Smart betting

[powerz]

Okay here is a situation for you...

Since 2011 I have tracked closing MLB line up until Sept. So April 1 -Aug 31.
I track the Fav Closing line by League - Home and Away - and wether it went up or down from opening

I also track by FAV RL - FAV ML - DOG ML - DOG ARL

SInce 2011
National League home Favs whose ML went up north of -200 are 54% on RL and 71.4% SU
National League Home Fave whose ML went down but stayed north of -200 are 47% On RL and 71.1% SU

American League home Fave whose ML went up north of -200 are 57% on RL & 71.7% SU
American League home Fave whose ML went down but Stayed north of -200 are 42% on Rl nd 68% SU

I can give you every closing line calculation and which one is the most profitable for example when an American league Home team moves up & closes in -170s they hit the RL 49% of the time

 

[GoSooners]

I became much more successful when I became a flat bettor. I'm not saying it's good for everybody, but it was good for me. There were a fair share of games last year that I would have lost my ass on if I had followed my past methods and bet big on and lost. As is, I came out ahead. I didn't make a killing, but I did good. Which is what I prefer over the long run.... I also don't like parlays. This is where the books clean up. Not considering the possibility of a "push," the correct odds of hitting a 3 team parlay are 1 in 8. The true payout on a 3 team parlay should be 7-1, since, theoretically, every 8 bets would produce 1 winner and 7 losers. Since the payout on a 3 team parlay is usually 6-1, the house has a 12.5% edge. Not a fair wager. But people seem to love these bets.

 

[Akillies]

fo anyone genuinely interested in
[h=1]Quantitative Analysis in Sports[/h]http://www.degruyter.com/printahead/j/jqas

 

[Akillies]

Each March, more than an estimated 50 million Americans fill out a bracket for the National Collegiate Athletic Association (NCAA) men’s Division 1 basketball tournament (Barra 2014). While paid entry into tournament pools is technically outlawed, prosecution has proved rare and ineffective; an estimated $2.5 billion was illegally wagered on the tournament in 2012 (Boudway 2014; Tsu 2014).
Free tournament pools are legal, however, and Kaggle, a website that organizes free analytics and modeling contests, hosted its first college basketball competition in the early months of 2014. Dubbed the “March Machine Learning Mania” contest, and henceforth simply referred to as the “Kaggle contest,” the competition drew more than 400 submissions. We submitted two entries, detailed in Section 3, one of which earned first place in the 2014 contest.
This manuscript describes how we combined standard statistical methods with relevant data to build a men’s college basketball prediction model. Formally, building on themes first suggested by Carlin (1996), we blend information from the Las Vegas point spread with team-based possession metrics by using a weighted average of the predictions generated from logistic regression models. The success of this algorithm reinforces longstanding themes of predictive modeling, including the benefits of combining multiple predictive tools and the importance of using the best possible data.

 

[omgbokie]

..very interesting articles, but why vegas spread? Why not your own spread predictions?


One more thing, I do understand, this was win - lose type of bet, but this has value only against offered odds. There's no point of trying to predict outcome when there's e.g. 57-43% it's still a coin toss... as there are only 63 games it's sure to say that luck governs...


I use bivariatePoisson when determining chances. head on head is in one way expressed through theta, and values I get I compare with ones offered by bookies.


Math half is not an issue, I'm having problems with non statistical data assessment regarding circumstances tied to the game, you know: garbage in - garbage out, like e.g. traveling distance, or a new coach, are these really an important factors... and if so, let's put a number behind them..


Obviously a problem for all generations
I can imagine that g.i.-g.o. was also main concern to the article crew..

 

[omgbokie]

Anyone finds this interesting?
http://www.doc.ic.ac.uk/teaching/distinguished-projects/2011/j.wozniak.pdf

 

[FOLLOWMYPICKS]

Consistently laying heavy juice is not a good idea but in most instances we have no choice. Props are often juiced.. Majority of books open week 1 NFL lines for both sides -120.

I think if anybody disagrees with the fact you have a better chance at winning long term flat betting probably isn't somebody you should take advice from because flat betting allows the gambler to survive the year unless you're that bad and just throw darts for the hell of it.

 

[omgbokie]

Hi, thanks for advice. The thing is, that although flat staking gets to be right choice, one shouldn't take that road as "default best". There are situations where flat staking can be "missed opportunity" and one must be aware of that..
Now, I don't know if you know how to calculate most profitable system and appropriate stake, but I wrote a lot about it and even made calculator for that for everyone who doesn't know how to calculate it manually... I'm not sure if putting a link is allowed so I won't, but you can google it if you are interested in (advancedbetcalc).
The question that I am interested in is "which parameters are relevant and which are not". The article about tennis explains how odds are created (I've also included this in my calculator) but in most cases they differ from the ones offered by bookies, a lot. This means that some factors beside "demand and offer" are present, and in my experience (when we ruled out "girlfriend problems, new coaches, drinking problems, rivalship performance boosters...") the answer was almost always hidden in statistics..

 

[RenoChazz]

Okay here is a situation for you...

Since 2011 I have tracked closing MLB line up until Sept. So April 1 -Aug 31.
I track the Fav Closing line by League - Home and Away - and wether it went up or down from opening

I also track by FAV RL - FAV ML - DOG ML - DOG ARL

SInce 2011
National League home Favs whose ML went up north of -200 are 54% on RL and 71.4% SU
National League Home Fave whose ML went down but stayed north of -200 are 47% On RL and 71.1% SU

American League home Fave whose ML went up north of -200 are 57% on RL & 71.7% SU
American League home Fave whose ML went down but Stayed north of -200 are 42% on Rl nd 68% SU

I can give you every closing line calculation and which one is the most profitable for example when an American league Home team moves up & closes in -170s they hit the RL 49% of the time

Pretty Sharp.

 

[omgbokie]

these odds seems to be as we expect them to be... I miss to see the point.

 

[gijoe]

Ill give you a couple of systems for me. I have a lot. Bet kc when facing a left hand pitching. Take over 6 or 6.5 unless you hae a top 5 pitcher going. You need to keep notes an get usa paper and sports weekly.(a lot of info)

 

[omgbokie]

Thanks alot for your advice.
I have no problem with lack of info, my question was how to sort them out by importance..
..are the average points more important than possession percentages.. ..is the new girlfriend as important as the new coach..
It seems that bivariate poisson gets to be more accurate than "supply and demand" or pythagorean method when determining correct odds...

 

[omgbokie]

...so, does that imply that we should value average points let's say 10 and ball possession 7? Or they are of equal importance when formulating "odds equation"?

 

[omgbokie]

.. so how about this for player (A) service points % in This Match (TM)=

(A)%ServicePoints(TM) = (**(A)ServicePoints%}/(1+**(A)RANK}/100))/((**(A)ServicePoints%}/(1+**(A)RANK}/100))+(**(B)ReturnPoints%}/(1+**(B)RANK}/100)))*100

e.g.
(A) - ATP Rank=5, ServicePoints%=67, ReturnPoints%=38
(B) - ATP Rank=20, ServicePoints%=67, ReturnPoints%=38
-------
(A)%ServicePoints(TM) = 66,83%
(B)%ServicePoints(TM) = 60,67%

and from that - we calculate 1st set win odds.
( (A)=1.52 , (B)=2.91 )

(and everything else...)

any thoughts?

 

[PatsFan1283]

Have you tested this vs odds?

A lot of ways to model that have no predictive value over market sentiment (Nate Silver writing about sports comes to mind)

 

[omgbokie]

.yes, but only on few hundred games. It has positive roi but d=sqrt((n/k - n/k^2))/n says that I need more tries... It is some sort of log function and i cannot get my hands on it (yet) so this has some problems on edges (e.g. Fognini F. vs Rublev A. -as this youngster is 329. @atp) but looks good when I input average contestants in.. (e.g. Garcia-Lopez G. vs Baghdatis M.)

 

[omgbokie]

Recently i have been doing some calculations on "probability collisions" and it has a lot in common with sports betting.. Let me explain.. if we have two teams which have not played against each other this year, but have won 75% of their games against common opponents... What are the odds against each other? This one is rather easy, but what if i said 73% against 56%?? This calculus is commonly used when determining starting Poisson values (attack&defence strenght) but the common way seems to be wrong... I have found correct equation and I will incorporate it into my calculator, so... fingers crossed, ..will inform of progress...
(also calculator will be publicly available so you may help in research ..)

 

[omgbokie]

btw, do you use any sort of "quantitative analysis" or you just go "by hunch"?..

 

[omgbokie]

It took me awhile (several weeks of calculations on i7 non-stop), but finally here it is.. a theory is ready for real testing. Matrix is made for english premier soccer league (don't know which one is next but for now let's make money on this)
I have 15-4 (even odds) for the first five rounds... fingers crossed
..i will keep publishing value tips as calculation are made..

 

[Darkbay]

I used to win a lot of horse bets by picking the names of the horses that looked fancy to me because I didn't know how to read the racing form.....until one day someone taught me how and that screwed me all over.