[usma1992]

My degree is in electrical engineer so I have a math background but not good enough. Could use Insight...

I believe that I have a statistical model that can predict 58% of the time or great what the total score of two college football teams would be +/-7 points. For instance, on the North Texas/La Lafayette game my model came up with 47 points and the final score was 53 points. Vegas had the over under at 56 and so my model said to take the under.


Here is my question: If I truly am able to predict the outcome at a 58% certain using the standard line, using a teaser, how much does my probability go up and does it make sense? i.e. Of the 55 games this weekend my model says to bet 25 games because most of the are neutral. Tonight it says bet the Houston/SMU over 60 points. Which means my model is at least predicting 68 points or more. If I tease that 60 down to 53 points... what percentage do I have to hit to make up for the difference 65% or 70%...

I am not sure I worded the question, correctly... but I appreciate any response...

[uvarunthetable]

I'm confused about the accuracy of your model-- Are you saying that 58% of the time you within 7 points (either + or -) of what the actual score ends up being?

[InTheRed]

With teasing, you're going to need to add something to the bet. I am not sure of any book that will offer a 1 team teaser.

By adding a 2nd item to the bet, you're EV is going to drop as will your 58% mark.

There are books that will allow you to buy points to help your number. The problem is that the juice on getting that number to move several pts is going to kill you in the long run.


Further, if you're hitting at 58% and your model is giving you half of the games in a week to play, I wouldn't even bother with a teaser. If it is a legit 58% over time and your money management is sound, you'll be very happy by the end of the season. I'd kill for a 58% hit rate.

[mcduggly]

Red is right, you can't tease a game by itself, you have to pair it with something. With that said, it's going to be harder to win 58% of your pairs even with the added help in the form of a teaser due to the sole fact you now have to win 2 games.

[LtDementia]

Red is right, you can't tease a game by itself, you have to pair it with something. With that said, it's going to be harder to win 58% of your pairs even with the added help in the form of a teaser due to the sole fact you now have to win 2 games.

If memory serves me right, you need about a 73% win rate on each leg of a 2 team teaser to be at breakeven.

[LT Profits]

With teasing, you're going to need to add something to the bet. I am not sure of any book that will offer a 1 team teaser.

By adding a 2nd item to the bet, you're EV is going to drop as will your 58% mark.

There are books that will allow you to buy points to help your number. The problem is that the juice on getting that number to move several pts is going to kill you in the long run.


Further, if you're hitting at 58% and your model is giving you half of the games in a week to play, I wouldn't even bother with a teaser. If it is a legit 58% over time and your money management is sound, you'll be very happy by the end of the season. I'd kill for a 58% hit rate.

Red is right, you can't tease a game by itself, you have to pair it with something. With that said, it's going to be harder to win 58% of your pairs even with the added help in the form of a teaser due to the sole fact you now have to win 2 games.

He would obviously tease 2 (or more) of his totals. Unfortunately, the more pressing issue is that 58% is not nearly good enough for a teaser leg. For a 7-point tease at -130, you need over 75% per leg.

[usma1992]

on a 3 game teaser at +120 to be equivalent to -110 at 58%. If you look at every event as singular event, and than you take into account the money.

The question really becomes by adding the teaser to the already 58% ratio... can I raise the percentages to 63%+...

I agree 100% that if I am hitting 58% on half of the games played... I'm in good shape but I'm always looking for addition advantages...


Dave

[VLR100]

If you're hitting -110 bets at 58% with your model over large samples, don't bother with teasers. You're sitting on a supply of endless money. If you want to see what your winning percentage would be like with the extra teaser points you'll need to look at the data.

[raiders72001]

Need each leg 72.37% at -110.

[Dom177]

I think trying to make up a system or model based around teasers is a recipe for disaster. I agree with VLR, if you are coming up with positive results (I dont know how many games you have back tested or have bet using this) I'd stay away from the teasers. Good luck!

[illfuuptn]

Let's also address the issue that he's talking about teasing TOTALS--College Football totals at that! And I don't really understand...are you saying your model wins 58% or that 58% of the time the actual total falls within 7 points of your prediction? As for your question you'll need to use historical half-point values and push probabilities to determine that because scoring distributions aren't normal. Just a guess, but I bet a college game with a total of 48 is more likely to go 20 points over or under than an NFL game with the same total.

[A4K]

Let's also address the issue that he's talking about teasing TOTALS--College Football totals at that! And I don't really understand...are you saying your model wins 58% or that 58% of the time the actual total falls within 7 points of your prediction? As for your question you'll need to use historical half-point values and push probabilities to determine that because scoring distributions aren't normal. Just a guess, but I bet a college game with a total of 48 is more likely to go 20 points over or under than an NFL game with the same total.

I think he/she is stating that 58% of the time his model can predict the total within 7 points of the final score.

[HedgeHog]

I think he/she is stating that 58% of the time his model can predict the total within 7 points of the final score.

No, I believe he's saying that he can win 58% of his straight bets when his model reveals a 7+ point advantage versus the posted Total. Personally I think he's delusional, but on the remote chance he's right, he would choose to ruin this +Ev opportunity by teasing his system plays instead of straight betting them. Teasing college totals is a horrible waste of money.

[LT Profits]

I think he/she is stating that 58% of the time his model can predict the total within 7 points of the final score.

No, I believe he's saying that he can win 58% of his straight bets when his model reveals a 7+ point advantage versus the posted Total. Personally I think he's delusional, but on the remote chance he's right, he would choose to ruin this +Ev opportunity by teasing his system plays instead of straight betting them. Teasing college totals is a horrible waste of money.

I interpreted the same way A4K did, which is why I mentioned the -130 7-point tease.