Newbie question about modeling

[UntilTheNDofTimE]

Sorry if this has been asked but i couldnt find it. I've posted here for a number of years but i havent sport bet in just as many years. From time to time i see people posting about their models and here's what ive collected.

You quantify information into a spreadsheet and give it a value blah blah and it spits out a pointspread. You then back test information to see if your model is a winner. You then place bets when your model spits out a number that is giving you a edge over the market.

Isnt this modeling that you do the same model that lines makers use to create the line? If so do all professional modelers believe they can create a model better than the linesmakers?

When your model spits out a number better than the market opener what does that mean? That you have a more accurate model or that the linesmen are shading their actual line?

Basically what my question is if your model spits out 400 plays a year is this because your model is more accurate or are you just betting the 400 games a year where the book is shading the line/ taking a position on the game.

[Waterstpub87]

Depending on your belief about the objectives of the Odds makers, you could have one of two answers.

1. If you believe the objective of the odds makers is to come up with the most accurate line, forgoing a balanced marketplaces, you will are attempting to out model the odds makers. Your model will most likely deviate from their model, and you are attempting to have more accurate predictions that the odds makers.

2. If you believe the odds makers are trying to come up with the number that will maximize the balance of action, and reduce the risk of the books, you are trying to out pick the public. Much simpler a task.

If you are consistently coming up with numbers that differ significantly from the line (everyone of your predicted spreads in points off in NBA) you are most likely doing something wrong. In the past, modeling NBA and NFL totals, 65% or so of my projected totals were within
3 points of the opener. I only bet games which I found to be off more then 5 points, so it hardly spit out 400 games a year. Baseball would be different.

[UntilTheNDofTimE]

Depending on your belief about the objectives of the Odds makers, you could have one of two answers.

1. If you believe the objective of the odds makers is to come up with the most accurate line, forgoing a balanced marketplaces, you will are attempting to out model the odds makers. Your model will most likely deviate from their model, and you are attempting to have more accurate predictions that the odds makers.

2. If you believe the odds makers are trying to come up with the number that will maximize the balance of action, and reduce the risk of the books, you are trying to out pick the public. Much simpler a task.

If you are consistently coming up with numbers that differ significantly from the line (everyone of your predicted spreads in points off in NBA) you are most likely doing something wrong. In the past, modeling NBA and NFL totals, 65% or so of my projected totals were within
3 points of the opener. I only bet games which I found to be off more then 5 points, so it hardly spit out 400 games a year. Baseball would be different.

Thank you. That was the answer i was looking for. Thanks again.

[EXhoosier10]

Depending on your belief about the objectives of the Odds makers, you could have one of two answers.

1. If you believe the objective of the odds makers is to come up with the most accurate line, forgoing a balanced marketplaces, you will are attempting to out model the odds makers. Your model will most likely deviate from their model, and you are attempting to have more accurate predictions that the odds makers.

2. If you believe the odds makers are trying to come up with the number that will maximize the balance of action, and reduce the risk of the books, you are trying to out pick the public. Much simpler a task.

If you are consistently coming up with numbers that differ significantly from the line (everyone of your predicted spreads in points off in NBA) you are most likely doing something wrong. In the past, modeling NBA and NFL totals, 65% or so of my projected totals were within
3 points of the opener. I only bet games which I found to be off more then 5 points, so it hardly spit out 400 games a year. Baseball would be different.

Nice response, waters. This pretty much covers it.

[HeeeHAWWWW]

Depending on your belief about the objectives of the Odds makers, you could have one of two answers.

1. If you believe the objective of the odds makers is to come up with the most accurate line, forgoing a balanced marketplaces, you will are attempting to out model the odds makers. Your model will most likely deviate from their model, and you are attempting to have more accurate predictions that the odds makers.

2. If you believe the odds makers are trying to come up with the number that will maximize the balance of action, and reduce the risk of the books, you are trying to out pick the public. Much simpler a task.

Nice post. I'd slant it slightly by saying that opening lines are closer to 1 (with an eye on 2 in certain high-interest matches), while closing lines are closer to 2. For backtesting purposes it can be quite important to have both.