A football rating system, discussions, ideas

[PAULM03]

Re: A football rating system, discussions, ideas

Mr Onemore How will you cope with anomalies such as, for arguments sake, midweek matches having a totally different ratio of homes/aways/draws to weekend matches?

Does that happen? if it does it's something i never knew:o Just before christmas there was only 1 home win on saturday, so i backed the 3 homes on sunday purely because of this, one of which was liverpools last minute win at home to arsenal which returned a tasty profit overall. Was that a sensible move?

[Sir Puntalot]

Re: A football rating system, discussions, ideas I've got a headache. :lol:(

[mr onemore]

Re: A football rating system, discussions, ideas

Mr Onemore How will you cope with anomalies such as, for arguments sake, midweek matches having a totally different ratio of homes/aways/draws to weekend matches?

That is an argument in favour of looking at longer term trends rather than just recent form - things like you mention, even if true, are too detailed to program into a basic model - so by looking at a longer period you would hope to factor out things like sendings off, injuries, etc which in the longer term can be assumed to equal out.

[PAULM03]

Re: A football rating system, discussions, ideas

Just because events are unlikely, does not mean that a model will have more problems dealing with it - what you described above would have been rated unlikely by anyone - including a good model. But you can still find value bets whether an event is likely or unlikely - it doesn't matter.

Anither point i agree on, but you are testing your model, in reality, against the bookie. Can any model find a sufficient edge over a long period of time to beat a team of full-time odds compilers? or does a degree of 'instinct' always have to come into play?

[gazabroad]

Re: A football rating system, discussions, ideas I have to say that was probably just luck! I personally dont think that midweek form is any different to saturday form (unless maybe cup games enter into the equation). I´m sure mr. onemore will tell me different:o

[mr onemore]

Re: A football rating system, discussions, ideas

Anither point i agree on' date=' but you are testing your model, in reality, against the bookie. Can any model find a sufficient edge over a long period of time to beat a team of full-time odds compilers? or does a degree of 'instinct' always have to come into play?[/quote'] No - it can be done (I've already done it truth be told - I have a link on at www.unmentionable-site - should I move it here too?? - serious question) A little personal history - in the mid eighties I really got into the maths of all this - pre internet, slow computers, no info on games etc. There was also a 10% betting tax, and because of no internet bookies, an average overround of close to 10% on any match. Overcoming a total overround of about 20% was a non-sensical project - but the work I did then and since has never been in doubt. Nowadays, with no betting tax, and more competitive overrounds, the edges are down to a few percent. I can also crank out in milliseconds, what I used to have to leave on a mainframe computer overnight. The bookies/compilers have lost a cushion of about 15-20%. You also have massive volumes of bets on football compared to then - with more money in the wrong places - the bookies shift their lines to maximise volumes of bets across the spread - they're not even necessarily looking for the same things we want to - they rely on their edge. All in all - I say it can be done;)

[Sir Puntalot]

Re: A football rating system, discussions, ideas

I have to say that was probably just luck! I personally dont think that midweek form is any different to saturday form (unless maybe cup games enter into the equation). I´m sure mr. onemore will tell me different:o

I've not got any proof, but I hate betting midweek, because I am almost certain the results are much more inconsistent. I always think that footballers are much more up for a Saturday match than they can be for a midweek match. Just the buzz of a Saturday match, when you know the whole country is playing at the same time, and the fact that they may be tired from the weekend match also, add to that travelling tiredness. Maybe its bollox.....

[Sir Puntalot]

Re: A football rating system, discussions, ideas Mr O, post what you like, just not their link. ;)

[PAULM03]

Re: A football rating system, discussions, ideas

All in all - I say it can be done

I'm not a sceptic, i'm sure it can. Post your method here then we can all have a look. p.s. that link in the context it's posted is spam and will be treat as such. Don't take it personally:ok

[mr onemore]

Re: A football rating system, discussions, ideas

Mr O' date=' post what you like, just not their link. ;)[/quote'] OK Paul - understood;) I'll put a table summarising the results that can be found there, and then keep an update here as well. Please - is the way of pasting up tables at the 'other' site the same as here?

[Osesame]

Re: A football rating system, discussions, ideas Eng Prem this year Mon

11	6	2
58%	31.50%	10.50%

Tue

8	6	8
36%	27%	36%

Wed

4	5	4
31%	38%	31%

Sat

77	41	47
46%	25%	29%

Sun

18	18	13
36.70%	36.70%	26.60%

[PAULM03]

Re: A football rating system, discussions, ideas just a thought, but could those stats be something to do with the way the fixtures are scheduled? i.e. the games that sky want to show on those repective days?

[Osesame]

Re: A football rating system, discussions, ideas I believe this has been discussed before along with other thoughts on the cause of these variations.

[gazabroad]

Re: A football rating system, discussions, ideas Isn´t that how the fixtures are drawn? At the SKY HQ?

[Osesame]

Re: A football rating system, discussions, ideas Why should that be causal of such variations anyway?

[Osesame]

Re: A football rating system, discussions, ideas I believe they have to give a fair share to each team and the fixtures are arranged a long time before.

[gazabroad]

Re: A football rating system, discussions, ideas Yeah right!

[mr onemore]

Re: A football rating system, discussions, ideas I've seen that data before - certainly surprised me the first time - think I posted to try and give an explanation - but truth be told I don't have a good one to explain that...

[Osesame]

Re: A football rating system, discussions, ideas Average odds

	home wins	all matches
mon	2.13	2.52
sat	1.99	2.44
sun	1.99	2.36
tue	1.88	2.42
wed	1.76	2.6

[gazabroad]

Re: A football rating system, discussions, ideas In other words, There´s no real difference, as we said all along:\

[Osesame]

Re: A football rating system, discussions, ideas There is not much difference between average odds for all matches on mons and weds but I would suggest that it is a big difference in the actual win odds

[gazabroad]

Re: A football rating system, discussions, ideas Sorry you´re right, I´m still trying to get used to these decimals

[mr onemore]

Re: A football rating system, discussions, ideas Maybe this thread was a little too ambitious - there is just so much to cover here - and I've not ever tried to explain everything in one go. It doesn't help that the ideas are all over the place in this thread at the moment - I'll need to stay more focused to get my own points across, and more importantly for me, have a chance to take on others views here - which is why I want to do this after all. Here is quite a well-known quote that I rediscovered tonight and kind of sums up my feelings about this thread at the moment - don't know what the dolphins have to do with it though:) A thoughtful and provocative quote from educator John Holt (1923-1985) from How Children Learn:

We teachers - perhaps all human beings - are in the grip of an astonishing delusion. We think that we can take a picture, a structure, a working knowledge of something, constructed in our minds out of long experience and familiarity, and by turning that model into a string of words, transplant it whole into the mind of someone else. Perhaps once in a thousand times, when the explanation is extraordinarily good, and the listener extraordinarily experienced and skillful at turning word-strings into non-verbal reality, and when the explainer and listener share in common many of the experiences being talked about, the process may work, and some real meaning may be communicated. Most of the time, explaining does not increase understanding, and may even lessen it.

Yes, that could be happening here!!

[mr onemore]

Re: A football rating system, discussions, ideas Right, so there's been a bit of an intro, a few of the big questions touched upon - will come back to those in due course. For those who want to carry on here, I would like to go back to this - I'm going to try and finish this bit before moving on anyway... So after the intro, I think I should call this Chapter 1 feels like could write a book about this anyhow...

Well here's the first big problem for any rating system - and it's unavoidable. I'm going to use this very simple metric, that the ratings depend on how many goals one team is better than another. Let's say Team A beats Team B by 2-0 Then later, Team B beats Team C by 3-0 Team C seem the worst, let's give them a rating of 0 The metric will produce a rating list like this; Team A 5 Team B 3 Team C 0 I hope that sounds reasonable so far. If that's all there is to go on then from any rating system, by our fundamental premise, will might predict that When Team A plays Team C, Team A will win by five goals. After all, the metric used here is to measure strength in relation to how many goals a team is better than another. If this happens, we keep the rating list, and confidently use for a later encounter between the teams. Of course, and this is why we like football after all, that's unlikely!!!! What if this happens... Team C beats Team A by 1-0 How do you update the ratings now? What would the rating list look like? Well, that's the gist of the problem. As I say, there is a good mathematical way to solve this problem - I first saw it over 20 years ago - and it's used extensively by all the serious ratings lists now produced. I'll post it up later - in the best traditions maybe those that haven't seen this before might like to have a go!!

[mr onemore]

Re: A football rating system, discussions, ideas Going to sneak in a bit of maths here - I think it's necessary to understand the method, but you can still apply the method even by skipping this bit. If anyone wants to follow, it's just a little algebra, and some matrix work, which can be done very easily on a spreadsheet even if you have forgotten or never really knew what matrices are.... Using Ra, Rb and Rc to be the ratings of each of the three teams, and a metric that measures performance in terms of the difference in goals scored between the teams in the match, we can form the following equations... Team A beat Team B by 2-0 Ra - Rb = 2 Team B beat Team C by 3-0 Rb - Rc = 3 Team C beat Team A by 1-0 -Ra + Rc = 1 It's also necessary to fix the scale, have a base rating. Of the many ways, I prefer this one - putting the rating of the last team on the list to zero. So there is a fourth equation, Rc = 0 So, there are 4 equations, and 3 unknown ratings here...this is an over-determined system of equations - and as such there is generally no exact solution. In other words, you won't be able to give a rating to each team here that exactly models the events as described. Paul & Paul - if you are reading this - that might demonstrate why no rating list will ever be a 100% accurate descriptor of the events that have happened - it can't be done (full stop!) Anyhow, we can still find a best solution to this system of equations, one that minimises the errors. The preferred tool of mathematicians on this situation is to use a least squares approximation, which is a little bit like drawing a line of best fit through a set of points that you suspect should lie on a straight line but don't. There is a little bit of theory I'd like to put in now for completeness...though this next bit can certainly be skipped if you're put off by it!!

[mr onemore]

Re: A football rating system, discussions, ideas (btw this is necessary theory if anyone wants to combine metrics later on - I never have though for reason given earlier...) So, to start off write the first three equation in matrix form - I hope this layout comes out right... | 1 -1 0 | |Ra| |2| | 0 1 -1| |Rb| = |3| |-1 0 1 | |Rc| |1| but, from the fourth equation, there is Rc = 0, that gives | 1 -1 0 | |Ra| |2| | 0 1 -1| |Rb| = |3| |-1 0 1 | |0 | |1| which effectively eliminates the last column of the first matrix, and the last row of the second matrix (that's matrix multiplication for you!!) | 1 -1 | |Ra| |2| | 0 1 | |Rb| = |3| |-1 0 | .........|1| Now a bit of matrix algebra!! It will be easier to use A for the first matrix, r for the second matrix, a column matrix holding the ratings of Team A and Team B, and g for the third matrix, a column matrix holding the goals difference in each match played. So, we now have the much easier to write, A.r = g To find the least-squares solution for this, first pre-multiply each side by the transpose matrix of A, I'll call that (At) as I can't do superscipts here - the PL is not really set up for writing maths Paul!! At.A.r = At.g Now pre-multiply each side by the inverse matrix of At.A - let me call that inv(At.A) inv(At.A).At.A.r = inv(At.A).At.g giving r = inv(At.A).At.g So there it is!!! To get the ratings, just do that series of matrix calculations. Any spreadsheet can be used to crank that out. Put the matrix A into some cells, find A transpose using the functions in the spreadsheet, then working from left to right in the final equation above, multiply the inverse of (matrix A transpose times matrix A) times matrix A transpose times matrix g. Maths is so ugly when written in English... Well anyway, I get this Ra = 1 Rb = 1 and don't forget we put Rc = 0 Have a go - it's not so hard really...it's good to experiment a bit once the matrices are set up on a spreadsheet too, to get a feel for the solutions. The rating list we have now would suggest that if Team A now played Team B, we'd expect a draw - even though Team A won the first game 2-0. (There's loads of refinements to add before using this though - but I really do want to leave that until later...)

[mr onemore]

Re: A football rating system, discussions, ideas Now, just for fun, I think that I'll post up the rating list for the Premier League for this season using this metric. Paul - this will put us at your stage with your performance ratings - simply a rating list that if anyone wants to use they have to use their gut feelings as to what the numbers mean. In the long run, this is not acceptable to me as a system, so we'll soon look at a way of using ratings like these to make odds predictions with a more objective basis. Also, you have to remember that this solution at present treats all games of equal importance - right from the start of the season to the one played last week. Some people like to refine that, weighting more recent games as more important than games played say 6 months ago. Still, that's all 'fine tuning' as far as I'm concerned at present. One other advantage of this least-squares solution is that it takes into consideration the relative strengths of the opponents a team has played - it's built in automatically in the theory by definition - any team that has played a harder schedule might find itself with a better rating than someone might expect - although as we're only just past the half-way stage of the season and everyone just about played everyone else at this stage the schedules have all been fairly similar. I going to run this to the start of the season too - someone might prefer last 6 games only, someone else 11 games - well, I'm just taking all the games this season for this example. But before posting the ratings, there is one quick refinement that we can make on the performance metric that I always put in - the home advantage factor - which I use to be a constant for all teams (before any comments about how it's more significant for some teams than others - yes, ok, you can do it like that if you want - just we're keeping this simple - fine tuning like that is something I just don't bother about at thie stage - actually I never do that - but you can do it if you must...)

[Datapunter]

Re: A football rating system, discussions, ideas

I am a teacher by profession after all

well who would have tought that ;) i'm loving this thread :ok will have to get the old algebra books out thou, :\ any chance of an example spreadsheet ? would be a great help.

[brookes]

Re: A football rating system, discussions, ideas Very interesting thread, been following from the start and was going to wait for more details , but can't wait . What about " new manager syndrome " and local derbies which last season I would have avoided. After joining P.L. and getting a basic understanding of value I follow these with interest and even considered a strategy Keep it coming Mr O

[mr onemore]

Re: A football rating system, discussions, ideas

well who would have tought that ;) i'm loving this thread :ok will have to get the old algebra books out thou, :\ any chance of an example spreadsheet ? would be a great help.

Yeah - it might be good for and especially later, to be able to give a spreadsheet out - how would I do that? Also, Datapunter, I'd like to put in some tables at some point but they never come up right here - how do I do that?