building models, as they come pre-packaged with all sorts of preconceptions about "what should work" and the "theory says this". With all the storms in their heads, they can't think, nor build anything that works.
Click to expand...
Valid points, but it's somewhat analogous to what he developed since the lines tend to move in the direction of his numbers, even if they differed ever so slightly (2-3 points) from the opening line. Even without waiting for the Sunday and Monday games to finish to include the previous week's stats in the equation, his models could predict the lines for the following week's games with amazing accuracy.

Here's an example of what I'm talking about: on Sunday night 11/28/10 he ran his models and produced lines and totals for the following week's games before the lines were available anywhere and only one spread differed by more than 3 points (http://forums.eog.com/showthread.php?t=306183&highlight=lines). Those results are typical for the past 3 years.
 
F

Foom

EOG Enthusiast
Re: Math Models

Makes sense fellas. Even if I were to generate a model, how can I expect to have a better grasp than the house? I most likely can't. I'll continue to try & gather as much info on a team/game & go from there. One last question if I may...Regarding bases, I often hear 'By my numbers, I have Team A at pick 'em yet pinny lists them at +10, therefore Team A is a play.' Is this a math guy? Or is he digging for info & 'feels' that Team A & Team B are so evenly matched, that +10 is too hard to pass up?What are you looking at to make Team A a pick 'em? I understand you look at starting pitching, home/away, pens, injuries, etc...my confusion is how does that get you to pick 'em?
 
raiders72001

raiders72001

1
Re: Math Models

WVU said:
Play the bonus game to build up your bankroll and then work on your model. Bonus play pays better than any model, but if you learn to do both you are golden
Click to expand...
you never cease to amaze me. Where do you come up with this crap?
 
raiders72001

raiders72001

1
Re: Math Models

LQQKer said:
Unfortunately a poor conception. The idea is to find the true value of the teams ON THE DAY OF THE GAME.

The initial line might be set by some math model, but it is driven by the market, which is mostly right by closing. As the culmination of many smart folks, it is thusly pretty good and hard to beat in terms of overcoming the vig.

One last note, PhDs are particularly bad in building models, as they come pre-packaged with all sorts of preconceptions about "what should work" and the "theory says this". With all the storms in their heads, they can't think, nor build anything that works.
Click to expand...
sharp post except I disagree with true value on the day of the game. You want the best line considering all lines between openers and closers unless limits prevent you from taking that number.
 
WVU

WVU

EOG Master
Re: Math Models

raiders72001 said:
you never cease to amaze me. Where do you come up with this crap?
Click to expand...

LOL, you are a complete tool. Most know this. You find me a model that can give the player over 15% advantage. Yeah...that's what I thought :shoot:
 
raiders72001

raiders72001

1
Re: Math Models

WVU said:
LOL, you are a complete tool. Most know this. You find me a model that can give the player over 15% advantage. Yeah...that's what I thought :shoot:
Click to expand...
you can't make that much bonus whoring long term. You just found a couple a books that you can do this with for peanuts and you haven't done it for years.
 
raiders72001

raiders72001

1
Re: Math Models

WVU- If you got $1000 bonus, which is high, are you saying that you made $150? Books with stop this bonus if you are beating them.

One of the books that you mentioned took the bonus money away from the player without paying him.

The player wasn't stiffed but wasn't paid his bonus.
 
WVU

WVU

EOG Master
Re: Math Models

raiders72001 said:
you can't make that much bonus whoring long term. You just found a couple a books that you can do this with for peanuts and you haven't done it for years.
Click to expand...

You have no clue now do you? Those of us who know bonuses will tell you otherwise. Exactly how much can you make modeling? Yes, I can generate 15% or more edge using bonuses currently available. The bonus game changes but I have been successful using them this year as well as in years past. Please don't comment on something you know nothing about. Thanks
 
royalfan

royalfan

EOG Dedicated
Re: Math Models

My god. WVU acting like bonus whoring is the way to go in the sportsbetting industry? You cannot make this up folks. That is about as small time as it gets.

Then you got an idiot like High Times calling him out for it when he could probably use a bonus or 15.

:LMAO
 
WVU

WVU

EOG Master
Re: Math Models

raiders72001 said:
WVU- If you got $1000 bonus, which is high, are you saying that you made $150? Books with stop this bonus if you are beating them.

One of the books that you mentioned took the bonus money away from the player without paying him.

The player wasn't stiffed but wasn't paid his bonus.
Click to expand...

no, I am not saying that. I would make much more than $150 on $1000 in bonuses. I have never been stiffed on a bonus.
 
H

High Times

EOG Master
Re: Math Models

WVU said:
didn't you get 16th in the field of 16? You should be very familiar with losing. I have won plenty. You are < 50% capper as has been proven with your posted picks.
Click to expand...



You are an internet joke.....

You (and the scum like you) are the sole reason these Books hate to give out bonuses.

You ain't bullshitting anybody.

If you were a winner you would never stoop to that level
 
H

High Times

EOG Master
Re: Math Models

royalfan said:
My god. WVU acting like bonus whoring is the way to go in the sportsbetting industry? You cannot make this up folks. That is about as small time as it gets.

Then you got an idiot like High Times calling him out for it when he could probably use a bonus or 15.

:LMAO
Click to expand...


Dude go the fuck away and join all the other little boys who you can bullshit
 
WVU

WVU

EOG Master
Re: Math Models

royalfan said:
My god. WVU acting like bonus whoring is the way to go in the sportsbetting industry? You cannot make this up folks. That is about as small time as it gets.

Then you got an idiot like High Times calling him out for it when he could probably use a bonus or 15.

:LMAO
Click to expand...

welcome back, Chad. You do your thing and i will do mine. I have done pretty well over the years including this past year
 
WVU

WVU

EOG Master
Re: Math Models

High Times said:
You are an internet joke.....

You (and the scum like you) are the sole reason these Books hate to give out bonuses.

You ain't bullshitting anybody.

If you were a winner you would never stoop to that level
Click to expand...

And yet books continue to give out bonuses... I won the invite that you got last in. I won many other contests too. You post losers yet claim to win. Hmmm :+clueless
 
You must log in or register to reply here.
Top