Storeytime
- Thread starter peterstorey
- Start date
And the handicap is:
Top 6: Arsenal, Chelsea, Liverpool, Man Utd, ManC, Spurs.
Bottom 5: Blackburn, Norwich, QPR, Swansea, Wigan.
Midtable: the other 9.
Win all homes (57)
Lose away against top 6 (0)
Point away against midtable (9)
Win away against bottom 5 (15)
81 points so probably won with a +4.
Top 6: Arsenal, Chelsea, Liverpool, Man Utd, ManC, Spurs.
Bottom 5: Blackburn, Norwich, QPR, Swansea, Wigan.
Midtable: the other 9.
Win all homes (57)
Lose away against top 6 (0)
Point away against midtable (9)
Win away against bottom 5 (15)
81 points so probably won with a +4.
So you're on target for a title?
Easy par we should have got an eagle.
10 year plan
***REQUEST TO THE MODS***
Can we please have a zero tolerance policy in this thread for gimps coming in saying, "Ha ha is this the table where Arsenal win the league, personally I use the actual table, which shows, y'know, lol, reality..." etc.
If you're not interested in the Storey Table, fine, feck off
Obviously a certain amount of gooner-baiting becomes appropriate later in the season when the Storey Table itself rules them out of contention and then pete claims it's an illusion caused by the vagaries of the fixture list.
Can we please have a zero tolerance policy in this thread for gimps coming in saying, "Ha ha is this the table where Arsenal win the league, personally I use the actual table, which shows, y'know, lol, reality..." etc.
If you're not interested in the Storey Table, fine, feck off
Obviously a certain amount of gooner-baiting becomes appropriate later in the season when the Storey Table itself rules them out of contention and then pete claims it's an illusion caused by the vagaries of the fixture list.
I like the concept but it's not nuanced enough
It's very useful in the early stages of the season when the table doesn't give you any perspective at all of how teams are actually doing. Plech's request should be upheld.
Let's give what Plech said a go.
Ignore the thread if you're not interested.
Ignore the thread if you're not interested.
Agree with Plech I think this table is a good barometer of early season form.
Will this be where pete is also keeping count of referee mistakes that give/cost teams points
I feel there needs to be more nuance for this to truly be a success - essentially what we're looking at is the points a team should expect to be getting from a game, and these more to that than the relative size/standings of the club
The problem is you!!!!!
It's aim is to be cheap 'n' cheerful. You could add complexity like allowing for travelling distance, factor in playing away after CL etc etc but I don't think it would be worth the effort. In my experience with other models increasing complexity doesn't bring the expected benefits until you get close to a model which approaches the entity you're trying to model.
Feel free to come up with your own system. Various Scandinavians of ambivalent sexual orientation did so last season, to widespread acclaim.
In an absolutely ideal world the standings would change over the course of the season as the real quality of the teams became clearer, with the points changed retroactively. But that's a pain in the arse to do.
Isn't that essentially impossible and defeats the purpose of attempting to understand the league in the fledgling stage.
I don't think so. In the early stages it would stay the same, then as the weeks went by you could start feeding the real standings into the ratings, and adjusting the scores from previous weeks. It would be a kind of hybrid of the real and Storey tables, which both canceled out fixture congestion and went some way to solving the problem of having to rely on ratings from last season that might not apply very well.
But I'm guessing there are fancy statistical tools which will get you pretty much the same effect without the legwork.
But I'm guessing there are fancy statistical tools which will get you pretty much the same effect without the legwork.
Yeah I guess so, it's just at what point does the table start to become accurate I guess around 19 games assuming people play an even split of teams from the respective divisions of top/middle/bottom home or away.
So we're top now. So far so good.
Tables.
You wait til I add my ref adjustment factor.
That's in the spirit of the Bayesian approach where you feed in a prior distribution (the old bands) and then use the collected data (the results) to calculate a posterior distribution (the new bands). It sounds complicated and I doubt it would improve the Storey table by much.
One thing I would say is that a home win against another top 6 side is better than par: maybe plus 1, with minus 1 for the defeated team.
Winning all home games and drawing all away games produced 63 points in the old 22 team league, and gives 76 points now. 63 points was usually sufficient to win back then but 76 points never is now. A sign that the top teams produce better results these days?
How does this work then?
Edit: I've just noticed it says Arse Champions 2011/12...it kinda sets the whole thread off on a delusional tack doesn't it?
Edit: I've just noticed it says Arse Champions 2011/12...it kinda sets the whole thread off on a delusional tack doesn't it?
So let me check if I understand Pete's system here.... the results so far give us:
United +2 (won at WBA where par is a draw)
Chelsea 0 (drew at Stoke = par)
City 0 (won a home game)
Spurs 0 (but they haven't played yet)
Arsenal 0 (drew at Newcastle = par)
Liverpool -2 (drew a home game where par is a win)
Is that right?
United +2 (won at WBA where par is a draw)
Chelsea 0 (drew at Stoke = par)
City 0 (won a home game)
Spurs 0 (but they haven't played yet)
Arsenal 0 (drew at Newcastle = par)
Liverpool -2 (drew a home game where par is a win)
Is that right?
I think so yes.
Yep, unless you do the ref-adjusted one where Chelsea and Arsenal are both +2 as well.
Don't get all bitter and twisted after the first game pete! I think there's one thing we can all be sure of this season, the refs will be rubbish againInteresting. I must say par for beating the top 4 at home has always seemed a bit stingy to me. A home draw with Chelsea feels intuitively like as good a result as an away win at Villa or Bolton.
pete, we're trying to keep the barbarians from the gates here but if you're going to keep on with this adjusting for refs talk it's not going to be an easy task.
I love you peterstorey
But the intention is to suggest how many point you should get from that game. What you're suggesting is that we should call that win as worth 4+ points, which completely changes what this table is trying to do, and takes it away from the realms of the realistic.
Also, when the mods go through and remove all the spanners complaining about Peter being a spanner, can they remove all the spanners complaining about the spanners complaining about Peter being a spanner?
I'm just thinking about this idea. I tried to follow it last year but struggled because my simple brain couldn't hack it. Seeing as this is seemingly becoming a Caf mainstay (and rightfully so!) could we have a large OP explaining the point system and also the rules from within the thread (such as Plech's).
If it works out and anyone would be interested, I wouldn't mind making a spreadsheet or program to work out the table for us automatically. To be honest, I think there are loads of people here more than talented enough to make the system automatic and spare pete the time.
If it works out and anyone would be interested, I wouldn't mind making a spreadsheet or program to work out the table for us automatically. To be honest, I think there are loads of people here more than talented enough to make the system automatic and spare pete the time.
Pete should like this. It sounds like another way of putting an Arse spin on it...
It's really not very complicated. You start by defining a set of results against specified competition which overall give a potentially title winning points tally.
Then you compare each actual result to the defined 'par' results, and therefore get a total of points versus par.
Comparing different teams points v par gives you an idea of how they're performing against each other, despite them not having played the same games as each other.
Dammit - I've actually just managed to make it sound really complicated. But it's not.
It's easy really. Although it's been described in mind bending terms at times...
From Pete's second post:
A title challenging team is expected to achieve 'par' results against the other teams as follows:
Win all your home games regardless of opposition
Lose all away games against the top 6
Win all away games against the bottom 5
Draw all the other away games (against the middle 9)
Compare actual results with 'par' to get a plus or minus score.
Simples...
The remaining discussion is about when or whether you should adjust the rules and start using the real league table to decide which teams constitute the top 6, middle 9 and bottom 5 (rather than Pete's reasonable but arbitrary starting position)
Cheers lads. I think I did myself a disservice in my other post because I actually seemed to understand all of that, I just wanted someone to hold my hand and tell me I was right.
Just to clarify, if you then lose at home to any opponent, it's minus one or two? Equally, winning away at a top six team is plus three? This is the bit I want to know in particular, is it based on the actual table or does the plus or minus sit in its own table environment.
Anyone interested in digitalising this system?
You could set up a 'spanners' table...
Cheers lads. I think I did myself a disservice in my other post because I actually seemed to understand all of that, I just wanted someone to hold my hand and tell me I was right.
Just to clarify, if you then lose at home to any opponent, it's minus one or two? Equally, winning away at a top six team is plus three? This is the bit I want to know in particular, is it based on the actual table or does the plus or minus sit in its own table environment.
Anyone interested in digitalising this system?
It's the difference between the points you would get for par and what you actually get. So losing at home is -3, regardless of the opponent. Drawing at home is -2. Winning at home is zero. Winning away at a top 6 club is +3 as you were expected to lose, drawing away at top 6 is plus 1, losing would be zero.
It's based on Pete's arbitrary 'final' table in post 2 but some of the discussion is about whether at some point in the season you should change it to the actual table (when you'd have to recalculate all the results so far against the revised top, middle, bottom definitions.)
Re digitalising the system, I'm not putting my fingers anywhere near it...