-
Comment -
arguing against value is like saying a casino shouldn't have taken an individual's bet(s) if they somehow came up ahead in a -EV game.Comment -
Joe,One estimate can be derived from closing ATS lines which are amazingly accurate at approx a 50-50 chance of one side or the other winning. If you were to wager on all games, you would probably come very close to 50-50 unless you were doing something terribly wrong (which also means you might have the seeds to do something very right). However, if you only wager on a subset of games that meets a validated criteria and those games give you a validated 55% chance of winning your wager, which estimate is correct for that one particular game, the 50-50 or the 55-45? As I wrote in previous posts, now imagine another player using different validated criteria but selects the opposite team as the 55% chance, Which estimate is correct for this particular game? The point is, that they are all correct, in context. There are no hard probabilities for sports events like there are for a coin or die etc. and therefore an estimate has to based on a sampling of games and not just one particular game. There is no one single % that one can place on a single sporting event without backing that up with a sample and then dependent on the criteria used, you will come up with different %'s. The only hard probabilities for a sporting event are after the game when you know that one wager had a 100% chance of winning and the other side had a 0% chance of winning. I think some believe that only one side of a wager in a single event can have value and that is simply not true without context.
+EV for a sporting event simply means (to me) that you have a validated expectation of collecting more on your winning wagers than you lose on your losing wagers. It is not about "picking" winners.
If you could always pick the winner, then of course that is a solution. However, the moment you make a wager and it does not win, then probabilities and odds come into play. And the question that needs to be answered prior to making a wager (assuming you are trying to "win" money) is: If I make this same type of wager many times, will I collect more on my winning wagers than I will lose on my losing wagers. The use of validated data is very helpful in answering that question.
I hope I have been clear and I want to thank those that are helping to make this a good discussion.
Joe.
I know we're fighting the same battle here, but let's look at the definition of EV... EV is roughly the probability an event would occur if it were played out an infinite number of times.
a +EV wager is a wager in which your return is expected to be positive.
All games have an expected value, and I agree with you that it is unknown, and can only be estimated. HOWEVER, that does not mean that a true EV does not exist. If the game were to be played in an infinite amount of universes at the same time, we could "know" the EV. We could then compare that EV to our wager, and determine if our wager was indeed +EV or -EV.
Since we can't do that, the best we can do is estimate the EV using past data, however this is just an estimation. Two people may estimate they have +EV on a single game [assuming the same odds,market efficiency, etc etc], but it's not truly possible.
However, at the system level, two systems that make multiple bets over the long run, may be +EV -- even if they are on the opposite sides of some games. It's two different things, game EV != system EV.Comment -
ljump12, Thank you for the reply.
Yes, we are fighting the same battle and I think we agree on the overall concept (“a +EV wager is a wager in which your return is expected to be positive”) but here is where I think we disagree.
EV, to me, is not about the event, it is about the wager. For example, if we flip a fair coin, the probability that it will land heads (or tails) is 50%. But that doesn’t say anything about EV. If I risk one dollar on heads @ +105, then my wager is +EV. If however I risk one dollar @ -105, then my wager is –EV. This is easy to figure out because there is a known probability for a single event, namely one flip of a coin. This single event probability does not exist for sports betting. The only way you can estimate the probabilities is to examine the whole set (or sufficient random sampling) or a subset selected with the use of certain criteria.
Now looking at closing ATS wagers at random, the probability that I will win is approx 50%. This again says nothing about EV. If my wager is @ +105, then my wager is +EV and if my wager is @ -105, then my wager is -EV. I hope we agree on this. If we don’t, then we have to agree to disagree.
Assume that there is a generous book that will let me wager on any one side ATS @ +105. I hope we can agree that every wager I make would be considered +EV. If not, what would you call it? Now imagine a player that uses certain criteria to select a subset of ATS wagers and has a validated win % of 55% @ -110. Again, I hope we can agree that every wager identified by his validated criteria that this player makes @ -110 would be considered +EV. Again, if not, how does one describe his wagers?
Therefore, even if these two players wound up on opposite sides of an ATS wager, they both are making +EV wagers. Obviously both cannot profit from this same event and equally obvious is that the majority of their wagers cannot be on the opposite sides of each other. This happens because there is no such thing as a single event probability for sports wagering. To come up with a probability requires a sampling of like wagers. The only thing that anyone can say for certain about a single event is after the game, one team had a 100% chance of winning and the other team had a 0% chance of winning. There is no single % that anyone can come up with prior to the event without a sampling and then dependent on the criteria used to create the sample, the %’s would be different.
I hope this is understood. This really could be a chapter to fully discuss or at least a couple of hours of discussion over a few beers.
One of the things I want to stress in all this is the fact that it is just not true that only one side of a wager can be +EV. Yes, that sounds ridiculous but it truly depends on the sample. I hope everyone can agree with me that they have made a sports wager that based on their criteria was +EV but lost. How far-fetched is it that someone else using different criteria identified the other side of your wager as +EV and won?
Have I converted anyone?
Joe.Comment -
looks like syracuse had no value tonight
keep laying that chalk guysComment -
Here's hoping Cornell +8.5 has value.
Comment -
This is 100% false... just because a player is a winning bettor doesn't make every single bet he places +EV. And just because it isn't readily apparent which side is in fact +EV doesn't mean we can just assume any side a winning gambler places a bet on is.
Only ONE SIDE AT MOST of any bet is ever going to be +EV under normal circumstances. Think of it like this... expounding on the "fair coin" example let's say a player is given +110 odds on every coin flip. Once out of every 100 bets the "book" will use a rigged coin that only lands on tails. Our player hasn't picked up on that... his "system" is to bet every flip and to always bet heads. He's a +EV player overall but once every 100 bets he is in fact making a -EV bet because of an anomally that his "system" has failed to recognize. Still a good and profitable system but it doesn't always make +EV bets.Comment -
Well put, you always have to determine value by how many points you receive or lay then what is expected after thorough analysis. Sports Center and CBS Sports don't know anything about value.Comment -
Tonight's St. Mary's/Baylor game has a line of Baylor -4.5. This is being assessed as a ML of 66% chance of Baylor winning.
It is possible that if this game were to be played in that infinite parallel universe scenario the true and actual probabilty of Baylor winning may be 60%.
Therefore, even if you could get Baylor at -4.5 +105 you would still not have a +EV bet, you would only be in position to lose less in the long run.Last edited by Art Vandeleigh; 03-26-10, 04:09 AM.Comment -
Let's try to clear up some semantics.
EV = Expected Value of one's wager (not the probability of which side will win though that is a component to calculate EV)
Now I know I have thrown a monkey wrench into the thought process by entering a 2 player scenario so let's stick with one player. If you have a validated 55% ATS win % @ -110 and you find a wager that meets your validated criteria, how do you refer to that wager? Do you call it +EV, -EV or do you say something else?
And please don't try to come up with an answer to support your position, answer honestly (not that I am accusing anyone of answering dishonestly).
Et tu, Justin?
Joe.Comment -
Joe, I'll make one more post and then leave it to the people who know what they're talking about...
I think you are incorrectly putting "ATS" as the center of your universe. The center of your universe should be the true and actual probabilty of each game, which may or may not coincide with the ATS line.
Let's say (sorry for this ridiculous example) there is only one sportsbook in the universe where you can play and only one linesmaker making the lines. If the line you were betting against was being made by a young inexperienced linesmaker, you may have been able to hit 55% while he was there making many mistakes and judging spreads far from their true and actual values.
If the book now hired a more astute linesmaker who could make lines much closer to the true and actual probabilty of these games, you would have a much more difficult time achieving this level of success.
If the book was still unsatisfied and hired MR. Perfect whose ATS lines ALWAYS coincided with the true and actual probabilty of the game, it would no longer be possible to have a long term edge and therefore long term profit.
Validated or not, the fact that you were able to have a series of bets whose summation turned out to be +EV against one set of past lines, will not turn out as well if a future set of lines appears which are much tighter and closer to the games' true and actual probabilty (infinite parallel universes).Last edited by Art Vandeleigh; 03-26-10, 07:29 AM.Comment -
Art, I apologize if you took my post personally, but my intent was not to offend you (or anyone else for that matter). We (as well as others) just have a disagreement over a topic we are discussing. Of course, a message board is a far cry from an efficient way to discuss certain topics. For example, in your reply, you make reference to efficient market theory which not only can be a chapter but a whole book. I too will try to respond one more time and if the topic gets dropped, so be it.
I am not putting ATS at the center of my universe; I was just using it as an example. I thought by using it that everyone could agree that the random long-term win % for ATS wagers was 50-50.
I was serious when I asked my question which is simply: How does one describe a winning bettor’s wager? Based on what has been posted so far, it appears that no one is willing to say that a validated winner’s wagers are +EV unless they “know” the "actual” probabilities. (Which makes me wonder, how can the term +EV be thrown about in all those other posts on this board?). Unfortunately, “actual” probabilities do not exist for a single sporting event (meaning one actual game). For sporting events, you either have an estimate based on the entire group of like wagers (or sufficient random sample) or you have estimates based on a sufficient sample drawn with the use of certain criteria. And based on different combinations of criteria, you will wind up with many different samples that will have some overlap of other samples, all of which will have different “validated” probabilities for that specific combination of criteria. But that estimate is not the probability of one specific game. What that estimate is telling you is that when you wager on the next 100 games using this criteria, you should win x% of the time.
One last point to expand on the sentence above. OK, hold down the snickers! It has been pointed out that the “actual” probability represents what would happen in the long-run if the same game was played over and over again. I don’t agree. This concept works for trials where there is a “known” probability such as a fair coin, a fair roulette wheel etc., but to me, applying this to sporting events is like comparing apples to oranges. There is no trial that can actually be repeated. All you can do is compare criteria. And at the end of your analysis, all you can say is that when a Team A had certain criteria facing a Team B with certain criteria, Team A won x% of the time. This point certainly needs to be expanded upon but I will stop here. Hold down the applause please!
I honestly hope that someone is able to benefit from this “discussion”. Even if it only makes you question what you believe or to look at or try different things, I am pretty sure it will benefit you if you do.
Joe.Comment -
No problems here Joe, I was serious when I said I need to leave this to others who know more I'm really not too well versed on this subject.
I remember Ganch using the infinite parallel universe analogy for single sporting events in a previous post(s) (hope I remebered correctly), so I am just following his lead.
Could be you are closer to the truth. If you are correct though this means that while rolling dice or spinning roulette wheels have a definite probabilty known before the event, as the complexity and chaos of an event increases, at some point this concept of definite probabilty breaks down and turns into something indefinite that can't be measured before the complex chaotic event begins (because as you stated in the above quote it doesn't exist). It doesn't sound too logical to me as a layman but neither does the concept of quantum physics or quantum mechanics, so there you go.Comment -
BD you make value wagers despite thinking otherwise- you just dont calculate your edge to the degree other do.
True, Losers are losers whether its LTProfits losing EV wager of a +150 dog getting +200 or any one of your 41% of losing 5k wagers. Hindsight is 20/20 and holding losing tickets is irrelevant to determining if wager had value
But to make wagers of 5k you must believe your winner has more than 52+% of winning to profit. It is unrealistic for you to think you will win 100% of those 5k wagers so what is the deciding factor for you to pull the trigger? (Pick will win 6 out of 10?, 6.5 out of ten? etc) It's also unrealistic to think you made these plays as if they were a 50/50 coin flip and you manage to hit 58%.
How many of the 41% losers would you make the same play again if you could do over? I would venture to guess quite a few of them. You would only play losers again is if you saw some value in the play and felt long term you would profit making that play over and over. Hence "expected value"
But if you think they all are 100% sure things, hitting 58% of "sure things" isnt that good and you are extremely overvaluing your edge on the plays.
In more real terms, if you are hitting 58% of 5k plays @ -110 you expect to win 60% of the time (we all overestimate our true edge)you are making money but you must ask yourself what is more important profits or percentage? Hitting that high of % you are costing yourself profits by not playing more games even if you win those extra games at a lesser 54%
While I agree any idiot can figure out that +200 is obviously a better price than +180 it doesnt necessarily make it a +EV wager- if when all the data is reviewed and interpreted a break even price of +210 is needed
Value in itself is a relative term(likely reason its overused) because there are no absolutes in sports betting and everyone has their own way to handicap. If Mn@ home is favored by -7 and line moves to -6.5, the half point is valuable to the MN backer willing to lay any price but there is no value to the handicapper who capped the game at MN -4.5. MN 6 pt teaser player will find value in the move
The Saints backer just lost the key 7 as well.
How can the same 1/2 point move affect this many people so many ways if "value" wasnt a relative term?
value to me is getting more than your fair share of a ML (laying less or taking more)based on expected probability of winning or laying less points than your method of handicapping indicates for a fair break even or better price
The reason everything is "expected" value is because of the lack of absolutes however this is the exact reason sports betting is beatable. If you're good at interpreting the available data better than most(catching slow line moves on hot teams with poor records etc), you can beat the line- unlike a dice outcome it doesnt matter how well you know the history of a die or history of past throws- where winning comes down to short term luck against negative odds
Also not every wager has at least one + EV outcome- IE Coin toss for Superbowl was -115 both ways.
Theres more to picking winners to be successful because one can pick 7 of 10 winners layin heavy chalk and either be close to even or even behind. Where is the benefit of making little or no money at 70% win pct?Last edited by chaka; 03-28-10, 07:53 PM.Comment -
good thread
Comment -
I understand why people hate hearing about value I do. How ever just because people over use the word does not mean it's not real.Comment -
Its hard to make the claim that the term "value" is overused. The reason being the lack of perfect efficiency in the market which allows there to be value on one side (and in a decent amount of cases, both sides) of every game during the lifespan of a game line.
A more accurate statement would be the term "value" is misused- by people that lack the ability to quantify the probability of a games outcome.Comment -
Bump for the thread asking about "value".
Joe.Comment -
When two winning systems have opposite sides of the same line, only one (at most) can be +EV.
Having a 55% win rate or a 5% ROI does not mean that every bet you make has that expectation. It means the pool of bets that they are coming from has that mean.
To continue with a rigged coin example (there's a book offering +105 / -110 on coin flips, depending on which side the money is on). Every 100 flips the flipper switches in a rigged coin that is either 100% heads or 100% tails.
A guy bets the +105 side every time. He is pretty clearly going to have a 2.5% ROI.
The flipper's friend only bets when the fix is in. Sometimes it's +105, sometimes it's -110 but he always wins. He is pretty clearly going to have a 97% (ish math is hard) ROI.
Although the first player is following a +EV strategy the times he ends up against the flipper's friend that individual bet is -EV.
This is an unknown unknown so there's not a lot he can do about it (although I believe unknown unknowns can be transformed in to known unknowns and then known knowns through brainstorming but that's another discussing) but that individual bet is -EV.Comment -
Shop for value or pass..they play the national anthem every day never force playsComment -
Real values are in Futures.
Getting teams like Flyers at 50-1, Bruins at 80-1 is just absurd.Comment -
This is a serious question.
If you have criteria that results in 55% winners at -110, when you find a wager that meets that criteria do you consider that wager to be +EV?
If not, what do you call it?
Joe.Comment -
Yes, with the information currently available I believe that wager to be +EV (or rather I believe the wager falls somewhere in an EV range with its central point at 5%).
The more times this criteria has been tested the smaller that range (you might want to call it a confidence interval) becomes and therefore the more confident I am that True-EV is >0.Comment -
But why so difficult? There are a lot of services that offers for free information about Value bets. You can try oddsportal, oddsfinder or the one of the best portals BMbets. They even have calculator for real value (extra added) checking.Comment -
College Basketball total opens 141...RAS releases Under
Screen moves to 137 but you find a slow moving book still hanging 141
Under 141 has VALUEComment
