21.74 – A Good Start (Pt 2 of 2) – Post 46

So we come to the first 10 games of the season. Now, as you’d expect, we cover about 50% of old ground here, but the two tables, at top and bottom, are far more revelatory of the seasons to come as, I guess, the better teams and weaker teams schedules ‘even out’ towards Christmas. By the end of 10 games, you’d expect a team to have gathered 21.74% of their points tally and, as before, the teams that are heading for promotion haven’t quite achieved that

ratio, and the ones that are heading for the drop haven’t quite fallen that far but, by 10 games, the gaps between the two are far closer which means that you’ve got a lot better idea by 10 games – particularly in seasons when there isn’t one team running away with it – of where you’ll end up.

So, looking at the table here you see that there’s only one season as before that the promoted teams were moving quicker than the average rate once (that – first – year that Leeds went postal and then collapsed) but the other four seasons has seen the top teams picking up on their first five games performance, but not quite looking like guaranteed promotion form (I find it strange

that the last two seasons have been the same, but that’s just me), a table that is repeated, in reverse, in the relegated teams table. The four fallers in 2006/07 fell away badly, as we’ve seen from the five game table, but other than that, they’ve not been far off the mark you’d expect – 24.55 is only around 3% above 21.74, and, as a secondary caveat to these teams, its also worth bearing in mind that a lot of these teams were recently promoted, so their early season form – particularly at home – is generally better than one would anticipate (or, in the case of Gillingham is 2009/10, is just good over the whole season).

What have we learned from these two posts, then?

Five games is still to early to draw any conclusions for the season, but by 10 games, the better teams will be beginning to show at the top part of the league, just as the weaker teams will be around the bottom somewhere.

Incidentally, I used the teams that got relegated for this chart, rather than those who gathered fewest points – Plymouth, notably, last season, skewed the figures a little, but only a little, with their creditable start.

10.89 – A Good Start (Pt 1 of 2) – Post 45

“I’m not going to look at the league table until 10 games in. Nothing’s really happened until then, and you can never tell after one game where anyone’s going to end up”

Well, now. Well, now. Is that the case? Remember Southampton’s slow start last season? Remember that Millwall were crippled by injuries and had to surge through. Remember also that Leeds were many points clear before their collapse, and Brighton were top of the league from very early on. So does that make a difference.

I’m going to take the answer to this in two parts and owe Jonathan Wilson (@jonawils) a massive debt of gratitude for the tools to solve this particular puzzle. I’m going to work with the seasons from 2006/07 to last year (five seasons, and fifteen promoted teams) to seek my answers, and hope to come to a conclusion.

How important is a good start? After 5 games, what sort of indication did we have that teams were going to prosper, or flounder? Mr Wilson worked his tallies out by comparing actual points tallies (% of final points totals) with the expected points percentages if they were gathered evenly throughout the season. I saw no reason to mess with his formula, though obviously, my sights were set a little lower than the top of the Premiership and my numbers a little more difficult to pin down.

As it happens, I was a little surprised by what I found – the variety was large but, by and large, the expected points tally (10.87%) was nowhere near matched, except in 2009/10, when Leeds’ start swept all before them, and led to the three

promoted sides gathering 10.90% of their total tallies in the first five games (as near as damn an average figure as you’re likely to find).

The other four seasons saw the promoted teams doing a bit of struggling while things levelled themselves out, from only 13 points (5.02%) in 2006/07 up to a more respectable 8.65% in 2010/11 – while Southampton dallied, Brighton and Peterborough’s houses got into order pretty quickly.

This can only help to illustrate that the extreme beginning of the season is really no indication at all of where teams will finish, as can checking what was happening with those teams who were going to drop out of the division in May, those four unfortunates. Well, I thought this was going to prove to be an opposite story at first, but it seems that Brentford’s glorious start and collapse was something of an anomaly, matched by the decent starts of

Bradford, Chesterfield and Rotherham to create the illusion that a good start generally meant a bad end to the season. Of course, as we’ve seen above, there’s a distinct regression to the mean with these early games, and the first five seem to go out of their way to prove that – 18.44% of their total tallies that the 2006/07 relegation vintage had gathered is almost double the rest of the tallies, and explains why the average figure (11.36%) is a touch over the 10.87% we’d expect if the games were equal across the season. The fact is, though, that 11.36% is a good deal higher than 7.60%, which indicates one thing very clearly – teams start a lot more evenly than they finish. The promoted teams gathered 100 point between 15 of them, while the relegated teams gathered the same figure amongst 20 – not a massive difference between 5 points and 6.67 points at the time, but that difference comes to translate into a lot at the end of the season – it would work out at a difference between 54 and 72 points, which is sizeable indeed.

Town then, don’t need to worry if they don’t set the world on fire at the start of the season. As we discovered last year, the best work is done after Christmas.

Next time, I see if 10 games into the season is more revelatory.