Is the Public Truly Wrong 75% of the time?

[IrishTim]

I bet plenty of times. My test for a good bet: can you convince an anti-gambling math professor that you have a +EV bet (even if it would take 6 hours to go through all the steps). While some people can bet on "feel", most "feel" bettors are just rationalizing a desire to gamble. If you are not using a solid mathematical approach to analyzing bets, you're very likely to lose going forward.

Are you speaking from experience?

[Firefox14]

I bet plenty of times. My test for a good bet: can you convince an anti-gambling math professor that you have a +EV bet (even if it would take 6 hours to go through all the steps). While some people can bet on "feel", most "feel" bettors are just rationalizing a desire to gamble. If you are not using a solid mathematical approach to analyzing bets, you're very likely to lose going forward.

Okay...In my post I meant, you don't bet - (minus) anything (the minus sign was difficult to see). What I am basically asking is; do you always bet with no vig/juice?

[MonkeyF0cker]

I bet plenty of times. My test for a good bet: can you convince an anti-gambling math professor that you have a +EV bet (even if it would take 6 hours to go through all the steps). While some people can bet on "feel", most "feel" bettors are just rationalizing a desire to gamble. If you are not using a solid mathematical approach to analyzing bets, you're very likely to lose going forward.

I think you should look around your place of employment once. It appears to me that most of the people who bet on feel bet because they think they'll win - not because of some gambling psychosis.

[Data]

My test for a good bet: can you convince an anti-gambling math professor that you have a +EV bet (even if it would take 6 hours to go through all the steps).

An excerpt from "Statistics Hacks" by Bruce Frey, Ph.D:

http://my.safaribooksonline.com/0596101643/170#X2ludGVybmFsX0ZsYXNoUmVhZGVyP3htbGlkPTA1OTYxMDE2NDMvMTY5

OK, Mr. Big Shot Stats Guy (you are probably thinking), you're going to tell us that we
should never play the lottery because, statistically, the odds will never be in our favor.
Actually, using the criteria of a fair payout, there is one time to play and to buy as many
tickets as you can afford.
In the case of Powerball, you should play anytime the grand prize increases to past
$146,107,962 (or double that amount if you want the lump sum payout). As soon as
it hits $146,107, 963, buy, buy, buy!

An excerpt from Elihu's D. Feustel "daringly" review of "Statistics Hacks" by Bruce Frey, Ph.D:

http://www.amazon.com/review/R3SBWH0H3JT8IH

From a gambler's perspective, this information is mandatory. If you aspire to be a serious gambler and haven't put in the time to master statistics, $30 is well spent on this book.

Yeah, how about those professors, huh? A tough crowd indeed.

[subs]

Okay...In my post I meant, you don't bet - (minus) anything (the minus sign was difficult to see). What I am basically asking is; do you always bet with no vig/juice?

i think that you are confusing to different things. +EV is positive expected value. a mathematical term which means if you made the same bet millions of times you would expect to win.
vig is the commission paid to the book. you will nearly always pay commission but can also have a +EV bet at the same time. the line +110 only gives you a % that a play should win. in this case about 48%. if you estimate you will win this play 55% then your play is +EV. this is what justin7 means.

a -110 line is about 52.3% so if you estimate 55% you are still +EV.

[dvsbmx]

Ok i think I understand the 99 % wrong thing, but then how long do you have to bet to realize your a losing bettor just hitting randomly or if your some kind of sharp who has an edge? Is it possible for someone to be a winning gambler over say a 1000 plays and still be just hitting randomly and somehow random variance has him winning say 56 % over a 1000 plays.

A sharp knows their edge, and figure it out using various methods. And yes, a winning player can hit 56% randomly over 1000 plays. If you flip a coin 1000 times, is it possible for the coin to land on tails 560 times, yes, you could have an amazing run for 1000 flips possibly even hitting tails at a higher percentage.

Its the Gamblers Fallacy just because tails hit 10 times in a row doesn't mean heads is surly going to be the next flip.

[CHUBNUT]

I know this is hard to comprehend by some of you living in fantasy land but were all part of the Betting Public.

[chipleader]

I know this is hard to comprehend by some of you living in fantasy land but were all part of the Betting Public.

This all makes sense now...No wonder i am wrong 75% of the time man WTF!

[CHUBNUT]

This is just another example of people obsessed by what others think and how it affects the market. Of course it would be complete nonsense for one to have their own opinion on a game.

[subs]

thanks firefox

[Santo]

There is actually a lot of info out there on this topic. On non-sports event prediction the public is right much more often than not especially if they are allowed to wager on the event. When they were in operation, Tradesport's real-world events %correct predictions were quite accurate.

I wonder if there wasn't some form of selection at work here. Sports betting attracts a large number of people for entertainment, your 'public'. Non-sports events (weather, politics) may only attract those to wager who have some level of knowledge on the subject.

I was speaking at an academic conference in Tampere, Finland a couple of years ago, where a senior academic was making the argument, in the context of real/virtual world convergence, that prediction markets weren't accurate because people were 'playing', and that basing public policy on them meant virtual worlds were no more virtual than the real one (poorly articulated by me, but I don't have my Mendeley database handy to check the reference).

Whilst I happen to agree with his main point, one shared by other scholars, his prediction market material was incredibly weak because he was concentrating on 'markets' established effectively by lobbyists to make a point, ignoring the markets that allow people to back their predictions with real money. We debated this at some length in the Q&A session, but I haven't seen any follow-up work that addresses it to date.

In effect I would argue that the 'free' sites will attract people multi-voting to support their political/ideological allegiance, those same people won't be 'gambling' on the outcome being the same as their allegiance, because the entertainment of "the game" isn't there.

Sorry for the tangent, but thought it might interest someone.

[gilly6864]

basically the public is picking a game based on a random selection..however i think the percentage is 49.8% that the random public is right

[wildbill044]

I wonder if there wasn't some form of selection at work here. Sports betting attracts a large number of people for entertainment, your 'public'. Non-sports events (weather, politics) may only attract those to wager who have some level of knowledge on the subject.

I was speaking at an academic conference in Tampere, Finland a couple of years ago, where a senior academic was making the argument, in the context of real/virtual world convergence, that prediction markets weren't accurate because people were 'playing', and that basing public policy on them meant virtual worlds were no more virtual than the real one (poorly articulated by me, but I don't have my Mendeley database handy to check the reference).

Whilst I happen to agree with his main point, one shared by other scholars, his prediction market material was incredibly weak because he was concentrating on 'markets' established effectively by lobbyists to make a point, ignoring the markets that allow people to back their predictions with real money. We debated this at some length in the Q&A session, but I haven't seen any follow-up work that addresses it to date.

In effect I would argue that the 'free' sites will attract people multi-voting to support their political/ideological allegiance, those same people won't be 'gambling' on the outcome being the same as their allegiance, because the entertainment of "the game" isn't there.

Sorry for the tangent, but thought it might interest someone.

I feel like the Aflac duck listening to Yogi Berra. The probability that posts like this can make the "average joe" like myself with an IQ somewhere in the 140's feel retarded is 99.5% Guess that's what I get for wandering into the think tank.

[TakeIt]

Compare it to roulette, 100% of the bets you make are -EV, yet you'll still win a side close to 50% of the time. Even if you win though, it doesn't mean it was a good bet.

That's because in roulette your decisions don't matter. In poker, sports betting, backgammon, gin rummy, etc, et al your decisions do matter.

Therefore, if a "public" makes a sports bet he can be making a correct decision even if he is guessing. He could be betting on inside information even if he doesn't know about it. He could be betting with sharps. In other words, he could be getting the best of it even though he is ignorant.

That's impossible betting on roulette, because you can never have the best of it.

[Arilou]

Justin7, I don't think it is useful to define a wrong side as a side in which laying -110 would be -EV, for several obvious reasons. That makes the vast majority of sides 'wrong' including sides in which one could profitably wager if you got low juice, and learning that a side is 'wrong' tells you very little even if you could know it. To me, a game can have at most one wrong side. Some games can be considered to have a wrong side and a right side, with right meaning profitable at low juice. Other games don't have a side.

Under Justin7's definition, the public is indeed wrong almost all the time, although 99% still seems a little high to me. I don't believe lines are as accurate as Justin7 does. However, much more interesting would be that 75% of the time the side the public chooses is worse than the other side, although of course that still doesn't justify a wager on its own.

[roasthawg]

Justin7, I don't think it is useful to define a wrong side as a side in which laying -110 would be -EV, for several obvious reasons. That makes the vast majority of sides 'wrong' including sides in which one could profitably wager if you got low juice, and learning that a side is 'wrong' tells you very little even if you could know it. To me, a game can have at most one wrong side. Some games can be considered to have a wrong side and a right side, with right meaning profitable at low juice. Other games don't have a side.

Under Justin7's definition, the public is indeed wrong almost all the time, although 99% still seems a little high to me. I don't believe lines are as accurate as Justin7 does. However, much more interesting would be that 75% of the time the side the public chooses is worse than the other side, although of course that still doesn't justify a wager on its own.

If a line is correct (a true 50% coinflip) then either side of a line that's paying juice is -EV. If that game is getting public action then they are all on the wrong side.

It's like casino games... everyone is making -ev bets pretty close to 100% of the time. In sports gambling I'd say it's closer to 70% of the time if I had to guess as there are many +EV opportunities out there imo.

[HauntingTheHoly]

Eventually, it seems as if the public would smarten up and fade more often. But then again, this may create a facade-like consensus, i.e. the public fades to fade, which then causes the unknowing fades to fade what they believe is the public, even though they don't know they're fading the original faders, which creates the exact opposite of what the faders are trying to achieve! It then becomes a faded fade. .

I use all of this information, but then I FADE it four additional times.

[TMoney33]

Most people in this forum will be dead in between 10-70 years, and even if its 70 years that is a blink of an eye. What will +EV matter at that point? My point is that a lot of people here need to start thinking long term as in eternity/where you stand with Jesus.

So in response to your point, +EV still matters. Wouldn't things like going to church/being a good person/praying/etc be +EV? EV applies to just about everything in life, not just gambling.

[EasyHustlin]

Wouldn't things like going to church/being a good person/praying/etc be +EV? EV applies to just about everything in life, not just gambling.

[Peregrine Stoop]

I'd say public is wrong on the best market line 51% of the time. Usually, they are getting much worse than the best market line.

[Justin7]

No more posts about religion in this thread. If you have a pertinent point to make relating religion and handicapping, start a new thread.

[subs]

I'd say public is wrong on the best market line 51% of the time. Usually, they are getting much worse than the best market line.

i'm prolly wrong, thought it would be a little higher. don't the books take unbalanced action on public teams?

just wondering.

[BRAVES1985]

75% ?? no

[frankthetank]

I look at volume of bets and the % sometimes. Tex AM and Texas Tech perfect examples of good games to bet ATP yesterday

[phillybul]

hell yeah, the uninformed public is lol, I laugh when people just do some stat research and expect to trump, it looking up infoyou woulndt normally think, of and taking advice from established handicappers...heres an excellent start

[Tomahawk]

Ok i think I understand the 99 % wrong thing, but then how long do you have to bet to realize your a losing bettor just hitting randomly or if your some kind of sharp who has an edge? Is it possible for someone to be a winning gambler over say a 1000 plays and still be just hitting randomly and somehow random variance has him winning say 56 % over a 1000 plays.

It is possible to win 56% of the bets over 1000 plays by betting randomly.

Let's calculate the chance for this to happen:

1. Let's asume a random player has a 50% chance to win a bet, so winning or losing has an equal possibility.

2. So we have to calculate the possibility of winning minimum 560 bets out of 1000 bets, this can be 560 wins, 561 win, ..., 1000 wins out of 1000 bet, we have to thing of everything that is theoretically possible.

3. So we have to count how much times can this happen (hit more then 56%) and devide it by all of the possibilities.

So chance to hit 56% out of 1000 bets with a 50% win rate=favorable possibilities/all of the possibilities

4. I think till this point it was easy to understand now comes the geeky part. How to calculate this?

all of the possibilities=2^1000 (this mean 2*2*2*....*2 1000 times)

favorable possibilities=SUM(i=560 to 1000) [ 1000! / (i!*(1000-i)!) ]

(! means factorial for example 3!=3*2*1)
(For example 1000! / (560!*(1000-560)!) is the number of possibilities to win exactly 560 bets out of 1000, so that's why I had to use the i and the SUM)

So the chance for a random bettor to hit 560 or more out of 1000 bets is:
SUM(i=560 to 1000) [ 1000! / (i!*(1000-i)!) ] / 2^1000

If you have time you can calculate this using excel. If you don't have even chance to win a bet or lose a bet then the above calculation is a bit more difficult.

If you are a pro bettor and place a lot of +EV bets and you have an average winning rate of 55% then your chance to win a minimum of 56% from 1000 bets is much more higher then for a bettor that has a winning rate of 50%.

[That Foreign Guy]

hell yeah, the uninformed public is lol, I laugh when people just do some stat research and expect to trump, it looking up infoyou woulndt normally think, of and taking advice from established handicappers...heres an excellent start

If I ran a book I would give all new accounts a free subscription to John Morrison.

[sideloaded]

It is possible to win 56% of the bets over 1000 plays by betting randomly.

Let's calculate the chance for this to happen:

1. Let's asume a random player has a 50% chance to win a bet, so winning or losing has an equal possibility.

2. So we have to calculate the possibility of winning minimum 560 bets out of 1000 bets, this can be 560 wins, 561 win, ..., 1000 wins out of 1000 bet, we have to thing of everything that is theoretically possible.

3. So we have to count how much times can this happen (hit more then 56%) and devide it by all of the possibilities.

So chance to hit 56% out of 1000 bets with a 50% win rate=favorable possibilities/all of the possibilities

4. I think till this point it was easy to understand now comes the geeky part. How to calculate this?

all of the possibilities=2^1000 (this mean 2*2*2*....*2 1000 times)

favorable possibilities=SUM(i=560 to 1000) [ 1000! / (i!*(1000-i)!) ]

(! means factorial for example 3!=3*2*1)
(For example 1000! / (560!*(1000-560)!) is the number of possibilities to win exactly 560 bets out of 1000, so that's why I had to use the i and the SUM)

So the chance for a random bettor to hit 560 or more out of 1000 bets is:
SUM(i=560 to 1000) [ 1000! / (i!*(1000-i)!) ] / 2^1000

If you have time you can calculate this using excel. If you don't have even chance to win a bet or lose a bet then the above calculation is a bit more difficult.

If you are a pro bettor and place a lot of +EV bets and you have an average winning rate of 55% then your chance to win a minimum of 56% from 1000 bets is much more higher then for a bettor that has a winning rate of 50%.

Thanks for the math. Really puts it all in perspective.

[Cookie Monster]

75% of the percentages posted in this thread are just guessed out of thin air.

[BeatingBaseball]

"While some people can bet on "feel", most "feel" bettors are just rationalizing a desire to gamble. If you are not using a solid mathematical approach to analyzing bets, you're very likely to lose going forward." - Justin7

Although I appreciate and respect your highly mathematical approach, Justin - I think you underestimate the number of guys who can make +EV calls in sports with a very high level of success against the number without employing a methodic, math model approach. I've known many over the years who seem to just have an innate talent for it absent any apparent detailed, mathematic calculation. They simply know the sport and the teams - and have a remarkable 'gestalt' type gift for analyzing the matchups by "feel." I'm sure there's some math operating in there somewhere with such cappers - but it's sure not formulaic and they certainly could never explain/articulate their process to anyone.

Unfortunately - in my experience - although these types are extraordinarily good cappers in terms of making individual selections - many of them are not good gamblers. They often have other holes, leaks, money mgmt problems, play parlays and in the overall end up losing despite their talent.

[bztips]

It is possible to win 56% of the bets over 1000 plays by betting randomly.

Let's calculate the chance for this to happen:

1. Let's asume a random player has a 50% chance to win a bet, so winning or losing has an equal possibility.

2. So we have to calculate the possibility of winning minimum 560 bets out of 1000 bets, this can be 560 wins, 561 win, ..., 1000 wins out of 1000 bet, we have to thing of everything that is theoretically possible.

3. So we have to count how much times can this happen (hit more then 56%) and devide it by all of the possibilities.

So chance to hit 56% out of 1000 bets with a 50% win rate=favorable possibilities/all of the possibilities

4. I think till this point it was easy to understand now comes the geeky part. How to calculate this?

all of the possibilities=2^1000 (this mean 2*2*2*....*2 1000 times)

favorable possibilities=SUM(i=560 to 1000) [ 1000! / (i!*(1000-i)!) ]

(! means factorial for example 3!=3*2*1)
(For example 1000! / (560!*(1000-560)!) is the number of possibilities to win exactly 560 bets out of 1000, so that's why I had to use the i and the SUM)

So the chance for a random bettor to hit 560 or more out of 1000 bets is:
SUM(i=560 to 1000) [ 1000! / (i!*(1000-i)!) ] / 2^1000

If you have time you can calculate this using excel. If you don't have even chance to win a bet or lose a bet then the above calculation is a bit more difficult.

If you are a pro bettor and place a lot of +EV bets and you have an average winning rate of 55% then your chance to win a minimum of 56% from 1000 bets is much more higher then for a bettor that has a winning rate of 50%.

Uh, ok. You know, you could just compute a z-score, much easier:

Given p=.5 and n=1000:

z = (x-mu)/sigma

x = 560
mu = 500 (= n*p)
sigma = sqrt(n*p*(1-p)) = 15.81

Given the above, z = 3.79, which means you're 3.79 std. deviations away from the mean of 500 -- that's very far away, meaning you have less than a .01% chance of hitting 560 or more bets out of 1000.

[Sawyer]

The public is wrong about 99% of the time. Fortunately for them, they win close to half their bets when they are wrong.

LoL. Well said Justin.

[the shadow]

Is +ev a fact or an opinion??

[Tomahawk]

Uh, ok. You know, you could just compute a z-score, much easier:

Given p=.5 and n=1000:

z = (x-mu)/sigma

x = 560
mu = 500 (= n*p)
sigma = sqrt(n*p*(1-p)) = 15.81

Given the above, z = 3.79, which means you're 3.79 std. deviations away from the mean of 500 -- that's very far away, meaning you have less than a .01% chance of hitting 560 or more bets out of 1000.

Yeah, this is a much more easier calculation, you can read more details about this in this book:

http://store.sbrforum.com/products/300-CONQUERING-RISK

I think this is the most detailed sportsbetting book when it comes to math.

[Wrecktangle]

hell yeah, the uninformed public is lol, I laugh when people just do some stat research and expect to trump, it looking up infoyou woulndt normally think, of and taking advice from established handicappers...heres an excellent start

Here's my religious quote for Justin7:

Whenever considering whether a handicapper is on the level, if you see any advertising photos from them of 1) partially dressed women 2) sports cars 3) handfuls of cash.

God has just sent you three signs that this handicapper is crap.

Learn some stat and do it yourself (self handicapping) you'll be much better off in the rest of your life having learned some stat.