Tips to a Newbie

[u21c3f6]

Ok, but your making this assumption basing your consideration on the result of 100 bets, which is not the good approach, since 100 is a little sample. I have the suspect that a mistake could around, but i would like to see it in the thinking process. Sadly at the moment there's none that found something.

No, I think your analysis is incorrect. Let me try an example to see if we agree.

The odds available for the same event are as follows:

+450 / -700
+400 / -600
+300 / -550

Is it your contention that the +450 is +EV in this scenario?

Joe.

[wrongturn]

SadZizou, if you all your 100 bets were on the huge dogs, it is obvious that there is something wrong with your method. Because by statistics, some of the values should be in the favorites. Sharpcat is right that on small market bets, books usually are not sure about the true odds, that is why they set low limit and low payout ratio. So using average is not the right way.

[SadZizou]

No, I think your analysis is incorrect. Let me try an example to see if we agree.

The odds available for the same event are as follows:

+450 / -700
+400 / -600
+300 / -550

Is it your contention that the +450 is +EV in this scenario?

Joe.

Using odd converter (sorry, i use decimal odds) I get

5.5/1.14
5.0/1.16
4.0/1.18

Then I calculate the no vig lines:

(1/5.5)+(1/1.14)*100=105.9%
(1/5.0)+(1/1.16)*100=106.2%
(1/4.0)+(1/1.18)*100=109.7%


Then I calculate the implied odds (original odd*no-vig)

5.82/1.22
5.31/1.23
4.38/1.29

The I calculate the averages:

5.17/1.246
or

+417/-406.5

So +450 is supposed to be +EV, even if 3 odds is a small sample, not enough to think that the average implied odds describes the real % of the event.

[SadZizou]

SadZizou, if you all your 100 bets were on the huge dogs, it is obvious that there is something wrong with your method. Because by statistics, some of the values should be in the favorites. Sharpcat is right that on small market bets, books usually are not sure about the true odds, that is why they set low limit and low payout ratio. So using average is not the right way.

I think that you are wrong when you say that some value should be in the favourites. Value, if there is, is always in underdog. But I might be wrong.
I understand the problem that small markets can bring, but let's consider the example I made, with the an average implied odds among seven books of 6.57/1.18,. Even if the market is distorted, I don't think that the real implied odd can by much higher than 8.00, which is the best odd I can bet on.
So I suppose that in the worst (and I guess rare) scenario betting on 8.00 I'm placing a 0EV bet or just little -EV. Nothing that can justify a -50% bankroll in 100 bets.

Mumble Mumble....

[sharpcat]

Using odd converter (sorry, i use decimal odds) I get

5.5/1.14
5.0/1.16
4.0/1.18

Then I calculate the no vig lines:

(1/5.5)+(1/1.14)*100=105.9%
(1/5.0)+(1/1.16)*100=106.2%
(1/4.0)+(1/1.18)*100=109.7%


Then I calculate the implied odds (original odd*no-vig)

5.82/1.22
5.31/1.23
4.38/1.29

The I calculate the averages:

5.17/1.246
or

+417/-406.5

So +450 is supposed to be +EV, even if 3 odds is a small sample, not enough to think that the average implied odds describes the real % of the event.

You are calculating your no-vig price wrong for starters.

[Optional]

Have you got good enough records to go back and check out how you would have done looking to lay against the chalk most under the odds, instead of backing the dogs at overs?

If your system is accurate, then your "hard luck" on the 50 bet sample should be reflected as good luck betting the other side. If it isn't, I'd take that as a strong hint your system is not giving you accurate lines. If it does work that way, then the system probably merits further analysis.

[u21c3f6]

So +450 is supposed to be +EV, even if 3 odds is a small sample, not enough to think that the average implied odds describes the real % of the event.

You cannot calculate EV like that.

Now let's imagine that the odds I gave:

+450 / -700
+400 / -600
+300 / -550

are for wagering on one number on the throw of a single die. The no-vig line is +500/-500 and therefore any wager made on any line above is -EV. That is in effect what I believe you are doing.

Joe.

[SadZizou]

You are calculating your no-vig price wrong for starters.

Ok but...where is my mistake? Because if I check the odds I get from the no-vig lines I calculate:


(1/5.82)+(1/1.22)= around 100%
(1/5.31)+(1/1.23)=around 100%
(1/4.38)+(1/1.29)=around 100%

[sharpcat]

-700=87.50%
+450=18.18%
total=105.68%

87.50%/105.68%=82.791234%
18.18%/105.68%=17.2028766%

82.791234 + 17.2028766=1

As you can see on this set of numbers I came u with a no-vig line of +/- 481 where you came up with +418/-455 (a no-vig line should be the same number on both sides. So now you can see just how far your numbers were off, when dealing with lines above +/- 200 a slight miscalculation can be deadly.

[SadZizou]

-700=87.50%
+450=18.18%
total=105.68%

87.50%/105.68%=82.791234%
18.18%/105.68%=17.2028766%

82.791234 + 17.2028766=1

As you can see on this set of numbers I came u with a no-vig line of +/- 481 where you came up with +418/-455 (a no-vig line should be the same number on both sides. So now you can see just how far your numbers were off, when dealing with lines above +/- 200 a slight miscalculation can be deadly.

I came up with 417/-406. The difference is due to approssimation. But these are averages. You can't compare first odd no-vig, with the average one.

[sharpcat]

On the set of numbers I used you came up with +418/-455 or 5.82/1.22 which is wrong your 2 probabilities need to equal +100% .

If you are going to average between books I would average each books no-vig line and not each side because each book is charging different juice.

I am not sure why you asked for help if you can not admit you are wrong, you have not been very open to advice so I am done with this thread. Good luck!

[SadZizou]

On the set of numbers I used you came up with +418/-455 or 5.82/1.22 which is wrong your 2 probabilities need to equal +100% .

If you are going to average between books I would average each books no-vig line and not each side because each book is charging different juice.

I am not sure why you asked for help if you can not admit you are wrong, you have not been very open to advice so I am done with this thread. Good luck!

It's not true I don't want to admit that I'm wrong. I know that I'm wrong, but I don't know where. It's not true I'm not open to advices. I just evaluate them.
I came with a different no vig just because I did not go deep on approximations. But the logic of my calculation is exactly the same as yours.

[SadZizou]

You cannot calculate EV like that.

Now let's imagine that the odds I gave:

+450 / -700
+400 / -600
+300 / -550

are for wagering on one number on the throw of a single die. The no-vig line is +500/-500 and therefore any wager made on any line above is -EV. That is in effect what I believe you are doing.

Joe.

Ok may be I got it. The problem is when I calculate (odd1/novig)+(odd2/novig).
Since the novig line represent the gain of the book, if I calculate the same novig on the favourite odd and the underog, I'm assuming that the book decided to gain the same on both odds. By there's the chance that the book decide to go break even on the underdog and gain everything on the favourite one.
Right or stupidity?

[wrongturn]

never mind

[u21c3f6]

Ok may be I got it. The problem is when I calculate (odd1/novig)+(odd2/novig).
Since the novig line represent the gain of the book, if I calculate the same novig on the favourite odd and the underog, I'm assuming that the book decided to gain the same on both odds. By there's the chance that the book decide to go break even on the underdog and gain everything on the favourite one.
Right or stupidity?

I don't know necessarily know the reasons for the odds (and your reason is as good as any) but that is not a concern of mine. The real concern I have for what you are doing is that you appear to be playing a market that has huge spreads. In gambling (as well as in the stock market) the spread is your enemy. In general, the larger the spread, the harder to make a profit.

Joe.

[luegofuego]

Göran?

[luegofuego]

That system seems to be like the Martingale, everyone has thought of it at some point it seems...

You have been told multiple times, the problem is trying to derive a correct line from 89% payout bookies. if the correct line is +800, five books have it at +550 and one book has it at +750, your system will tell you it's a profitable bet.

[WendysRox]

An efficient market would be a market that gets a lot of action like mlb spreads and totals. In a high volume market the market will correct itself to reflect the true win probability, in small markets that do not get a lot of action the market can not adjust itself.

I would focus on certain books which offer low juice and high limits when calculating no-vig lines.

Now wait a minute. I don't mean to pick on your post and maybe I'm understanding this completely incorrectly.... BUT, "In a high volume market, the market will correct itself to reflect the true win probability AS PERCEIVED BY THE PUBLIC" is the way this should read to me. Can we not agree that spreads/money lines/etc have absolutely nothing, zip, nada to do with the probability of that event occurring, but instead has everything to do with the books wanting even action on both sides?

This is also a reason that I believe the OP can't obtain a proper win probability just from looking at lines. I'm not saying that you can't find value by some form of middling, but to use a line as a means to a win probability seems absolutely absurd to me as the line has nothing to do with probability and everything to do with the public's perception of probability.

nevermind.. it's late and I'm tired...

[That Foreign Guy]

I'm interested in this thread as one of the things I do is try to use bookie lines to try to derive the line I have to beat to be +EV. I'm hoping that my method is working because I use a different technique to OP and not just because we are on different sides of variance. I can't imagine how he's down to 50% of starting bankroll if he's following Kelly. I haven't had a bet above 2% yet.

Most of my action is on the underdogs because that's where there tends to be the greatest divergence in odds.

A random football (soccer) match I was looking at today:

Highest 1.38 5.87 10.5
Lowest 1.26 4.00 6.00
Ratio 1.10 1.47 1.75

Clearly the chance of finding a +EV bet is much higher when the range of odds offered is higher as it tends to be on the dogs. That he is only betting longshots is not in itself a bad sign.

I came with a different no vig just because I did not go deep on approximations. But the logic of my calculation is exactly the same as yours.

That seems unlikely given you came up with different results (and results that were clearly wrong as they weren't no-vig lines).

Can we not agree that spreads/money lines/etc have absolutely nothing, zip, nada to do with the probability of that event occurring, but instead has everything to do with the books wanting even action on both sides?

It's a mixture of the two - the days of the perfectly balanced book are gone IMO. Money coming in serves as a new source of information for the bookies but they are willing to accept unbalanced action in the sake of increased EV. I hear stories all the time about whether sporting events were good for certain bookies or bad for them. The only way this can be true is if they have unbalanced books.

For example, the football world cup was great for English bookies because all the lemmings bet on England at terrible odds. Why didn't the books increase the prices on other countries to balance their action? Because they knew that they were getting more value in taking England bets at 7-1 than they would in having a balanced book (which would require offering better than true odds on Spain etc and taking less action on England by only offering 5-1).

[sharpcat]

Now wait a minute. I don't mean to pick on your post and maybe I'm understanding this completely incorrectly.... BUT, "In a high volume market, the market will correct itself to reflect the true win probability AS PERCEIVED BY THE PUBLIC" is the way this should read to me. Can we not agree that spreads/money lines/etc have absolutely nothing, zip, nada to do with the probability of that event occurring, but instead has everything to do with the books wanting even action on both sides?

This is also a reason that I believe the OP can't obtain a proper win probability just from looking at lines. I'm not saying that you can't find value by some form of middling, but to use a line as a means to a win probability seems absolutely absurd to me as the line has nothing to do with probability and everything to do with the public's perception of probability.

nevermind.. it's late and I'm tired...

You would be amazed to see just how accurate the public is if you actually took time to back test their accuracy

No!!! We can not agree that moneylines and spreads have nothing to do with win probability as action is split on each side a book is simply narrowing their line down so that the no-vig line can reflect the most accurate win probability.

If moneylines and spreads have nothing to do with win probability how exactly do you know if you are getting a good price or not when making a wager? how good of a price are you getting? do you just take a wild uneducated guess when you handicap?

It may be in your best interest to stick in the "players talk" forum if you can not understand that knowing your win probability is most important when gambling.

[WendysRox]

You would be amazed to see just how accurate the public is if you actually took time to back test their accuracy

No!!! We can not agree that moneylines and spreads have nothing to do with win probability as action is split on each side a book is simply narrowing their line down so that the no-vig line can reflect the most accurate win probability.

If moneylines and spreads have nothing to do with win probability how exactly do you know if you are getting a good price or not when making a wager? how good of a price are you getting? do you just take a wild uneducated guess when you handicap?

It may be in your best interest to stick in the "players talk" forum if you can not understand that knowing your win probability is most important when gambling.

hahaha... way to sneak a jab in there at the end.... I think we'll just have to agree to disagree that lines have anything remotely to do with probability. Where I went to school, a bookie's goal was to have equal action on both sides, as to not be a gambler themselves, but to win no matter the outcome of the match. If we can't agree on this cornerstone principal, then all other discussions are pointless. Let me pose it to your simple mind in this way: if a book's purpose is not to have equal action on both sides and reap a sure profit, then is he not just a gambler with multiple outs and not a book at all?

[WendysRox]

If moneylines and spreads have nothing to do with win probability how exactly do you know if you are getting a good price or not when making a wager?


and to answer this question... I know I am getting a good price when I cap a game to be won by team A by 7 points, and i can get them at -102 at -3 points. I would never consider a released line/ml to be any indication whatsoever of an actual win probability, only a reflection of the public action.

[sharpcat]

If moneylines and spreads have nothing to do with win probability how exactly do you know if you are getting a good price or not when making a wager?


and to answer this question... I know I am getting a good price when I cap a game to be won by team A by 7 points, and i can get them at -102 at -3 points. I would never consider a released line/ml to be any indication whatsoever of an actual win probability, only a reflection of the public action.

So the fact that when paying -110 you need to win 52.38% of the time to break even means nothing????

So the fact that when paying -200 you need to win 66.67% of the time and at +200 you need to win 33.33% (notice how these 2 numbers add up to exactly 100%) this means nothing to you??????????

You really need to go back to school if you can not understand that as the bookie is balancing his book he is narrowing the "NO-VIG" line down to a close representation of each teams win probability. If team A has value the public bets it the book moves the line but moves it too far the public now bets team B and back and forth until the book is getting equal action on each side.....................

What amazes me even more is that your handicapping has absolutely no way of knowing how much of an edge you have

[sharpcat]

Honestly do you ever bet moneylines or just point spreads?

If you bet a team at -300 would this be a good bet if you have a 68% chance of winning?

[byronbb]

If moneylines and spreads have nothing to do with win probability how exactly do you know if you are getting a good price or not when making a wager?


and to answer this question... I know I am getting a good price when I cap a game to be won by team A by 7 points, and i can get them at -102 at -3 points. I would never consider a released line/ml to be any indication whatsoever of an actual win probability, only a reflection of the public action.

Must be nice find a 33% edge on an NFL spread.

[That Foreign Guy]

I would never consider a released line/ml to be any indication whatsoever of an actual win probability, only a reflection of the public action.

So if your model says a team should be +0 and the books have them at +7 you're going to assume you know more than the rest of the world and not go, "hang on a second, what did I miss?"

Bookies make most money long term when they set accurate lines, not when they balance the books (admittedly that's the low variance route). Therefore a well run bookies wants to be accurate to reality not to perception.

[wrongturn]

The market price of one game may or may not reflect the true win probability. But for a random set of games, their collective market prices are very accurate to how they perform collectively.

[WendysRox]

ok, I see what you guys are saying, I think. you are telling me that the natural flow of bets on one side or another tends to progress a line towards the actual probabilities of either team winning. Is this correct? Frankly, I've never tried to determine just how accurate NFL spreads are. And, frankly, I'm not too interested in doing so. But, I may just out of curiosity.

What I proposed, and still stand behind for the time being, is that indeed spreads are set entirely based on what will bring in equal money on both sides. I know that if I were a bookie and not a gambler, I would want to take as little risk as possible. I'd take the sure vig over my capping ability. And, this is exactly what I believe the big books do.

I think I've even heard before that in something like 65% (or 80% or whatever)of NFL games, the straight up winner covers the spread, whether they were a dog or a favorite. So, no, I don't think the spread is any indication whatsoever of actual win probability. It is solely a reflection of what Joe Public perceives the win probability to be. And, to answer the foreign guy... no I do not question my numbers. When I find a favorable spot, I assume that the public has been watching too much sportscenter and has been brainwashed into making a bad bet.

[JustinBieber]

So you are still an idiot then basically?

[sharpcat]

If Apple releases a brand new device and decides that they will sell it for $900 and nobody buys it so they reduce the price to $700 and more people buy it bot not a lot so they drop the price again to $600 and finally everybody is starting to buy the device.

Would you disagree that $600 is the most accurate reflection of the true value of this new device?

This is how every market known to man operates the public is like a gigantic consensus of millions of people who 1 by 1 are clueless but when you take a consensus from all of them they will buy and sell until the market reaches its true value.

As Wrongturn stated this does not mean that every single closing # tells you the exact win probability but that over a large sampling of games of teams at -200 the win probability is more likely than not going to reflect a win probability in the range of 66.67%.

It is funny that you state that you "have frankly never tried to determine just how accurate NFL spreads are" yet you are quick to come into a thread and attack me and tell me how wrong I am Please at least make an attempt to make an educated argument the crap you learned from Brandon Lang and J.R. Miller is not going to cut it here.

Most importantly you should realize that each half point on a spread has value and if your model is telling you that team A should be -7 but the line is -0, you should be very concerned about the accuracy of your model. considering that everybody including the books have team A at -0 at 50% and you are basically saying that Team A at a -0 has a 75% chance to win. Someone is way off on their # and it would be safe to assume that it is not the books and the millions of people who help them adjust their lines.

[sharpcat]

ok, I see what you guys are saying, I think. you are telling me that the natural flow of bets on one side or another tends to progress a line towards the actual probabilities of either team winning. Is this correct? Frankly, I've never tried to determine just how accurate NFL spreads are. And, frankly, I'm not too interested in doing so. But, I may just out of curiosity.

What I proposed, and still stand behind for the time being, is that indeed spreads are set entirely based on what will bring in equal money on both sides. I know that if I were a bookie and not a gambler, I would want to take as little risk as possible. I'd take the sure vig over my capping ability. And, this is exactly what I believe the big books do.

I think I've even heard before that in something like 65% (or 80% or whatever)of NFL games, the straight up winner covers the spread, whether they were a dog or a favorite. So, no, I don't think the spread is any indication whatsoever of actual win probability. It is solely a reflection of what Joe Public perceives the win probability to be. And, to answer the foreign guy... no I do not question my numbers. When I find a favorable spot, I assume that the public has been watching too much sportscenter and has been brainwashed into making a bad bet.

You don't sound very confident in all of your un-educated theories sounds more like a lot theories and I heard from somewhere else's and I don't feel like researching.

[That Foreign Guy]

And, to answer the foreign guy... no I do not question my numbers. When I find a favorable spot, I assume that the public has been watching too much sportscenter and has been brainwashed into making a bad bet.

So you're the only genius in a world of idiots? :P

I suggest you read a book called the wisdom of crowds. It explains how and why a consensus will often beat an expert.

[WendysRox]

Let me propose you this: Bookie has NYG -3 @ NYJ. Bookie gets $500,000 worth of bets on NYG and $100,000 on NYJ. What do they do? They move the line down to -2 in hopes of getting more equal action. Does this mean that NYG -2 is closer to what may actually occur? HECK NO! It means that THE PUBLIC PERCEIVES THIS TO BE closer to what may actually occur! That's it! It does not have ANYTHING to do with the probability of that event occurring!

Your point of view, to me, is like taking a coin flip and saying that because 60% of the public are on heads and 40% are on tails, then heads is more likely to occur. IMO, the odds have nothing to with probability, only action.

Lastly, unless something new develops, I think I've made my thoughts clear. I apologize for sounding a little frustrated, but that's what I am. Perhaps I just need someone to explain it to me better.

[jgilmartin]

Let me propose you this: Bookie has NYG -3 @ NYJ. Bookie gets $500,000 worth of bets on NYG and $100,000 on NYJ. What do they do? They move the line down to -2 in hopes of getting more equal action. Does this mean that NYG -2 is closer to what may actually occur? HECK NO! It means that THE PUBLIC PERCEIVES THIS TO BE closer to what may actually occur! That's it! It does not have ANYTHING to do with the probability of that event occurring!

If you believe that, why not just blindly bet whatever team the market moved against? If the line is Team A -10 / Team B +10, and then moves to Team A -12 / Team B +12, you'd be getting two free points on Team B, since the line move provides no indication of the actual probability, right?

[u21c3f6]

If you believe that, why not just blindly bet whatever team the market moved against? If the line is Team A -10 / Team B +10, and then moves to Team A -12 / Team B +12, you'd be getting two free points on Team B, since the line move provides no indication of the actual probability, right?

In fact, applied properly using valid criteria (not blindly), there are many times when you are actually getting two free points on Team B. It all depends on the reason the line moved.

Not all line movements are created equal.

Joe.