[Juret]

Hi,

I have a data set of goal totals that are like a positively skewed Poisson distribution (tail to the right) due to a couple of outliers. The variance is for example larger than the mean and it doesn't pass a Chi-Square goodness-of-fit test for Poisson, failing heavily because of the tail to the right (otherwise fitting quite good). Any ideas how to deal with this? I don't think ignoring the outliers would be a good idea, it seems consistent that sometimes the score drifts away. When the score has drifted away, it is always due to a team winning huge, with only one exception when there was a high-scoring tie. It seems to me that sometimes a team just gives up, and this disturbs the Poisson process.

Thank you a lot

[Justin7]

How many of those outliers involved either red-cards, or gross mismatches?

[bztips]

You should try using the negative binomial distribution instead. With the Poisson, mean = variance by definition. The negative binomial relaxes this feature, and is specifically appropriate in the case you described of "overdispersion", where the variance exceeds the mean.

[Juret]

How many of those outliers involved either red-cards, or gross mismatches?

The outliers are almost always gross mismatches, top teams vs bottom teams. I am thinking of finding a fair coefficient to adjust large expected spreads to be even larger just as my distribution seems to suggest is the case. Match penalties are rare in bandy, it works more like in hockey.


I actually read about the negative binomial distribution before posting, but I am not sure what values to plug in for r and p; p being the probability of a goal I find hard to grasp.

[CHUBNUT]

not sure about a skewed poisson but i've had a twisted bollock and i know how painful that was. its all about the position you take on a bet, when sitting its best with legs apart.

[CHUBNUT]

Hi,

I have a data set of goal totals that are like a positively skewed Poisson distribution (tail to the right) due to a couple of outliers. The variance is for example larger than the mean and it doesn't pass a Chi-Square goodness-of-fit test for Poisson, failing heavily because of the tail to the right (otherwise fitting quite good). Any ideas how to deal with this? I don't think ignoring the outliers would be a good idea, it seems consistent that sometimes the score drifts away. When the score has drifted away, it is always due to a team winning huge, with only one exception when there was a high-scoring tie. It seems to me that sometimes a team just gives up, and this disturbs the Poisson process.

Thank you a lot

I know people arent going to like this on here but your not trying to land a man on the moon, your deciding whether a soccer game has more than 2 goals. your not going to be right all the time and when your wrong the degree will play havoc with your thinking. In my extraordinary experience betting goal totals is not for the faint hearted. its far to easy for the books to lean the vig in the expected mathematical result leaving the bettor to either take a bad price or be on the considered mug side.

[sharpcat]

I know people arent going to like this on here but your not trying to land a man on the moon, your deciding whether a soccer game has more than 2 goals. your not going to be right all the time and when your wrong the degree will play havoc with your thinking. In my extraordinary experience betting goal totals is not for the faint hearted. its far to easy for the books to lean the vig in the expected mathematical result leaving the bettor to either take a bad price or be on the considered mug side.

Your a waste of oxygen.

[bztips]

Chub can't help himself, he just loves to rub our noses in his ignorance.

[sharpcat]

Chub contradicts himself in every post he makes and sounds like a moron.

First he suggests that you can not find an accurate probability of the outcome of an event and than he goes on to say that the books can shade the vig to lean towards the true mathematical result......WTF is this idiot rambling about?

[CHUBNUT]

Your a waste of oxygen.

Whats your beef the guy bets an under and the game finishes 6-2. now he's scouring the math books to look for an answer. all I'm telling him is shit happens, move on.

[CHUBNUT]

Chub can't help himself, he just loves to rub our noses in his ignorance.

thats good i'll have to remember that one. thanks.

[Juret]

Whats your beef the guy bets an under and the game finishes 6-2. now he's scouring the math books to look for an answer. all I'm telling him is shit happens, move on.

I'm not crying out.. I'm just trying to understand a sport here and THEN bet it..

[CHUBNUT]

I understand your quandry but when it comes to total goals in soccer no matter what stats or math you use the books will have the vig on that favored side. in other words you work out the under is the play, if your figures are correct the books will have that under at -120. as random as baseball totals are soccer is worse. stick to asian betting on soccer.

[sharpcat]

I understand your quandry but when it comes to total goals in soccer no matter what stats or math you use the books will have the vig on that favored side. in other words you work out the under is the play, if your figures are correct the books will have that under at -120. as random as baseball totals are soccer is worse. stick to asian betting on soccer.

So what if his numbers say the under has a 56% chance of winning and the book has -120?

Do you understand push probabilities?

[CHUBNUT]

So what if his numbers say the under has a 56% chance of winning and the book has -120?

Do you understand push probabilities?

then he does what all the other idiots do, take the -120 and wonder why they cant make the game pay.

[sharpcat]

then he does what all the other idiots do, take the -120 and wonder why they cant make the game pay.

The game won't pay if he has a 2.5%+ edge?

Why do lines move if the bookmakers always have the right number CHUB?

[Juret]

I understand your quandry but when it comes to total goals in soccer no matter what stats or math you use the books will have the vig on that favored side. in other words you work out the under is the play, if your figures are correct the books will have that under at -120. as random as baseball totals are soccer is worse. stick to asian betting on soccer.

LOL, where did you get I was betting soccer from? I would never bet soccer, I'm doing this for bandy, a sport you are very unlikely to have heard about

[CHUBNUT]

bandy? i thought that was a leg problem. better a leg problem than what sharpcats got.

[sharpcat]

bandy? i thought that was a leg problem. better a leg problem than what sharpcats got.

It would be nice if Justin7 would clean the think tank up and start handing out infractions for players talk style of posting. It is clear that your intentions in the think tank are to just act like a clown all of the time and dumb down the subforum.

The think tank does not need a JJGold take your nonsense to players talk.

Most of the more educated posters like MonkeyFocker, Data, and Thremp have abandoned this sub-forum. This is sad

[ChuckyTheGoat]

Not sure this is the answer u want, but my 2 cents:

One way to adjust for blowout results = adjust the raw data to a more rational level. IE, a team wins 13-0. Identify a formula to make it more reasonable. For margins of > 6.0, maybe your Adjusted Margin = 6.0 + .60 * (X - 6). So, a 13-goal net would translate to an Adjusted 10.2 Net. I have found this to be a helpful adjustment.

You do see it in mismatches. Attacking teams will go for a ruthless final. Defensive teams are in just as much control, but don't go for an embarrassing scoreline.

[Juret]

Not sure this is the answer u want, but my 2 cents:

One way to adjust for blowout results = adjust the raw data to a more rational level. IE, a team wins 13-0. Identify a formula to make it more reasonable. For margins of > 6.0, maybe your Adjusted Margin = 6.0 + .60 * (X - 6). So, a 13-goal net would translate to an Adjusted 10.2 Net. I have found this to be a helpful adjustment.

You do see it in mismatches. Attacking teams will go for a ruthless final. Defensive teams are in just as much control, but don't go for an embarrassing scoreline.

Yes, this is what I am looking to do..