[SteveRyan]

Many times this NFL season, I had correctly figured out which team would cover the spread. Unfortunately, I lost several of those bets because I parlayed the spread with a bet on the total....and did not choose the over/under correctly. So, I decided to build a tool to help me determine if the total will go over or under.

It occurred to me, that given any spread and total, if I choose correctly which team will cover the spread, then there is a certain number of final scores where the spread is covered and the over is achieved. There would also be a certain number of final scores where the spread is covered and the under is achieved.

So, I made a calculator in Excel, that shows how many final scores cover the spread and the total goes over. It also shows how many final scores cover the spread and the total goes under.

If one is significantly larger than the other, then it seems obvious which to choose. If they are close, it may be better to pass.

I have attached the calculator in this thread for anyone to review.

It is a Beta version. It only calculates for the FAVORITE. It calculates the results based upon spreads with a hook (.5) and totals without a hook.

For example:

New Orleans -3.5
San Francisco 47

If you intend to use the calculator, please read the following explanation.....even though it is a very simple calculator.

How to use the calculator:

1. On your own, choose a match-up where the favorites spread has a hook, the total has no hook, AND you think the favorite will cover the spread (You can use the example I give above if you like.)
2. On your own, estimate how many points you think the favorite will score in the game.
3. Enter the spread into the green box under the "Spread" cell.
4. Enter the total into the green box under the "Total" cell.
5. Enter the estimated number of points that the favorite will score into the green box under the "Fav est pts" cell.
6. Click anywhere outside the calculator (Do NOT click inside the grey cells because those cells contain formulas that cannot be messed with!!)

Explanation of results:

"Fav req pts": This cell shows the minimum points the favorite must score in order to cover the spread and achieve the over. If the favorites estimated points is SMALLER than the result in this cell, then only the spread and under can be achieved.

"Dog pts": This cell shows the maximum number of points that the dog can score if the favorite scores EXACTLY what is in the "Fav req pts" cell.

"Covered by": This cell shows how many points the favorite covered the spread by, given the two scores in the previous cells. It is an easy way to check and see if indeed the spread was covered.

"Total pts": This cell shows the total points scored given the two scores in the previous cells. It is an easy way to check and see if indeed the over is achieved.

"Total OV": This cell shows the total number of final scores where the favorite covers the spread and the over is achieved.

"Total UN": This cell shows the total number of final scores where the favorite covers the spread and the under is achieved.

If you find any errors with this calculator, please let me know. If it turns out to be a useful tool, I will make other ones like it for all variations of the spread and total with or without hooks.

[SteveRyan]

Any thoughts on this calculator?

I was using it today and it seems to be pretty useful.

Lets take for example the match-up of New Orleans vs San Francisco

New Orleans -3.5
San Francisco 47

Many people might decide to bet New Orleans -3.5 with the over. However, if New Orleans manages to score only 27 points, the calculator then shows that there are only 3 possible ways for the Saints to cover the spread and hit the over!

27 - 23
27 - 22
27 - 21

However, there are 19 different ways to cover the spread and hit the under.

27 - 19
27 - 18
27 - 16
etc...down to
27 - 0

If New Orleans managed to score 31 points, there are then 11 ways to cover and hit the over, and 15 ways to cover and go under. At first glance, it seems that the under might be the better play, but consider this:

Several of those unders include blow-out scores where San Fran scores very low.

31 - 0
31 - 2
31 - 3
31 - 4
31 - 5
31 - 6
etc....up to
31 - 15

So if you really think that New Orleans is gonna score 31 points, then San Fran better not score more than 15 or you lose. Not likely San Fran will score that low....so the over would be the better pick.

I think the calculator might be more helpful if it displayed the actual final scores possible so that you can have a little more information to judge which way to go.

[mathdotcom]

Dogs and Unders
Favs and Overs

Both are correlated and betting otherwise is dumb unless you're circumventing limits or something 'unusual'.

Quiz:
If the Saints were -46 what would you bet with that total?

[SteveRyan]

Dogs and Unders
Favs and Overs

Both are correlated and betting otherwise is dumb unless you're circumventing limits or something 'unusual'.

In the AFC this season, in all games where the dog won and covered the spread, 42% of them went over. 3 pushed.

So it's not as cut and dry as you make it sound.

[RickySteve]

What a silly and irritating thread.

[SteveRyan]

What a silly and irritating thread.

Typical statement coming from someone who posts an overwhelming amount of crap on this forum.

[mathdotcom]

In the AFC this season, in all games where the dog won and covered the spread, 42% of them went over. 3 pushed.

So it's not as cut and dry as you make it sound.

Exactly dummy.
Most of the time when the dog covered the game went UNDER. 58% of the time.

ie. bet dogs and unders

[SteveRyan]

Exactly dummy.
Most of the time when the dog covered the game went UNDER. 58% of the time.

ie. bet dogs and unders

Again, it's not as cut and dry as you make it sound. If the ratio was 70%-30% then fine...I see what you are talking about. But seeing how the ratio is nearly half (48% including pushes), I think that with a little more analysis you could pick off a few that you would normally just pass.

Why would you be opposed to analysis like this anyway?? If you have a lean on the dog or the favorite and an accurate projection of points, why would you not want to know the number of ways it can hit on either side of the total? Are you just gonna rely on a blanket concept and hope that it ends up over 50%?

[mathdotcom]

Your analysis is wrong because, among other reasons, you haven't attached probabilities to the likelihood of each 49ers score. With that spread and total the market is basically guessing the Saints score 25 and the 49ers score 22.

Think of a normal distribution around the market's guess of 49ers scoring 22. The probability of the 49ers scoring 22 is greater than the probability of the 49ers scoring 21, 20, 19, etc., all the way down to 0.

So those scores you count on 'unders', like 27-0, 27-1, 27-2 should be discounted because they are less likely to occur than scores like 27-22 or 27-21.

[In the above analysis I am assuming NFL scoring to be essentially continuous, which ignores the fact that scores like 22 are unlikely to occur.]

A better way for to think about these parlays is to say: suppose I like the Saints -3.5, then if that cashes, it means the Saints offense probably did well, and if the Saints offense did well then the Saints probably put up a lot of points, and so the game was probably high scoring. If this doesn't make sense to you suppose the total were still 47 and the Saints were -46.

[SteveRyan]

So those scores you count on 'unders', like 27-0, 27-1, 27-2 should be discounted because they are less likely to occur than scores like 27-22 or 27-21.

I did mention in my 2nd post that "Several of those unders include blow-out scores where San Fran scores very low." "Not likely San Fran will score that low...."

While there is no actual probability attached to each of San Frans possible scores, we can at least eliminate the ones that are most unlikely to occur and narrow things down a little bit.

A better way for to think about these parlays is to say: suppose I like the Saints -3.5, then if that cashes, it means the Saints offense probably did well, and if the Saints offense did well then the Saints probably put up a lot of points, and so the game was probably high scoring. If this doesn't make sense to you suppose the total were still 47 and the Saints were -46.

I have a hard time buying this idea because that's like saying anytime a team covers -3.5 points it means that they played well and it was a high scoring game....which isn't always the case.

Anyway, thanks for the input...rather than just ROFLing and calling me a dummy.

I did however recalculate the dog covering the spread and going over.

As it turns out, in the AFC, the dog covered and went over 31 times (48%) and under 33 times (51%) (Thanks to Kansas City covering and going under 8 times...which was about 3 times the average). Earlier, I had only counted those where the dog won the game....so obviously the spread was covered in all those games.

[mathdotcom]

Your numbers are saying when the dog covers it is more likely that the game went under, which is consistent with what I've been saying.

You're right the 3.5 correlation is very low. It is very possible the Saints can win a 14-10 or 20-14 game etc etc.

That's why the closer the spread and total are the correlation increases.

Your estimates are very rough calculations; please do not take those %s as anything accurate, and PLEASE do not bet based on them.

[mathdotcom]

However you are correct on one thing... your avatar