How Would You MATHEMATICALLY Calculate the "FAIR" value of NFL 1/2 point ?

[Kramer]

How would you “Mathematically”
calculate the FAIR VALUE of an NFL
½ point?

For example ,
a gambler at -110 on a fair line needs to hit 52.4% to break even .

If you lay 120, you can buy a half point at most sports books
If you bet -120 , what’s your break even point ?

In the nfl , an average season has 10-15% pushes so
this has to be taken into consideration into the formula.



3 points (the big one) - close to 17% of all NFL games fall on this number.

If you have a team at -3 and they win by 3 it is a push so no big deal,
but taking 3.5 is something you need to think about carefully.
Conversely, getting +3.5 points on an underdog might be excellent value.

MANY BOOKS OBV DON’T OFFER ½ POINT ON OR OFF THE 3


7 points - 10% of all games fall on this number.

If 10% of NFL games fall on 7 points is it worth
betting -110 plus 10% on this 1 game??

10 points - A little over 6% of all games fall on this number.

Over the past five seasons of NFL action 33% of all games fell
on either 3, 7 or 10 points.
These are the three most significant key numbers.

6 points - Not surprisingly, almost 6% of NFL games
ended on this number.
14 - 5% of all games fell on this number,
but this is not as important a key number because
spreads rarely get to a 2 TD level.

4 points - A little over 4% ended on a spread of 4 points.


All other victory margins not mentioned above occurred less than 4% of the time during the past five seasons.

Would we need more database of 5 seasons ?\

E.g. 500 or 1000 games of Margin of victories,
pushes ..

THOUGHTS?

[Romanowski]

the push frequencies you stated are a little off

from a data set that I trust

push frequency on 3 is around 10%, on 7 is 6%...still very large but nowhere close to 17%

now to calculate the value of a half point, spread dependent of course

assuming -110 lines on fav and dog

off of -7/+7 to -6.5/+7.5 is worth 14 cents to -6.5 (-124)/+7.5(-124)

off of -3/+3 to -2.5/+3.5 is worth 25 cents to -2.5(-135)/+3.5(-135)

[MCP313]

Originally Posted by Romanowski

the push frequencies you stated are a little off

from a data set that I trust

push frequency on 3 is around 10%, on 7 is 6%...still very large but nowhere close to 17%

now to calculate the value of a half point, spread dependent of course

assuming -110 lines on fav and dog

off of -7/+7 to -6.5/+7.5 is worth 14 cents to -6.5 (-124)/+7.5(-124)

off of -3/+3 to -2.5/+3.5 is worth 25 cents to -2.5(-135)/+3.5(-135)

Basically buying points is a waste of money especially with the reduced juice options out there now. If someone is buying points say for a specific play it can somewhat be justified....come to think of it its a waste of money. If you feel the need to buy points you shouldnt be playing it to begin with. Just my opinion.

:peace:

[GFootsXtreme]

Here's my problem with this analysis and I very much welcome Romo and everybody else to comment --

This type of analysis never takes into account the difference between dogs and favorites, or how big the spread is.

For example, if Romo's #s are correct, 10% of the NFL games land on 3. That means that some percenetage of FAVORITES win by 3, and some % of DOGS win by 3, but it does not simply mean that buying a half point gives you the full value of the % chance the game lands on 3...because you only have "half" of the "3", either the dog or the favorite.

The analaysis also does not break down the %'s based upon the spread...I would think statistically buying the half-point has more value on a low-spread game than a high one.

if the Pats are favored by 20 in a game, the chances of that game landing on 3 are less than if the packers play the vikings and the spread is set at pick em (as an example)

of course my analysis may be wrong but it occurs to me that buying points really is a scam and although you do win sometimes doing it, if you do it all the time you lose overall in doing so

comments?

[upinya83]

Originally Posted by GFootsXtreme

Here's my problem with this analysis and I very much welcome Romo and everybody else to comment --

This type of analysis never takes into account the difference between dogs and favorites, or how big the spread is.

For example, if Romo's #s are correct, 10% of the NFL games land on 3. That means that some percenetage of FAVORITES win by 3, and some % of DOGS win by 3, but it does not simply mean that buying a half point gives you the full value of the % chance the game lands on 3...because you only have "half" of the "3", either the dog or the favorite.

The analaysis also does not break down the %'s based upon the spread...I would think statistically buying the half-point has more value on a low-spread game than a high one.

if the Pats are favored by 20 in a game, the chances of that game landing on 3 are less than if the packers play the vikings and the spread is set at pick em (as an example)

of course my analysis may be wrong but it occurs to me that buying points really is a scam and although you do win sometimes doing it, if you do it all the time you lose overall in doing so

comments?

rome will have to correct me if I'm wrong, but when he says push frequency is 10% - that means the line is 3 and the end result is 3 - not that a team is favored by 20 and it lands on 3. Im guessing that this number only takes into account games lined at 3, nothing else

Because it also states that push frequency on 7 is 6% - and I'd have to guess that more than 1 in 17 or 18 games end up on 7 - the % stated is in regards to games lined as that number

am i reading that wrong?

[Romanowski]

^ exactly, the 10% push frequency implies, the line was 3 and it lands on 3

[ZoomVT]

this then begs the question; is it profitable or do you gain an edge by buying the hook to give yourself a shot at a push?

[Kramer]

If I read Rome correctly , the 7 point spread is worth
14 cents.

If the book charges -120 , that's a 6% house edge,

but if the half point changes the bet so you are laying 12 to 10,
YOUR BREAK EVEN percentage increases to 54.55!

So you're paying 6% house edge , your break even point increases
by basically 3% ,

Therefore not mathematically worth it,
either find another game off key #,
or just bet who you like ,

why pay the man + 6 % to increase your b/e by 3%?

[smitch124]

Much of the statistical analysis done for push percentages will have a certain "radius" in order to include more data. For example, the push percentage for 3 will take into account all teams that are -2 through -4, therefore having a radius of 2. This includes many more data points for statistical significance while still being applicable.

[uva3021]

you would need to find the correlation coefficient between push frequency and total number of games convened on the relevant key number. I'm guessing with each number (3,4,7,10,14,21) it won't be perfectly linear, might actually be logarithmic, with the 3s having the highest percentage of push frequency

once you find the relation, whether logarithmic or linear, you could probably place it on a graph and use the unknown variable in the equation as the value you are looking for

[GFootsXtreme]

First let me say this is a great discussion, very good input all around.

Second, let me say, this reminded me to be confident, not to buy 1/2 points and that alone has saved me probably $100

There's a few different pieces of data being discussed at once.

I think what I have read is that in NFL football 17%, or 14% (not sure which) actually have an ending result on 3, regardless of what the actual line of the game was.

Romo's stats are telling us that if the game is lined at 3, the chances the game falls onto 3 is 10%.

Take romo's stats -- my question is about dog versus favorite. if I take a dog at +3, I would push on the fave winning by 3 and win if the dog wins by 3. But if I take the favorite, I would push on 3 with a win by 3 and lose if the dog won outright.

In trying to analyze whether or not it is worth buying a 1/2 point, don't we need to account for the different sides i.e. the dog getting 3 points is much more valuable potentially?

[Kramer]

Add'l info from a database I trust


In the NFL home underdogs have historically done well against the spread.
The following table shows the probability of winning, losing, and pushing
according by various breakdowns.
The table covers every game from the 1983 to 2008 seasons.


The expected return is for the standard laying 11 to win 10.
The expected return when laying 105 to win 100 is about 2.2%
more than when laying 11 to win 10.



1. Expected Return against Point Spread in the NFL
Event Win Lose Push Exp. Ret. (-110)
All favorite 47.19% 50.05% 2.76% -7.15%
All underdogs 50.05% 47.19% 2.76% -1.69%
All visiting 48.71% 48.57% 2.72% -4.29%
All home 48.57% 48.71% 2.72% -4.56%
Home favorite 47.56% 49.75% 2.7% -6.51%
Visiting underdog 49.75% 47.56% 2.7% -2.33%
Home underdog 50.7% 46.39% 2.91% -0.3%
Visiting favorite 46.39% 50.7% 2.91% -8.53%





Over 1975 games during the 1993 to 2000 seasons
the overall house edge by purchasing the extra half point is 4.13%.

[ProlinePlayer]

Originally Posted by Kramer

1. Expected Return against Point Spread in the NFL
Event Win Lose Push Exp. Ret. (-110)
All favorite 47.19% 50.05% 2.76% -7.15%
All underdogs 50.05% 47.19% 2.76% -1.69%
All visiting 48.71% 48.57% 2.72% -4.29%
All home 48.57% 48.71% 2.72% -4.56%
Home favorite 47.56% 49.75% 2.7% -6.51%
Visiting underdog 49.75% 47.56% 2.7% -2.33%
Home underdog 50.7% 46.39% 2.91% -0.3%
Visiting favorite 46.39% 50.7% 2.91% -8.53%

Great info in this thread.

Just as an observation ...

I would guess (have no proof of this) that a lot of the variation shown in the above is no longer there.

I believe that today the influence of the smart money is a lot stronger and forces the books to post more accurate numbers.

They can no longer do as in the 80s and post what they knew were incorrect lines because they could predict how joe public would bet the games.

PLP

[Steel]

Enjoy the read thanks for the tech training.

[msearcy]

wow, I will say most of this is above my head. But I think I get the point. Stay away for buying 1/2 point.