97 – An Uneven Battle – Post 37

It struck me, when I was looking through the winning percentage of football clubs, that Chelsea’s % since 2000 would be incredibly high; so it proved – the sample I took was .737 over a few years, which is high indeed. Then I was reading my Bill James book on the way into work and this idea, the idea of league dominance crept into my head. One of the stats Bill uses is a five-year total of Win Shares to rank who has performed the best over a certain length of time. I decided to bring that idea over to football, but used both 3 years and 5 years as my bases, just to see if there was a marked difference.

Here is the Top 10 3 Year Winning Percentages, all time, in the English league – the year you see is the final year of the middle season.

Chelsea	2006	114	0.820
Manchester United	2008	114	0.803
Chelsea	2005	114	0.803
Chelsea	2007	114	0.794
Manchester United	2009	114	0.789
Arsenal	2003	114	0.789
Manchester United	2007	114	0.785
Chelsea	2008	114	0.776
Chelsea	2009	114	0.776
Arsenal	2004	114	0.776

That’s 10 teams, all in the last 10 years, more dominant than ANY teams, EVER. The first non-big four team is the Preston of 1890 (the invincible Preston the season before, and the title that season too). It all seems crazy that now, more than ever, we’re watching football that is almost written before we start. Look at the ratios of the top 3 – above .8 – that means they’re going to win over four out of every five games (draws not-with-standing); how is that possibly good for the game?

We’re here for Huddersfield Town though. I thought that we might have a chance, actually, of being pretty high in this comparison – winning three league titles in a row, and having two runners-up spots after that; but 97th is the highest spot we muster – the 3 years of the 3 championships – though admittedly the first one was horribly close. What we see then is that since, effectively, Jose Mourinho took over at Chelsea, there is very little interest in anything beyond the top two teams, and their paths to glory have been unimpeded until…well, until about now. We know Manchester Utd won the Premiership this season with a stonking home record and a lot of away draws – they actually sit level with the bottom of the table above on 0.776. Chelsea, though, are down to 0.732, and I think we can say with some certainty, that they’ll face more of a challenge from Manchester City than ever before next season.

I wanted to give the figures a bit of context, so I looked abroad for successful teams, my gaze first falling on the feats of the Grande Torino. Over two three year periods (centred on 1947 and 1948) the Grande Torino were 0.817 and 0.810 – which would bring them in second and third to the Chelsea team of 2006. The 1946 vintage were 7th, at 0.793. That confirmed my knowledge of how superlative that team were and much of a travesty it is that they’re not better known in this country. Never-the-less, there’s other movements going on, currently, that might just be as dominant as the Chelseas and Uniteds of the mid-noughties.

FC Barcelona have, focusing on their most recent La Liga season (2009/10) a winning percentage of 0.851; which, put into context means that, if their games aren’t draws (of which they’ve had 18) there’s an 85% chance they’ll win ANY game, not just home, but away games as well. That’s a level of dominance we’ve never seen before. The tragedy, and its a tragedy that is something like the argument tennis players have that they were playing in the time of Federer, is that Real Madrid’s score for that same period is 0.816 – good for third on the list by a mere 0.001 places.

This high-falutin’ nonsense isn’t something I often spend a lot of time thinking about, but the number astounded me when I saw them – for the teams today to be getting more and more dominant is a worry to me, its a worry not just about how bored I am seeing the same teams winning things, its a worry about the very state of the game that all things will run through these elite clubs – as if it doesn’t already.

Anyway, back to Huddersfield Town.

Outside the 1920s, the best performance over three years by a Town team surrounded the 1979/80 season, unsurprisingly, and saw the team having a % of 0.612; bang on level with Bradford City’s highest mark – centred a season later. On that same 0.612 figure, however, we see Lee Clark’s team of 2009/10. Time and again when I work out the numbers, this team is as successful as any Town team outside the 1920s and that 1980 Championship winning team. Another strong season, and that mark might even be surpassed – they’ll have to go some to get up to the 0.683 of the 1925 team, but I don’t see that its entirely out of reach.

Meanwhile, the lowest Town mark was centred on the 1987 season (old ground, isn’t it) at 0.383 – that’s a pretty atrocious mark, to be honest – only 287 three-season spells, out of more than 8,000 have been worse. Never again, let’s hope.

2.03 – Kevin Kilbane And AMMP – Post 34

I saw Kevin Kilbane play a few times last season, and although there was a couple of amazingly good passes in the game against Sheffield Wednesday, his other performances were pretty forgettable – I thought he was awful in the playoff final; so much went astray, and he never seemed to be anywhere he needed to be.

One of the things I’ve been working with lately (which I posted some information from a wee while ago) is the minutes each player features for Town. Using that data set, now complete for the 46 league games, I can see how many points each player won; in terms of minutes on the field. Peter Clarke, having played all 90 minutes of all 46 league games, can claim 87 points, as can the two goalkeepers between them.

The formula I’ve used to work this ratio out is one I’ve called AMMP. It might sound complicated, but it only stands for Appearance Made / Minutes / Points, which is basically how I worked it all out. There is a maximum AMMP of 3 (because there’s three points for a win) and a minimum, as you’d expect, of 0. Given that most teams average about 1.5 points per game, I’d expect an average AMMP for a first team player to be about 1.40; allowing for substitutions throughout the season – Damien Johnson was the closest Town player to that mark, with an AMMP of 1.44.

That’s useful in a few ways, but most of all it can be used as a direct comparison. Ian Bennett registered 47 points, and Alex Smithies 40 – which means Bennett’s appearances counted for an average of 1.96 points, and Smithies for 1.82. These numbers for the goalkeepers are helped by each one only playing full matches, but I don’t think its too controversial to say that it suggests what people already knew – that Bennett had been a little better than Smithies when he got into the team.

So far, so obvious. There was never any great discussion about who should be playing in goal, though, that decision was pretty much made for Lee Clark. The more controversial positions were elsewhere; up front, at centre back (Kay or McCombe?) or in midfield. This is where we should be able to draw some differences with AMMP.

Looking at the second graph here (players arranged by squad number)

of the strikers AMMP; there’s a definite points advantage to Rhodes and Afobe both playing (though note how low the figures on the left are there – nothing higher than 1.2 for Jordan Rhodes. There’s a relatively straightforward reason for the lower scores in that Lee Clark loves to change his forward line relatively early in the second half – there’s lot of 76, 65 and 45 minute appearances made. After Afobe and Rhodes, though, the other four – Novak (31 apps), Lee (28), Garner (16) and Cadamarteri (11) – are much of a muchness, ranging from 0.82 to 0.94 AMMP – the highest being Lee Novak.

I can’t help but think that the fact that there is no striker even close to 1.5 may well be one of the reasons we seemed to struggle up front. Rhodes scored a lot of goals, but was regularly substituted. Afobe played a lot of (successful) games, but scored – relatively – few goals. Alan Lee and Lee Novak would be regarded as having disappointing seasons, yet they were equally effective as Garner and Cadamarteri, one of whom was shipped out in January.

Moving onto centre backs, then. Peter Clarke is a representation of the team, as I mentioned, as he played all 46 games. Kay played 27 and McCombe played 34, so we’re not looking at small sample sizes there, and there’s a very noticeable difference. Town were more successful when John McCombe played. I quite like Antony Kay as a player, and accept how gangly and ungamely Jamie McCombe looks, but the numbers don’t lie; Antony Kay’s 1.57 AMMP is a long way lower than Boom Boom’s 1.81 – both are a little below Peter Clarke’s 1.89, but it certainly seems that McCombe and Clarke would’ve been the better centre-back pairing over the course of the season – not sure how they would’ve been different given those three goals against Peterborough. I only used Clarke, Kay and McCombe because there wasn’t really evidence that any other players were anywhere close to first choice in this position throughout the season.

This, then, leads us onto the midfield, which is, I’d say, the positions that have caused the most debate. I’ve included,

because of their playing time/importance, eight midfielders. Four I’d class as wide (Roberts, Ward, Pilkington, Carey) and four central (Gudjonsson, Johnson, Arfield, Kilbane). There’s a happy mirroring with this, too. With an AMMP of 2.03, Kevin Kilbane stands alone as the midfielder who’s performances have earned the most points this season, followed by Danny Ward, Gary Roberts and an incredibly close call between Joey Gudjonsson and Scott Arfield (1.60 v. 1.59). Those four (you’d have to take Joey, I guess, being that 0.01 higher) would make a pretty impressive midfield four, particularly if Gary Roberts was in his early season form rather than later. It is worth noting that for all his goals, Anthony Pilkington didn’t get to play in too many winning Town teams, and Graham Carey was the least successful of all the midfield. There’s mitigating circumstances, mind you, in that Gary Roberts was, for a while, the only player who COULD play on the left flank, so he was the only chance.

For the record, I’ve include an AMMP graph for the whole squad here at the bottom. Chris Atkinson gets a rough ride, in my opinion, but there’s big performers in the defence – Naysmith, Clarke, Peltier… it was a good season.

image007

image001

image005

18 – Why Does Steve Claridge Hate Us? – Post 23

Steve Claridge, it seems, doesn’t like Huddersfield Town. Every time he gets the chance to talk about us on the Football League Show, he shies away, mumbling something about budgets and how he prefers Southampton – despite being a legend at their hated rivals, which is where I remember him best.

But, I hear you cry, why doesn’t he like us? He has good memories, surely, of avoiding the drop with Pompey, winning Division 2 with Birmingham; although that is counter-balanced by Town relegating his Walsall team in 2006.

I decided to pull out all the Steve Claridge stats I could and see if there was anything that stood out. I was surprised.

In all, Steve played (or was at teams who played in the early days for when I can’t get line-ups) 18 times against Town – 9 home, 9 away. He scored 4 goals – 2 home, 2 away (again, possibly – probably – a couple more). The outcome of those games is as follows.

Home – P9 W3 D3 L3

Away – P9 W3 D3 L3

Its all very even, and strangely so. A team against whom your results have been exactly mediocre don’t shriek out as someone to take against – not to me anyway.

The only thing I can possibly think is that Aldershot played Town on 22nd October 1988 – which may well have been Claridge’s debut for them after 3 years in the wilderness. They lost 1-2. Does he bear a 23 year old grudge? I wouldn’t put it past him.

If anyone can provide opposition line-ups/scorers for Town games before 1995; I’ll update more accurately.

39 – Weighted Top Of The Table Update – Post 21

Things have changed at the top of the Weighted League One table, and the news is things have swung in favour of Southampton; their games in hand/points potential probably bears this out, but we now see the following for the top four teams. Brighton are pulling away now – they’ll beat anyone and everyone at the moment. Then, where previously, Town have shaded it, Southampton have crept ahead of them.

Brighton & HA	9.297
Southampton	7.083
Huddersfield T	6.372
Peterborough U	6.092

On the strength of that turnaround, I make Southampton narrow favourites – their problem, clearly not having the talent to win games, but having the depth of talent to win as many games as they need to within the month.

For the record, had Southampton played 39 games, they’d still be a hair ahead of Town on 6.532 – this is what happens when you beat Bournemouth but lose to Walsall.

6.629 – The League Table Never Lies – Post 14

It’s a cliché, but often held to be true, that the league table never lies after 46 games. I’m contrary by nature, but wouldn’t necessarily go completely against it. I would say ‘the league table can’t be argued with after 46 games’; it is final, it is recorded, and it goes in the books. The league tables would differ, for example, against those from the days of two points for a win, and if goals for was used instead of difference. It is these things that crop up during the season (Peterborough’s massive goals haul, the dearth of draws for certainly teams in League One this season) that I focus on now, and I’ll explain why.

Late last year, Leicester City were taken on by Sven Goran Eriksson and have bombed up the table since – the acquisitions of Yakubu and the likes – and sit pretty well now. As such, although I can’t argue that wins against Leicester should be worth any more than they used to be, they’re a far less common commodity, and there ought to be some way of weighting that – to see who has got more points against better teams, and who is just a flat-track bully. To which end, I’ve been thinking of a few ways of weighting the league table to represent the strength values of the points acquired.

This is one of the methods I came up with. Brighton have been top of the league for a long, long time; wins against them are less likely than wins against Walsall, who’ve been struggling all season. Southampton, though, started pretty poorly, and have come on since then. Wins against them earlier were more common than they would be now. To cover this changing team status, I’ve used the league table as it stood at the end of the games between two teams and worked with the following method.

If a team in twenty-fourth place beat a team in third, they would receive 22 points (basically the league position inverted); the same as a team in first place would if they beat the same team; by being beaten, that team in third would probably drop positions, so the next time they were beaten the team who beat them would probably receive only 20 or so points. Draws are tricky to incorporate – I thought about leaving them out as null matches, but settled in the end on using half the points that would be gained for a win – this has resulted in Oldham’s position changing dramatically – as explained below. The two figures; wins and draws are combined, and divided by the amount of games played – defeats have to be taken into account, too.

This should flag up, then, if there’s any teams who are generally beating better teams and losing to worse ones – their PSV (I needed something to label it – Points Strength Value) would be higher than someone on the same points who only beat teams in the bottom half. What, then, does this weighted league table show? Well, its no surprise to see Brighton at the top; they have been consistently beating anyone and everyone and would be top of most tables you could think to throw at them. Its no surprise, either, to see that Walsall, Yeovil and Dagenham are near the bottom; Dagenham’s recent upturn in fortune has seen them probably a little better off than they would have been at Christmas.

The big surprises to me are in 2^nd, 17^th, and 21^st place.

Charlton aren’t a bad team, obviously, but their league position has been pretty varied throughout the season. Their 2^nd spot here must be on the strength of their good runs – when they’re good, they’re very very good, but when they’re bad they struggle just as much – they’re currently on a good run. Oldham, who are similarly placed in the league table to Charlton and (by all accounts) quite a good team on their day, but have drawn two more games than they’ve won; that suggests that as the gaps between the top eight grow wider, Oldham might find that a point against the better teams isn’t enough, and their one win against a top half team would suggest they’d struggle in the playoffs (Charlton, by comparison, have five wins against the top eight). MK Dons position is similarly low compared to where they find themselves in the league; they’re in the same boat as Oldham but buoyed by fewer draws against the top half. Of course, there is an element of chance in this – MK Dons couldn’t beat Brighton if they weren’t playing against them, so the fixture list, and the time of year is critical.

One thing this PSV table doesn’t do, sadly, is sort out the battle for second spot; Bournemouth seem to be a little over-valued in the league, but Huddersfield and Southampton are just as equally matched. In this as, one suspects, the league itself, the two games Southampton and Charlton have against each other may well prove pivotal.

League One PSV Table (at 25/02/2011)

Brighton	7.793
Charlton	6.759
Huddersfield	6.629
Southampton	6.603
Bournemouth	6.516
Hartlepool	6.468
Exeter	6.097
Leyton O	5.897
Peterborough	5.790
Colchester U	5.484
Brentford	5.483
Carlisle	5.400
Rochdale	4.828
Sheffield W	4.483
Plymouth	4.453
MK Dons	4.424
Swindon	4.406
Bristol R	4.323
Tranmere	4.167
Notts C	4.148
Oldham A	4.109
Yeovil T	3.833
Dagenham	3.655
Walsall	3.219