Let-s fight with their weapons

[giraldi]

I'm not sure it's an original idea. It came to mind after I lost in 3 weeks almost all the advantage I gain for 2 months If we consider true the following statements: 1. Payout of bookmakers and our loss is based on the calculation of probabilities for events on long-term 2. Bookmakers approximates probabilities much better than us for a specific event 3. The laws of statistics are available for all bookmakers and bettors For simplicity let’s take the case of an event for which it is offering odds of 2.00. This listing suggests a probability of 50%. According to bookmakers margin we can assume that the true probability is 45%. No one knows the exact probability of an event and everyone can go wrong, including them, let's say + or -10%. Let's assume that the mistake is on their advantage and so we can easily assume then that a real probability for 2.00 offered is around 40% This means that betting on 10 units, win 4 games so 8 units. On the whole, forever losing 2 units every 10 invested. So probably the system works. Now, of course the statistics allow us to win the first 300 bets in a row that we have 600 won at 300 units invested. If we continue to play the laws of statistics work and when we draw the line at 1000 units we have lost 200. So we want or do not want the results are according to the theory. But that's not the same for them? Statistically they can earn 200 at a time and so we have 0 units won and 200 units invested. Statistics will continue to work and to 1000 invested we lose again 200 units. The idea is if you don't play the first 200 bets? We have 1,600 won at 800 units invested. This example is extremely. I put it so you understand. To see in practice: Let-s say that we have the Pinncale odds between 2.50 and 3.00. Considering what I've said above, considering their margin and approximation errors, in the worst case the probability, for this range of odds, is about 25% They are very good on calculated odds so we expect 25 matches out of 100 to be winning for odds in this range. If you are at the beginning of a season and 10 games with odds in this range that have not come out winning is correct to think that statistics will work and will have more events won to achieve the level of winning 25 percent? I have the tools to check this theory and I make a small program for that. I know it seems like kind of a Martingale but I think is far more subtle. What is your opinion?

[zbrochu]

Re: Let-s fight with their weapons These events are unrelated so it is not conditional probability (http://en.wikipedia.org/wiki/Conditional_probability). It is gambler's fallacy (http://en.wikipedia.org/wiki/Gambler%27s_fallacy) because you understimate what "long-term" means. It is few milions of bets, not few hundreds.

[mrsnoopy]

Re: Let-s fight with their weapons in my opinion you have made a fundamental erronius assumption that many bettors make, and this one fact is one of the most important things that a succesfull bettor needs to know. 1.) you assume that the bookmakers prices reflect that bookmakers prediction for the outcome of the match ie that odds of 2 represent the probability of that outcome being 50% etc. The reallty is that bookmakers make profit from the overround as in for over/under 2.5 typical odds would be under 1.68 and 2.16 over this gives a 106% book and is how they make money. HOWEVER it is important to realise that they are only garanteed to make money if they have an equal (related to the odds) amount of punters betting on each side of the bet. For example if a bookie calculates that Man Utd have a 91% chance of beating QPR at old trafford and QPR have a 3% chance of victory the draw by default is 6% the fair odds would be 1.098 hw 16.66 draw and 33.33 for QPR. The problem with this for the bookie is that all the mug punters that walk into the bookies will see that they have to risk £100 on man utd to win £9.80 and that is just not acceptable for thye average punter! nearly all the money would go on QPR at 33-1 as the thinking would be that for £5 I could win £165 if there is a shock away win. This would mean that the bookies would have an uneven book for the match and that would force the bookie to gamble!!!! bookies don't like to gamble they prefer to take the garanteed overround profit from as many matches as possible. The solution for them is to offer odds that will attract bettors on both sides of the bet so odds of say 1.18 hw 8 draw and 16 away win. Enough punters would back Man Utd as a "sure thing" and enough would be willing to bet £5 or £10 on QPR for a bit of fun and a nice win if there is an upset. IT can be seen that the bookies job is to post a price that is designed to evenly split the betting action of the betting public (95% of the money that is wagered) It is not in the bookies interest to post correct odds to defeat the "smart punters" as they represent only about 5% of the money wagered (at most). The bookies odds are designed to defeat JOE PUBLIC using the overround to generate profit from as big a turnover as possible. That is why sports betting is a beatable proposition if you can accuratley predict outcome probability percentages (And that's a BIG if btw!!) But it CAN BE DONE!

[giraldi]

Re: Let-s fight with their weapons Hi, zbrochu Thanks for your intervention and relevant observations 1. Of course the events are not so obviously related, though there might be a whole discussion especially if we consider the events in a certain well established area. Lotto games are absolutely unrelated events but when we speak about football and sports games in general, especially when limit the area of selection may be a discussion in my opinion. For example in a particular League, for a specific team event "team win" is not completely independent in my opinion of previous teams events. Otherwise, everything would turn to Lotto and then we don’t have any chance to beat the bookmakers. So I guess it's not about a pure “gambler fallacy” here as long as the events are not 100% independent and the event “goal” (the base for all events) is not really random. (more chances to appear on some circumstances) If you think so then, for a team very unlucky, as you say, the next event is absolutely independent and so can't say anything about that or that team may be forever unlucky 2. The number of events, again I agree with you, I used small numbers just to understand the idea. The idea is to catch the trend. 3. Maybe is stupid think but I am willing to try this idea (and you now I have the tools) on narrow enough areas to keep minimum connection between events. Will see what happens Thank you again for your intervention. Maybe you are perfectly right but don-t destroy my little dreams.

[giraldi]

Re: Let-s fight with their weapons Hi, mrsnoopy I really don-t think I am wrong here. Of course they adjust the odds so they become attractive for all events but they first have to calculate the probabilities as close as they can These are starting odds and after that bets are placed they move the odds but almost never they will go much under probability calculated. In my opinion is a big risk for them to show odds with big value no matter the amount of money is involved. I totally agree that after this movements appears greater opportunities for value bets but very hard to find I don-t know, it is just my opinion

[mrsnoopy]

Re: Let-s fight with their weapons I'm a bit confused as you state in your other thread (elo and poisson system) that your system regularly finds value that is 20%+ better than the odds suggest. In this thread you suggest finding value is very hard, your thinking seems out of sync with your aspirations! maybe your recent losing streak has affected your confidence??? My advice is to stick with your original system as losing streaks are completely normal, even if you have a solid edge long term edge of say 8% some of the losing streaks can be quite brutal, that is why bankroll management is so important. I personally use flat stakes and never bet more than 1.5% of my bankroll no matter how convinced I am that I have found value as I know I will have to survive some losing streaks.

[zbrochu]

Re: Let-s fight with their weapons

I'm a bit confused as you state in your other thread (elo and poisson system) that your system regularly finds value that is 20%+ better than the odds suggest. In this thread you suggest finding value is very hard' date=' your thinking seems out of sync with your aspirations! maybe your recent losing streak has affected your confidence??? [/quote'] And whats wrong with that? he can stick to his original system that I personally like a lot by the way and continue to develop new ideas and startegies. If he likes sports-betting and modeling sports, he is free to check different approaches and methdos. It is what I do all the time, at this points I have 3 strategies for football and I am developing new ones for US Sports. And I do not think anything wrong in loosing confidence, it is normal Giraldi might be confused and overhelmed by recent streat that happened after so many weeks of nice profits. I completely agree - bookies set odds to reflect true probabilites of events. After this they use tools to balance the actions. The tools are: 1) moving odds as a result of bets placed 2) low limits early in the week 3) very sensitive odds that it is very easy to affect early in the week But this tools do not change the fact that original odds are based on their sophisticated econometrical model that is designed to reflect fair estimates of teams' chances. And if you check all bets in region of 1.95-2.00 you realize that indeed this events have had ~50% hit rate, exactly the same as bookies opinon on them.

[zbrochu]

Re: Let-s fight with their weapons

1. Of course the events are not so obviously related, though there might be a whole discussion especially if we consider the events in a certain well established area. Lotto games are absolutely unrelated events but when we speak about football and sports games in general, especially when limit the area of selection may be a discussion in my opinion. For example in a particular League, for a specific team event "team win" is not completely independent in my opinion of previous teams events. Otherwise, everything would turn to Lotto and then we don’t have any chance to beat the bookmakers. So I guess it's not about a pure “gambler fallacy” here as long as the events are not 100% independent and the event “goal” (the base for all events) is not really random. (more chances to appear on some circumstances)

You are right, streak of football games of one team is not a streak of unrelated events but I still do not think that with this approach would have a chance to get an edge over bookmakers. You need to know the strenght of both teams to evalute if bookmaker is right or wrong in his odds.

2. The number of events, again I agree with you, I used small numbers just to understand the idea. The idea is to catch the trend.

What do you mean catch the trend?

3. Maybe is stupid think but I am willing to try this idea (and you now I have the tools) on narrow enough areas to keep minimum connection between events. Will see what happens Thank you again for your intervention. Maybe you are perfectly right but don-t destroy my little dreams.

I am far from desire to destroy your dreams Giraldi, I am sorry if you got this impression. I think it is great thing you have lots of ideas and want to check them in reality. Just start checking this and develop in time but do not bet real money yet. I am always happy to respond to your question if you have any. And I keep fingers crossed bro.

[giraldi]

Re: Let-s fight with their weapons Indeed, I made some test at it seems that is not a good idea. Considering previous and current season there are some teams or group of teams that may qualify to test on but the number of events is far too low and not worth the risk or effort to try. It was just an idea. Sorry for the space wasted on server :)