Beat the Mathematical Model

[Darren]

Ok let's adjust it.

Again you go with your farcical comments. The model is predicts what the scores should be based on past performance, form, players la de da. But (as I've always maintained) there will be some element of randomness. That is the beauty of football Darren. You know that. If the best teams won all their games and no one could beat them it would be boring. Just think how boring the premiership was last year when Man U only dropped 2 points all season against non top 4 teams.

I still don't understand on what basis you are translating your friend's predictions on goals scored per team (which predictions are themselves impossible given that they predict fractions of goals) into results.

You don't seem to be grasping my point about the circular nature of the model. For about the 632nd time my point is this: your friend has defined certain characteristics as factors that make up a good team. He has ascertained the extent to which various teams have these factors and ranked them on this basis. Logic would suggest that where team A has your friend's factors to the power of 10 they must be a better team than team B which has these factors to the power of 5 and hence should beat them to a head to head contest.

However, you or your friend have provided yourself with a get out of jail free card by saying that, where Team A does not defeat Team B, it is not due to any problem with the model but due to "randomness" or "luck" or "exogenous factors". It is consequently impossible to prove the model to be wrong or incorrect and hence is circular and self confirming.

 

[Darren]

Your logic doesn't necessarily make sense. See below.

That is illogical. Many people simply can't be bothered with 'systems', or don't believe in them (like you). You might say the same about stockbroking. There are people making vast sums of money from that - so why doesn't everyone do it?

Fortunately for the bookies, there are enough people that don't follow systems and, hence, the bookies can make profits. It's fairly difficult for the bookies to know who is using systems and, although they have techniques, there are so many systems as to make things difficult to spot. Some punters win, some lose. Most of this is pure luck, so you can't guarantee that a 'winner' must be cheating.



This isn't necessarily true. The bookies don't always try to calculate the 'true' odds. They sets their odds such that they can make a profit. Think of the following example. Suppose that the 'TRUE' odds of tossing a coin are 50% heads and 50% tails. If the bookies took bets on this, and 95% of punters bet on tails (maybe based on spurious 'hunches'), the bookies would be facing a loss if it did come up tails. As such, the bookies would offer longer odds on heads in order to entice some more bets on that. If you were an astute gambler, you would spot that the long odds are offering a decent reurn to you, and you'd stick the money on. You'd be happy (remember that you're not guaranteeing that heads will win - just that you're getting good odds on it). However, the bookies would also be happy, as they'd be hedging their own outputs and will make a profit no matter what, even though the odds are favourable to you.

I think that most 'systems' don't aim to actually predict WHO will win each game. They simply aim to spot where bookies are offering slightly generous odds. In reality, these odds will only be generous to a degree of around 1.5% or 2.0% but if you do this over a long time you'll get a nice return.

All fair points, but isn't a fair summary of the points that you will only make any significant money from gambling if you spend a lot of time studying the thing you are betting on and have a lot of money to play with?

That's true, but spend a lot of time and money on any human activity and you will most likely be successful at it.

 

[RichJay]

I think most have missed the point of the system, which is to provide an edge when assessing the relative odds.

I agree completely. For anyone who didn't read the paper...

"We bet on all outcomes for which the ratio of model to bookmakers' probabilities exceeds a specific level. For sufficiently high levels, we have shown that this strategy yields a positive expected return".

This makes all these predictions pointless because the model wouldn't suggest betting on all of these games.

 

[Geordie Blade]

All fair points, but isn't a fair summary of the points that you will only make any significant money from gambling if you spend a lot of time studying the thing you are betting on and have a lot of money to play with?

That's true, but spend a lot of time and money on any human activity and you will most likely be successful at it.

Yes - but that's why systems can exist! Ollesandro's mate has (potentially - I don't know enough about that system to judge it) seemingly done that. You don't necessarily need a lot of money to succeed though. A return of 1% per day will cumulatively turn £10 into £374 within a year. A 2% return would make it £13,500! That's theoretical, of course, since you cannot guarantee a regular return every day. However, if you have enough time, the more bets you place, the more predictable your return will be.

I actually did this a few years ago, without using real money. I followed a system called the 'favourite-longshot bias' (look it up!) and made a return of about 5% over three months. It took 527 bets in total though! If I'd started with £1000, I would have made about £50 and it would have taken a load of effort! That's why I never went on to do it with real money. Still, I did prove that long-term money can be made from the bookies.

 

[Darren]

I actually did this a few years ago, without using real money. I followed a system called the 'favourite-longshot bias' (look it up!) and made a return of about 5% over three months. It took 527 bets in total though! If I'd started with £1000, I would have made about £50 and it would have taken a load of effort! That's why I never went on to do it with real money. Still, I did prove that long-term money can be made from the bookies.

But isn't that why you do need a load of money? You would have made a 5% return on your investment and it would have taken a load of effort to make £50. You quite rightly judged it was worth the effort. However, if you had £1 million you would have made £50,000 and might then have thought it was worth the effort (though couldn't you probably get a better long term investment rerturn on that sort of money by investing more conventionally?)

 

[Geordie Blade]

But isn't that why you do need a load of money? You would have made a 5% return on your investment and it would have taken a load of effort to make £50. You quite rightly judged it was worth the effort. However, if you had £1 million you would have made £50,000 and might then have thought it was worth the effort (though couldn't you probably get a better long term investment rerturn on that sort of money by investing more conventionally?)

I was only betting on 6 or 7 horses per day. After 527 bets, however, the return was 5%. It's not quite right to think of it as '5% over three months', it should really be '5% after 527 bets'. If I could make 527 bets in a day (if I had enough time and if there were enough horse races!), then my wealth would (in theory) grow by 5% per day. Starting with £10 would turn into big £££s pretty quickly indeed.

My problem is that I'm an academic and, whilst I love the theory, I can't be bothered to find 500 bets per day! Other people can though - and they do!

 

[Darren]

I was only betting on 6 or 7 horses per day. After 527 bets, however, the return was 5%. It's not quite right to think of it as '5% over three months', it should really be '5% after 527 bets'. If I could make 527 bets in a day (if I had enough time and if there were enough horse races!), then my wealth would (in theory) grow by 5% per day. Starting with £10 would turn into big £££s pretty quickly indeed.

My problem is that I'm an academic and, whilst I love the theory, I can't be bothered to find 500 bets per day! Other people can though - and they do!

But isn't that a bit like the famous compound interest fallacy which goes something as follows: if my ancestor had invested one shilling at 2% compound interest in 1755 (or whenever) and the results had been passed down through the generations I would, in theory, now have more wealth than exists in the entire world.

The point being that, whilst in theory, it might be possible to make a 5% profit per day, due to the various constraints you mentioned it is, in fact, practically impossible.

 

[Selly]

Starting with £10 would turn into big £££s pretty quickly indeed.

At which point the bookies will most likely suspend your account and/or refuse to take bets of the size you need to maintain a 5% return.

 

[Geordie Blade]

But isn't that a bit like the famous compound interest fallacy which goes something as follows: if my ancestor had invested one shilling at 2% compound interest in 1755 (or whenever) and the results had been passed down through the generations I would, in theory, now have more wealth than exists in the entire world.

The point being that, whilst in theory, it might be possible to make a 5% profit per day, due to the various constraints you mentioned it is, in fact, practically impossible.

Of course it's not practically likely that this would happen to that degree. But I don't think there's any doubt that systems can work and help punters make reasonable sums of money. You're kind of right about the compound interest fallacy though. However, real interest is a lot slower than that (5% per year, not 5% per day). The problem with the betting version is that the 5% isn't guaranteed. Some days it'll be negative and the compound thing won't really work.

As I pointed out above, there's no reason why the bookies would suspend an account of someone making money (as SellyOak Blade suggested). As long as the sums are offset by other punters betting on another outcome, the bookie will actually be happy to accept bets from 'system' betters.

 

[Ollessendro]

I still don't understand on what basis you are translating your friend's predictions on goals scored per team (which predictions are themselves impossible given that they predict fractions of goals) into results.

You don't seem to be grasping my point about the circular nature of the model. For about the 632nd time my point is this: your friend has defined certain characteristics as factors that make up a good team. He has ascertained the extent to which various teams have these factors and ranked them on this basis. Logic would suggest that where team A has your friend's factors to the power of 10 they must be a better team than team B which has these factors to the power of 5 and hence should beat them to a head to head contest.

However, you or your friend have provided yourself with a get out of jail free card by saying that, where Team A does not defeat Team B, it is not due to any problem with the model but due to "randomness" or "luck" or "exogenous factors". It is consequently impossible to prove the model to be wrong or incorrect and hence is circular and self confirming.

It's because there isn't really one. All I've done is awarded the result to the team with the highest goals providing there is a significant margin (e.g. more thn 0.3 goals). I'm willing to listen if you have a better suggestion?

And Darren I am grasping you point about the circular notion, I just don't agree with it. You can go on and on and on and on (In fact you have been doing) but i am not going to roll over and agree with you. My friend uses a model that incorporates all the relevant factors (past performance and la de da). This is not just something he has dreamed up but something the industry created. The best statistical/probability minds and geeks with footballing knowledge have come up with a system of beating the bookmakers and predicting the probability results will happen. From this a ratings table can be developed.

As for your 3rd point then bravo. I clearly stated several times that this excercise was futile because of externailties. The likelehood is that my friends model will predict better results than you, Tom, Dick and Harry but it's not guaranteed. Randomness and luck will affect predict power. It's what makes football the beautiful game! You have no cause for complaints because you suggested this and steamrollered ahead with this. You cannot blame me for making this a personal vendetta to discredit myself and my friend .... a task which you've finally realised you cannot do.

 

[Darren]

It's because there isn't really one. All I've done is awarded the result to the team with the highest goals providing there is a significant margin (e.g. more thn 0.3 goals). I'm willing to listen if you have a better suggestion?

And Darren I am grasping you point about the circular notion, I just don't agree with it. You can go on and on and on and on (In fact you have been doing) but i am not going to roll over and agree with you. My friend uses a model that incorporates all the relevant factors (past performance and la de da). This is not just something he has dreamed up but something the industry created. The best statistical/probability minds and geeks with footballing knowledge have come up with a system of beating the bookmakers and predicting the probability results will happen. From this a ratings table can be developed.

As for your 3rd point then bravo. I clearly stated several times that this excercise was futile because of externailties. The likelehood is that my friends model will predict better results than you, Tom, Dick and Harry but it's not guaranteed. Randomness and luck will affect predict power. It's what makes football the beautiful game! You have no cause for complaints because you suggested this and steamrollered ahead with this. You cannot blame me for making this a personal vendetta to discredit myself and my friend .... a task which you've finally realised you cannot do.

It's a system for beating the bookmakers where there is some mathematical/probability error in their odds. It may well work on that basis. I neither know nor care.

The point you don't seem to grasp is that it cannot provide some neutral judgment of how good teams "really" are. You have obviously failed to understand my last point, it is not the concession you appear to think it is: theories that are circular and cannot concievably be proven wrong are obviously useless. I have a theory that if you produce someone to me who has fathered children, I will predict without speaking to them or seeing them, that they will be male. I will also guarantee you that my prediction will be 100% accurate .

That theory is self evidently pointless, circular and impossible to disprove because it is a necessary condition of being a father that one is male. Hence the necessary 100% correctness of the theory is built into the theory itself. On a more complicated level, the same applies to your interpretation of your friend's theory:

If, according to the theory, Team A is better than Team and beats them in a game, that proves the theory is correct.

If Team A fails to beat B then that was due to random factors and it proves the theory is correct.

Can't you understand that a theory that cannot possibly be disproved is useless?

 

[Selly]

As I pointed out above, there's no reason why the bookies would suspend an account of someone making money (as SellyOak Blade suggested). As long as the sums are offset by other punters betting on another outcome, the bookie will actually be happy to accept bets from 'system' betters.

My point related to the fact that Geordie had compounded his returns. Assuming that the size of the bets grows (which it has to to maintain a 5% compound return) you will eventually reach a point at which a bookie will be suspicious that you know something they don`t.

Try walking intro any bookie and putting any significantly large sum of money on (£1000 for example) they will probably have to OK it with HO.

Continue to do that (especially if you end up "up" in the long term) and you will be banned eventually.

Unless you use Betfair of course in which case you only have to find enough mugs to match your bet...

 

[Geordie Blade]

My point related to the fact that Geordie had compounded his returns. Assuming that the size of the bets grows (which it has to to maintain a 5% compound return) you will eventually reach a point at which a bookie will be suspicious that you know something they don`t.

Granted! I only meant that they wouldn't necessarily stop you using a system. They'd certainly limit your stakes, though.

 

[happyhippy]

As I pointed out above, there's no reason why the bookies would suspend an account of someone making money (as SellyOak Blade suggested). As long as the sums are offset by other punters betting on another outcome, the bookie will actually be happy to accept bets from 'system' betters.

They do all the time. I know of several people who can rarely 'get on' on course, and also people whose accounts have been severely restricted by turf accountants due to being very good.

Unfortunately, I'm not one of those people

 

[alcoblade]

I still don't understand on what basis you are translating your friend's predictions on goals scored per team (which predictions are themselves impossible given that they predict fractions of goals) into results.

You don't seem to be grasping my point about the circular nature of the model. For about the 632nd time my point is this: your friend has defined certain characteristics as factors that make up a good team. He has ascertained the extent to which various teams have these factors and ranked them on this basis. Logic would suggest that where team A has your friend's factors to the power of 10 they must be a better team than team B which has these factors to the power of 5 and hence should beat them to a head to head contest.

However, you or your friend have provided yourself with a get out of jail free card by saying that, where Team A does not defeat Team B, it is not due to any problem with the model but due to "randomness" or "luck" or "exogenous factors". It is consequently impossible to prove the model to be wrong or incorrect and hence is circular and self confirming.

I don't think there's a get out of jail free card - the randomness, or noise, will average out over time. On any Saturday, the noise could well swamp the signal. But given enough repetitions, the noise should balance out and the signal will be clear. I'm sure you understand this, so I'm probably missing your point.

UTB

 

[happyhippy]

I don't think there's a get out of jail free card - the randomness, or noise, will average out over time. On any Saturday, the noise could well swamp the signal. But given enough repetitions, the noise should balance out and the signal will be clear. I'm sure you understand this, so I'm probably missing your point.

UTB

Which still ignores the point of the model. It seems that both Ollesendro and Darren completely misunderstand the point of the model.

Remember the model? It's on another thread, somewhere.

Darren - the statistical model does not purport to give an accurate representation of a match's outcome. It gives a statistical likelihood.

Ollesendro - given that 'good' is an abstract, a model can only be used to show how you, personally, define 'good' to be. Given that randomness is inherently built into the model, on several different levels, it is obvious that a theoretical table built on 'goodness' cannot be extrapolated.

Neither can the 'goodness' of a side be devined by the model.

Please, for the love of God(TM), will you leave the bone to be found by future archaeologists instead of chewing on the fucker.

 

[happyhippy]

I agree completely. For anyone who didn't read the paper...

"We bet on all outcomes for which the ratio of model to bookmakers' probabilities exceeds a specific level. For sufficiently high levels, we have shown that this strategy yields a positive expected return".

This makes all these predictions pointless because the model wouldn't suggest betting on all of these games.

I did intimate this before

I had to embolden those three words from the quote though. They speak volumes.

 

[Darren]

Ollesendro - given that 'good' is an abstract, a model can only be used to show how you, personally, define 'good' to be. Given that randomness is inherently built into the model, on several different levels, it is obvious that a theoretical table built on 'goodness' cannot be extrapolated.

I haven't misunderstood the model at all, given that what is quoted above is exactly what I have been saying to Olly!

I am well aware that the model is just a betting tool and cannot be used to establish how good teams "really" are; that is what Olly seems to be arguing and that is what I , for the past 723 years, have been arguing against...

 

[BarcaBlade]

This set of results was arrived at by attaching a value to each letter of the alphabet, A=1, B=2, etc then adding up all the values to get a total. As with a cyclic redundancy check I derived a number by seeing what value needed to be added to this total such that it is divisible by 11 without any remainder. This is the value that each team was assigned.

It's then a straight-forward process of awarding a win to the team with the highest value.

Blackpool v Ipswich AW
Bristol C v Doncaster D
Cardiff v Middlesbrough HW
Palace v SUFC D
Derby v Watford HW
Newcastle v Barnsley HW
Forest v Swansea HW
Peterborough v Coventry D
Plymouth v Preston HW
QPR v WBA HW
Sheff Wed v Leicester AW

(Actually I just guessed)

 

[happyhippy]

I haven't misunderstood the model at all, given that what is quoted above is exactly what I have been saying to Olly!

I am well aware that the model is just a betting tool and cannot be used to establish how good teams "really" are; that is what Olly seems to be arguing and that is what I , for the past 723 years, have been arguing against...

You're also asking for the model to be used in an inappropriate way to try and disprove it. It isn't there to predict results. Comparing just one week's worth of results from one division against others' predictions proves absolutely nothing at all.

While the model will undoubtedly produce the most likely outcomes from the data it is fed, it may suggest that absolutely none of them are betting propositions.

The last bit is the salient point with regard to the model.

 

[PeterNdlovu081]

Blackpool v Ipswich AW
Bristol City v Doncaster HW
Cardiff v Middlesbrough D
Crystal Palace v Sheffield United AW
Derby v Watford HW
Newcastle v Barnsley HW
Nottingham Forest v Swansea AW
Peterborough v Coventry D
Plymouth v Preston HW
QPR v WBA AW
Sheff Wednesday v Leicester AW

I got 2/11.

 

[BarcaBlade]

Blackpool v Ipswich AW - wrong
Bristol C v Doncaster D - wrong
Cardiff v Middlesbrough HW - correct
Palace v SUFC D - wrong
Derby v Watford HW - correct
Newcastle v Barnsley HW - correct
Forest v Swansea HW - correct
Peterborough v Coventry D - wrong
Plymouth v Preston HW - wrong
QPR v WBA HW - correct
Sheff Wed v Leicester AW - wrong

5 correct - not too shoddy.
Noticably, by simply choosing all home wins you would have got 8 correct.

 

[tomtheblade]

Blackpool v Ipswich D
Bristol C v Doncaster HW
Cardiff v Middlesbrough D
Palace v SUFC AW
Derby v Watford HW
Newcastle v Barnsley HW
Forest v Swansea D
Peterborough v CoventryAW
Plymouth v Preston HW
QPR v WBA AW
Sheff Wed v Leicester AW

I got 3/11 right. Pretty terrible. Don't think i'll be changing professions to a pundit anytime soon.

 

[Casper]

lol just spent an hour or so of my time reading 'The case of Darren V Ollessendro'.

Verdict:

Ollessendro's mates system is fit for purpose in so much as it is proven to make him money. However it should not be confused for predicting one off results or league tables at any particular point in time.

There you go, you are both right!

Systems do work - i made a tidy amount myself trading on Betfair. However, i stopped when i came to the conclusion that the amount of time and effort i was putting into it was not really worth the amount i was winning. Oh and i didnt have the guts (or strong enough heart lol) to stake larger amounts. £600 at a time was my limit.

Always follow the golden rule - if you cant afford to lose it, then dont do it!

 

[RIP Niftos]

Is there just me who can't wait to hear Darren tell us he got 5 right, and the ensuing same points about Olly's mate's system being made over and over again as (as we all know) noone else can have a view that differs to his without being bored into submission?

Wooooooo

 

[Ollessendro]

It's a system for beating the bookmakers where there is some mathematical/probability error in their odds. It may well work on that basis. I neither know nor care.

The point you don't seem to grasp is that it cannot provide some neutral judgment of how good teams "really" are. You have obviously failed to understand my last point, it is not the concession you appear to think it is: theories that are circular and cannot concievably be proven wrong are obviously useless. I have a theory that if you produce someone to me who has fathered children, I will predict without speaking to them or seeing them, that they will be male. I will also guarantee you that my prediction will be 100% accurate .

That theory is self evidently pointless, circular and impossible to disprove because it is a necessary condition of being a father that one is male. Hence the necessary 100% correctness of the theory is built into the theory itself. On a more complicated level, the same applies to your interpretation of your friend's theory:

If, according to the theory, Team A is better than Team and beats them in a game, that proves the theory is correct.

If Team A fails to beat B then that was due to random factors and it proves the theory is correct.

Can't you understand that a theory that cannot possibly be disproved is useless?

Who do you think you are Richard Dawkins?

I see what you are getting at, but it doesn't relate to this situation. The idea of a giant floating invisble marshmallow floating around earth is ridiculous but yopu can't prove it wrong. Hence you can randomly make up a theory about soemthing or other and if it can't be proved wrong it it is bullshit. But how many times do I have to tell you that this is not a theory Darren? This is a mathematical model cretaed by the best in the business. These footballing experts have come up with a method of estimating probability that works. It's not just some twat stating that there is an invisible elephant underneath the table.

Very clever Darren: I can see exactly what you are doing, but it will not work. You are trying to coax me into saying something along the lines of if by the end of the season my mates not top of the table then you are right. But you should know by know that I am never going to agree with you. For example if my mate was clear winner of the league table, would you accept that his professional judgement (based on the model) is a better indication of how good a team is than the league table? I think not.

 

[Ollessendro]

Is my mates initial prediction of United finishing in the bottom half of the table starting to look more realistic??