POLL: Does a line reflect probability?

[WendysRox]

Example 1: Atlanta Braves are playing the Houston Astros. The lines are: Atlanta -128 and Houston +118. Does this mean Atlanta is more likely to win? If so, at what probability? Explain.


Example 2: I'm going to flip a coin. The lines are: Heads -128 and Tails +118. Does this mean that Heads is more likely to occur? If so, at what probability?

I'm interested in people's opinions. Please let me know what you think.

In my opinion, once you establish that lines have nothing to do with probability, then you can throw out all theories involving that premise.

[djiddish98]

Why would the lines have nothing to do with probability? How would you establish that? Explain.

[bztips]

Given the OP's prior uninformed rants about the same topic, I guess he's just trying to stir the pot; but he's not going to convince many regulars in the think tank. Suggestion to Wendy -- you'll be much more at home in the Player's Talk forum.

[lasker]

[WendysRox]

Excellent question, dj. I'm glad you asked. To establish how lines have nothing to with probability, we need to understand that a linesmaker's goal is to have as close to equal action on either side of a bet. If done properly, he collects a vig, independent of that event's result (otherwise, he'd not be a Linesmaker, but a gambler). How does he get equal action? By fluctuating the line. I propose that an event's probability does not fluctuate based on public opinion. Therefore, a line cannot, by definition, be any indicator of the probability of an event, but only of the public perception of the probability of an event.

I'll pose it to you as a question... How can a fluctuating public perception be used as a determining factor of the probability of an event?

[WendysRox]

Whatever, guys. If I'm such a complete moron, explain to me how a bookie's line can determine probability.... I suppose next you are going to tell me that a company's stock price determines it's true value.

[Justin7]

Wendys,

The line is a reflection of what the best mathematician/sports bettors think.

There are two kinds of sports bettors:
1. Those that win (1%)
2. Everyone else.

Those that win in the long run tend to bet more and more, and their winnings accumulate. everyone else's bets might fluctuate, but they aren't getting bigger in the long run.

Ultimately, the winning bettors are the ones controlling the line. If Atlanta closes at -128/+118, it is because no winning bettor wants to bet at -128 or +118. At that point, the true odds of Atlanta winning that game are somewhere between -128 and -118.

The public is almost irrelevant. They merely contribute money to the pool, which feeds sportsbooks and professional players.

[WendysRox]

Wendys,

The line is a reflection of what the best mathematician/sports bettors think.

There are two kinds of sports bettors:
1. Those that win (1%)
2. Everyone else.

Those that win in the long run tend to bet more and more, and their winnings accumulate. everyone else's bets might fluctuate, but they aren't getting bigger in the long run.

Ultimately, the winning bettors are the ones controlling the line. If Atlanta closes at -128/+118, it is because no winning bettor wants to bet at -128 or +118. At that point, the true odds of Atlanta winning that game are somewhere between -128 and -118.

The public is almost irrelevant. They merely contribute money to the pool, which feeds sportsbooks and professional players.

Thanks for a non-judgmental response. But, I tend to disagree that the public is irrelevant and that the sharps set the line. But, seeing that you are posting this type of thing and apparently you own/run this forum, I couldn't imagine getting anyone to disagree with you, right or wrong.

Fair enough. I guess we'll have to agree to disagree regarding lines and probability.

[sharpcat]

Whatever, guys. If I'm such a complete moron, explain to me how a bookie's line can determine probability.... I suppose next you are going to tell me that a company's stock price determines it's true value.

If you are betting a -110 line you need to win 52.38% to break even if you are betting +140 you need to win 41.67% to break even.

If you disagree that these numbers represent a close approximation to the probability of outcome of an event so be it but I do not understand your need to argue it.

since you did not actually post a poll here are the results so far:

think tank: 5
you: 1

[lasker]

If the closing lines didn't reflect probability with a high degree of accuracy, then beating sports betting would be a piece of cake.

Just check a big enough database and see what percentage of bets at -110 hit, then at -120, then at -130, and so on. If the line is not usually a reflection of probability, you'll need to find another reason for the correlation (besides coincidence).

[wrongturn]

Yes, most of time the line won't reflect the true probability. The hard part is to know which side is off. It does not really matter to the books if you can't win long term.

[JustinBieber]

Make that 6, we can agree to disagree Wendys but then that would just make you wrong.

[WendysRox]

Taken from nflpickles.com (link:http://www.nflpickles.com/2007/09/ve...ing-point.html)

Is Vegas getting better in predicting the point spread? In short, no, the following table shows the mean and standard deviation from 1992 to 2006.




Taken from therx (link: http://www.therxforum.com/images/sta...n/post_old.gif)



Maybe this info will help --

In the NFL, just throw out the spread ... It rarely matters.

As a matter of fact, I have tracked this for many, many years and the point spread comes into play an average of just 17% of all NFL games. In other words, the S/U winner covers 83% of all NFL games .... Every year, without fail. Even in the preseaon.

For example: This past preseaon, the S/U winners went 55-5-5 ATS. In other words, if you picked all the S/U winners this past preseason, you cashed an amzing 92% of your tickets.

Sounds impossible, right? Go check it out.

Currently, through Week 3 of the current NFL season, the S/U winner is 41-7 ATS (30-2 ATS the past two weeks & 11-5 Week 1).

This is not an abnormal ocurrence. Year in and year out 82-83% of ALL S/U NFL winners cover the spread. If you like a dog in the NFL, you should only bet them if you feel they can win the game.






..... My point being, the spread rarely matters....

[WendysRox]

If the closing lines didn't reflect probability with a high degree of accuracy, then beating sports betting would be a piece of cake.

Just check a big enough database and see what percentage of bets at -110 hit, then at -120, then at -130, and so on. If the line is not usually a reflection of probability, you'll need to find another reason for the correlation (besides coincidence).

POST #13 seems to go against your idea that line = probability

[WendysRox]

further....



I am not a bookmaker nor do I intend to be. Hence, I do not know exactly how they arrive at the point spread. I do know that their intentions are not to predict the true point spread but to separate the population in half. This way, the bookmaker can guarantee 5% profit regardless of the outcome of the game (10% from half of the betting population).


(again from nflpickles.com)

[djiddish98]

With 35.15% of spreads within the pick-3 range, and 76.32% of the spreads within the pick-7 range (based on my crappy data), plus the fact that most football scoring occurs in 3s or 7s, you'd expect a team that covers to spread to also win.

Aren't MLs for football dervied from the spread? So the ML is going to take into account these probabilities anyway. Unless a book forgets to move its ML after moving its spread, there's no real advantage here.

Plus, this ignores the fact that the vig is likely higher on MLs than spreads (I believe), so you're going to most likely pay more than necessary to get the straight up winner as opposed to just the spread winner.

[WendysRox]

is there a ::crickets:: smiley?

[djiddish98]

further....



I am not a bookmaker nor do I intend to be. Hence, I do not know exactly how they arrive at the point spread. I do know that their intentions are not to predict the true point spread but to separate the population in half. This way, the bookmaker can guarantee 5% profit regardless of the outcome of the game (10% from half of the betting population).


(again from nflpickles.com)

If you got a bunch of experts in a room and debated over what the spread should be for a specific game, don't you think you'd arrive at a pretty good guess once you took everyone's insights into consideration?

Bookmakers don't "create" the final point spread. The market essentially does. The bookmaker is just trying to keep up.

[sharpcat]

further....



I am not a bookmaker nor do I intend to be. Hence, I do not know exactly how they arrive at the point spread. I do know that their intentions are not to predict the true point spread but to separate the population in half. This way, the bookmaker can guarantee 5% profit regardless of the outcome of the game (10% from half of the betting population).


(again from nflpickles.com)

Why do all books across the board offer nearly identical lines than?

If this was how it worked don't you think that a few books would be taking more action on a different side than others?

If this was the case it would be impossible for every book to offer the same number.

A bookmakers best move is to set the line to where the no-vig price closely reflects the true win probability. This way it is like flipping a coin at +110 since he is playing a game with a 50% win probability and he only needs to win 47.62%.

How do you think local bookies profit long term when they take action from 1 player on 1 side of a game for $5,000? The answer is "THEORETICAL HOLD"

[jgilmartin]

Wendys,
Do you acknowledge that players who beat the closing line a large majority of the time will be long-term winners?

[djiddish98]

POST #13 seems to go against your idea that line = probability

Finally, if there is a very small theoretical difference between the winner of the spread and the ML, then why does the ML price != the spread price? Don't you think, MAYBE, there's some adjustment due to....probabilities?

Also, care to explain why, taking a decent sized sample, the dog covers roughly 50% of the time?

[dogman]

I compare sports betting and the racing parimutuelsas all in the same. The public and sharps together make the final odds. The favorite to the longest priced horse. The favorite usually looks like the winner and the longest odds looks like a last place finisher. The odds in the LONG RUN will reflect this. The shorter odds animals will win the most and the odds get longer from there. The trick is to find those dogs or horses in certain races where the odds don't match your opinion on what the track odds dictate. The favorite doesn't always take first and the longest shot on the board does win sometimes and all those horses in between finish first sometimes also.

So it's the same thing with sports betting. Find those teams where there is value in the odds.

[WendysRox]

With 35.15% of spreads within the pick-3 range, and 76.32% of the spreads within the pick-7 range (based on my crappy data), plus the fact that most football scoring occurs in 3s or 7s, you'd expect a team that covers to spread to also win.

Aren't MLs for football dervied from the spread? So the ML is going to take into account these probabilities anyway. Unless a book forgets to move its ML after moving its spread, there's no real advantage here.

Plus, this ignores the fact that the vig is likely higher on MLs than spreads (I believe), so you're going to most likely pay more than necessary to get the straight up winner as opposed to just the spread winner.

What? I thought this thread was about whether lines and probabilities are connected?

[bztips]

The stats cited in Post #13 don't do anything to refute the basic known empirical fact that a higher percentage of 10-point NFL favorites win than 7-point favorites, who in turn win at a higher rate than 3-point favorites, who in turn win at a higher rate than 3-point underdogs, etc. etc.

[WendysRox]

If you got a bunch of experts in a room and debated over what the spread should be for a specific game, don't you think you'd arrive at a pretty good guess once you took everyone's insights into consideration?

Bookmakers don't "create" the final point spread. The market essentially does. The bookmaker is just trying to keep up.

#1: Yes, I think, given a bunch of football experts, you'd reach a reasonably accurate number. What does this have to do with having equal action on each side of a wager in order to collect a vig independent of the outcome?

#2: "market"? What "market"? ....Oh, you must be referring to "the market" composed of THE PUBLIC F*ING ACTION which is what I've been trying to say for this entire thread!

[WendysRox]

Why do all books across the board offer nearly identical lines than?

If this was how it worked don't you think that a few books would be taking more action on a different side than others?

If this was the case it would be impossible for every book to offer the same number.

A bookmakers best move is to set the line to where the no-vig price closely reflects the true win probability. This way it is like flipping a coin at +110 since he is playing a game with a 50% win probability and he only needs to win 47.62%.

How do you think local bookies profit long term when they take action from 1 player on 1 side of a game for $5,000? The answer is "THEORETICAL HOLD"

Then he is a gambler, not a bookmaker. He is betting that his line is more accurate that your line. The above posts regarding the inaccuracies of NFL lines tend to disprove his theory.


And, don't patronize me with silly questions about available lines. You know as well as I do that the larger online books release their lines and everyone else follows suit. However, you can sit at your computer just like I can on Sunday morning and see lines ranging from -1 to -3.5 on one particular game. The fact is that there ARE minor line discrepancies. This convinces line shoppers to lay off at different books. So, it really does get the books more equal by just differing from the mainstream as little as 1/2 of a point, I believe.

[sharpcat]

Then he is a gambler, not a bookmaker. He is betting that his line is more accurate that your line. The above posts regarding the inaccuracies of NFL lines tend to disprove his theory.


And, don't patronize me with silly questions about available lines. You know as well as I do that the larger online books release their lines and everyone else follows suit. However, you can sit at your computer just like I can on Sunday morning and see lines ranging from -1 to -3.5 on one particular game. The fact is that there ARE minor line discrepancies. This convinces line shoppers to lay off at different books. So, it really does get the books more equal by just differing from the mainstream as little as 1/2 of a point, I believe.

OK but many bookies may find themselves only taking action on 1 side of a game.........you can not deny that this can and will happen.

How is this profitable to take 1 bet on 1 game for a book?

Please just give me 1 educated answer to this question and please do not post some silly article from a tout and expect it to hold water.

[WendysRox]

Finally, if there is a very small theoretical difference between the winner of the spread and the ML, then why does the ML price != the spread price? Don't you think, MAYBE, there's some adjustment due to....probabilities?

Also, care to explain why, taking a decent sized sample, the dog covers roughly 50% of the time?

#1: Yes, perhaps. I rarely play ML's so I don't know first hand if there is or isn't a difference in price between spreads and ML's. Perhaps this warrants further study on my part.


#2 How often does a +140 dog win outright?

[sharpcat]

#1: Yes, I think, given a bunch of football experts, you'd reach a reasonably accurate number. What does this have to do with having equal action on each side of a wager in order to collect a vig independent of the outcome?

#2: "market"? What "market"? ....Oh, you must be referring to "the market" composed of THE PUBLIC F*ING ACTION which is what I've been trying to say for this entire thread!

So you would agree that since so far in this thread 8 people disagree with your theory and nobody but yourself agrees with your theory that this is a strong indicator that you are likely wrong????

Nobody agrees with you, this thread is done

[bztips]

#1: Yes, perhaps. I rarely play ML's so I don't know first hand if there is or isn't a difference in price between spreads and ML's. Perhaps this warrants further study on my part.


#2 How often does a +140 dog win outright?

#1 This says it all, you don't know what you're talking about. You can't find a single example anywhere of a book offering the same price on a moneyline as it does on a non-zero spread for the same contest (excl. the vig of course)

#2 About 41.7% of the time.

[djiddish98]

#2 How often does a +140 dog win outright?

You would expect the +140 dog to win 100/(100+140) or ~42% of the time.

If we look at Pinnacle's moneylines for NFL preseason, we can see that this roughly equates to a spread of +3/-3. Yes, this isn't ideal, but I don't have historical moneylines along with my spread info. Someone please correct me if a +140 dog is not equal to roughly a +3 spread.

However, Lo and behold, looking at my (again, questionable) historical data, I see -3 faves winning outright at a ~58.85% clip.

How are probabilities not involved here?

[Justin7]

Thanks for a non-judgmental response. But, I tend to disagree that the public is irrelevant and that the sharps set the line. But, seeing that you are posting this type of thing and apparently you own/run this forum, I couldn't imagine getting anyone to disagree with you, right or wrong.

Fair enough. I guess we'll have to agree to disagree regarding lines and probability.

You're welcome to disagree. I hope you can support your position, and teach me something.

When a professional player sees a line that is off, he bets it repeatedly until the line is close to where it should be. The public does not have the bankroll or the will to keep firing bets sufficient to move the line, except in very rare circumstances (such as the Super Bowl, March Madness, and a rare Yankees/Lakers game).

[WendysRox]

OK but many bookies may find themselves only taking action on 1 side of a game.........you can not deny that this can and will happen.

How is this profitable to take 1 bet on 1 game for a book?

Please just give me 1 educated answer to this question and please do not post some silly article from a tout and expect it to hold water.

I disagree. Any book larger than a high schooler taking action from friends is not going to take action solely on one side of a wager. So, I guess I do deny that this can and will happen. Further, IF this ever were to happen, they'd simply lay off to another book. I know a local that lays off whenever he gets more than, say $2g lopsided within 15 minutes of kickoff. I guess that is just his personal threshold of how much he's willing to be unequal. I suppose it's like this with the larger offshore books as well, just with a higher threshold.

Now I have a headache.

BTW, nflpickles isn't a tout to my knowledge (read this: http://www.nflpickles.com/2008/05/fi...into-this.html).

Neither is the guy from therx forum, as far as I know.

[Justin7]

Moved this to PT. It might have been HTT-ish, but I think Wendy is just stirring up stuff.

In this thread, you can unleash a little more anger if you like.

[tomcowley]

I disagree. Any book larger than a high schooler taking action from friends is not going to take action solely on one side of a wager. So, I guess I do deny that this can and will happen.

LOL. If you even knew a single local, you'd know that they have an unbalanced position on most NFL games. They are gambling, but +110 beats coinflippers in the long run. Furthermore, since they accept parlays, do you really think they have balanced action on every possible parlay combination that they take? ROFL.