What a website! On 15th June, in response to my “who went down, who’s going down?” Dave left a comment which included:
“… I think Pearson will, at some point, have a big row with a player or a fan, and the team’s spirit that kept them up will dissipate…”
Two weeks later Leicester City’s website contained the following:
“…it has become clear to the Club that fundamental differences in perspective exist between us. Regrettably, the Club believes that the working relationship between Nigel and the Board is no longer viable…”
The effect on team spirit remains to be seen but as I said at the start, what a website!
A week ago I left you with a poll whose results were suggesting the targets of my dodgy numbers “relegation watch” would be the three promoted clubs and Leicester. Since then another couple of thousand votes have been cast but they didn’t make much difference, despite the turmoil at the Kingpower . Here are last Friday’s and this Friday’s results:
So AFC Bournemouth, Watford , Norwich and Leicester it remains – plus SAFC, of course, although with a month to go there’s plenty of time for Dick Advocaat to move into the top half of the Premiership’s longest serving managers list, for Villa to lose Benteke and for all sorts of weird things to happen.
If you can remember as far back as last month you might also remember that I was going to track whichever clubs do get selected by means of win loss ratios.
I didn’t tell you why but I did suggest you might ask yourself what tenuous link connects Sunderland, Aston Villa, Wolves, Wigan, Sheffield Utd, Portsmouth, Derby, Coventry, Southampton and Man City. The answer is 38 points – well part of the answer is, anyway. Have a look at this table:
Premiership clubs scoring 38 points since the formation of a twenty club league:
Season | Club | Won | Lost | Position |
2014-15 |
Sunderland Villa |
7 10 |
14 20 |
16 17 |
2013-14 |
Sunderland Villa |
10 10 |
20 20 |
14 15 |
2011-12 | Villa |
7 |
14 |
16 |
2009-10 | Wolves |
9 |
18 |
15 |
2006-7 |
Wigan Sheff Utd |
10 10 |
20 20 |
17 18 |
2005-6 | Portsmouth |
10 |
20 |
17 |
1999-00 | Derby |
9 |
18 |
16 |
1995-6 |
Coventry So’ton Man City |
9 9 9 |
18 18 18 |
16 17 18 |
That’s eight seasons, ten clubs, thirteen instances when 38 points have resulted. Only two of those clubs have been relegated, and both times it was on goal difference. That vindicates my previously expressed opinion that 38 points is usually enough to stay up, but it’s not what I’m about today.
Look at the win/loss figures. (And have a double look at Sheffield United, while you’re there. Are they the only team to have 10 or more wins and still go down? That’s a must for a pub quiz).
All of these clubs have lost exactly twice as many games as they have played. In fact, any team that loses exactly twice as many games as it wins must end up with 38 points. It doesn’t matter if it’s 0 wins and 0 losses (the lowest possible), 12 wins and 24 losses (the highest possible) or any number in between, the result is the same – 38 points.
I know that’s true because I’ve taken the time to go through all of the possibilities. I don’t have a proof that the converse is true – i.e. that any team with 38 points must have lost exactly twice as many games as it has won (I’ll give it some thought on holiday when the sun’s to hot to go out, the TV’s too boring to watch and the beer’s just right for drinking) but I think it must be so, given that on the thirteen times when 38 points have been achieved every club has done just that.
Losing exactly twice as many games as you win translates to a win/loss ratio of 0.5:1, which can be written as 0.5. So there we have it. Any team with a win/loss ratio of 0.5 or more will achieve a minimum of 38 points and, with that, the expectation of safety. Chelsea, by the way, had a win/loss ratio of 8.6 last season, and Arsenal reached infinity in 2003-4 (26 wins, 12 draws, no losses), so 0.5 isn’t really anything to write home about.
Safety isn’t 100% guaranteed, however. There has to be a bit of leeway as Sheffield Utd and Man City both went down with 38 points, and let’s not forget that as recently as 2010-11 Birmingham and Blackpool both went down with 39. On the other hand, in 2012-13 Hull and West Brom stayed up on 37 and 36 points respectively.
Having taken this into account and looked at the figures for the last few seasons I’ve decided that I’ll set a win/loss ratio of 0.45 as the death line – any team doing worse than this is definitely going down. I’ll take a risk and set a win/loss ratio of 0.48 as the safe level. It’s a risk because a few clubs have done better than this and still gone down (Birmingham managed a win/loss ratio of 0.53). However, there’s often a points gap between the 17th and 18th clubs (e.g Villa’s 38 points and Hull’s 35 last season, West Brom’s 36 and Norwich’s 33 the season before) so I think I’ll be OK with a prediction that any club which consistently has a win loss ratio above 0.48 at the reviews will stay up.
And in between? That’s the unknown, the danger zone. Will your team be in it?
I’m not sure exactly when I’ll do these reviews because it will take some time for ratios to get going but the international break at the end of August may be a good time to begin, and maybe I’ll include other clubs from time to time if they look like they are falling by the wayside. I’ll probably also review each of the club’s transfers around the start of the season to add a flavour of expectation, and with this in mind we’ll keep the poll running for the time being.
If you want to keep up with your team’s progress or the lack of it point your RSS feed at this site, bookmark the page, or visit regularly. You’ll be welcome.
[polldaddy poll=8923571]
Wasn’t that safc’s plan last year?
I just thought we couldn’t win because we had three forwards who couldn’t hit a barn door between them and a midfield that couldn’t create openings . Defoe & Giaccherini’s return bode well – I hope
The proof’s simple, too. I’m now satisfied there are no exceptions to 38 points requiring twice as many losses as wins That’s my holiday ruined.
Sorry about ruining your holiday plans. You’ll have to think of another project like working out what the optimal outlay is on Bargain Hunt or something. My theory is spend as little as possible to have the best chance of winning.
Back to the footy – going through a season undefeated with thirty eight 0-0 draws gets fewer points than thirteen 1-0 wins and twenty five 0-4 defeats.
One of my criticisms of Poyet’s safety first tactics was always that by not being positive enough we were always more likely to concede a goal than score and therefore less likely to even get 38 points. The fact we got 12 points from the 9 games after Dick’s appointment shows we weren’t on course for a point a game.
With a worse GD than Hull we would have gone down under Poyet, even if he had managed a point a game in the last 9 fixtures which I don’t think he would have done.
I don’t think we would have got more than 4 the way we were playing at the time.
“I don’t have a proof that the converse is true – i.e. that any team with 38 points must have lost exactly twice as many games as it has won”
Of course it’s true because with a win worth 3 points and a draw worth 1, any combination of one win and two defeats is the same as three draws (i.e. 3 points). The total will always equal the number of games played.
So if you do it for the other Divisions, any team with a win:loss ratio of 1:2 will have 46 points.
Sometimes the maths is a lot easier than it first appears.
Try this problem: 467 teams enter a straightforward knock out Cup competition. How many ties (that is matches excluding replays) will be needed to determine the winner? The solution can be determined in less time than it takes Rachel Riley to say “one big one and five little ones.”
Correct, and very simple. I’m put in my place.
Now do it as a mathematical proof.
We have five variables: games played (g); won (w) drawn (d) and lost (l); also points (p)
We know p=(3w+d) and we know g = (w+d+l).
Derive a proof to show that when g = p (in this case 38) then l=2w.
If you do it then there can be no possibilities where 38 points can be earned and there aren’t exactly twice as many losses as wins. Don’t give me the answer until I’m back from my holidays.
In answer to your question, every team loses one match except the winners; = 466 lost games
Correct – a simple solution to what at first glance looks more complicated.