Statistics for the All Time League Table

Statistics used for the All Time League Table

There are lies, damned lies and statistics.

To be sure, changing the way these football tables have been calculated would have changed the outcome. However, since most variations come about depending on how you treat time outside the league, the nearer the top you get the less the wiggle room.

If your team is one of the West Broms and Blackburns that have been in the league throughout then your average position is a cert. There is, however, no clear way of dealing with the period when Division 3 (south) and Division 3 (north) ran concurrently, but I’ve been as fair as I can be.

How the table was calculated

Statistical process

For each year of the league each team’s place has been recorded, from 1-12 in 1888-89, through to 1-92 after 1950-51. From here it is straightforward to calculate the mean and median positions.

The mean is the sum total of all league positions to date (this can be seen on the left of each table) divided by the number of seasons. It’s the most commonly used ‘average’. Its strength and weakness is that it accommodates unusual positions (like Man City’s time in the 3rd Tier in the late 90s), or can be distorted by them, depending on your view.

The median shows the middle position. Line your seasons up from the best to the worst, and the median is the one in the middle. So if your record is 1, 1, 1, 2, 3, 3, 17, then the median is 2 but the mean is 4. Choose what you like. I refer mainly to the mean, but personally prefer the median.

The SD (standard deviation) shows how much variation in position there’s been. The SD number shows how many places above and below the mean you need to allow to include 68% of the team’s positions. Consistent Everton are often within 7 places of 10th position. Meanwhile Ipswich, with an average position of 52, has a massive SD of 33. This reflects both the time spent outside the league and their time at the top.

Years outside the league

Here’s where it gets interesting. How do you account for positions outside the football league? Nailing this down would require delving into non-league records, and fails entirely for the years when a club doesn’t exist.

I’ve aimed for a compromise and allocated a ‘presumed position’.

Where a team joins the football league after 1950-51 (when the league went up to 92 teams) I have arbitrarily decided the presumed position throughout their time outside the league was 100.

Before this time the football league was enlarged on several occasions. I have assumed that the teams that were embraced were the next best teams in the country at that time. So when the league added a second division in 1893-94 I’ve assumed that the 14 new teams were (on average) the 21st best in the land up to that date (21.5 to be statistically accurate). The following year the league expanded again and three new teams joined, presumably the 29th, 30th and 31st best teams. I’ve assumed these were all, on average, the 30th best team up to that time.

Where a team drops out of the football league I’ve assumed, similarly, that they’re the best team not in there. As the league expands without them their presumed position drops, and reflects the positions of teams yet to join. Darwen, for example, are presumed to be at 13.5 for the first three years of the league before they joined. They then have genuine places for their eight seasons in the league. After relegation they are presumed to be the 37th, then 41st etc, until they join the 100 club in 1950-51 with the Morecambes and Maidstones yet to join.

There are a couple of exceptions. For example, I know for a fact that Thames Association were formed in 1928 and dissolved in 1932. It seems a nonsense for their two years in the league’s bottom three to put them above teams like Workington, who spent some seasons at the giddy heights of 50th-ish. Therefore I’ve put Thames Association in the 100 club for the long periods when they didn’t exist.

Any more examples of undeserved positive treatment will be examined on merit (Chelsea fans beware!).

Bias toward early clubs

One clear outcome of this policy is that if you join early you end up higher. A year outside a league of 28 clubs is less of a penalty than if you wait to muscle your way into the 92.

This is seen most clearly with one-season-wonders Bootle and Middlesborough Ironopolis (yes, they’re real) being higher than regulars Wycombe Wanderers and Hereford Utd. This is simply because they joined a league with fewer than 30 clubs in it.

Despite their title challenging teams of the 70s, meanwhile, Ipswich pay heavily for not joining the league until 1946.

Two division threes

In 1920 the southern League joined the football league, adding a third division (south). A year later a third division (north) was added in parallel.

I have treated these as simultaneous additions to the league, and allocated all of these teams an average position of 63.5.

Also, I’ve made the top team in both divisions =45th place, the 2nd team in each =47th place, the 3rd team in each =49th and so on. This goes down to  =83rd, =85th (and =87th and =89th after 1950). That’s why all of these numbers are odd until 1958 when Division 4 was created. The exception is the bottom team, which was made =88th / =92nd.

Don’t like how I’ve done it?

Well, all you Kidderminster Harriers fans who feel you should be above Burton Wanderers, tell me how you’d do it.

Go to:
All time league table
Premier League era table
2010s league table
2000s league table
1990s league table
Statistics used in the ATLT

All material on this site copyright 2016 - 777ok Ltd