05 March 2014

Forecasting the Finish 2013-14 [Infographic]

I was asked to reprise a post from last season, where I tried to divine Liverpool's finish with both 13 and nine matches left based on points-per-game averages and the comparable results from the season before.

The 'nine games left' update was surprisingly close to where Liverpool finished, aside from the 'best case scenario' option which was further proof that I shouldn't be allowed to guess the results of Liverpool matches, even if they weren't actual predictions. But using the previous season's comparable results, as well as Liverpool's points-per-game over the last six matches, after the first five matches, and over the course of the season gave us a range from 59 to 63 points, with two of the points-per-game averages forecasting a 61-point finish. And Liverpool finished with 61 points.

So let's do that again.

As usual, click to open larger version in new window

The 4th place average since 1999-2000 is 68.357 points. The 1st place average over the same time frame is 87.643. I used that season as the cut-off because points totals were noticeably lower from 1995-96 through 1998-99, especially for the league winners, while the Premier League had a 42-match schedule in the three seasons prior to that.

This season's fourth place total will almost certainly be higher than that average. Liverpool and Arsenal need just 10 points from their final 10 games to surpass it; Tottenham and Everton need to average 1.6 and 1.9 points per game respectively to do so. Tottenham have averaged 1.9 points per game so far this season, Everton 1.78. United would need at least eight wins from its last from matches to get to 69 points, so we'll just assume that ship has sailed. Wave goodbye to it. Bon voyage, jerks.

There's also a decent chance that first place will be lower than the above average, but Chelsea's points per game this season isn't far behind that mark. The league winners have failed to reach 87 points just four times since in the last 14 seasons: United with 80 in 2010-11, Chelsea with 86 in 2009-10, United with 83 in 2002-03, and United with 80 in 2000-01.

So we've got a range from 77 to 85 points, a larger range than Liverpool had with nine games left last season, made that large because of Liverpool's incredibly impressive (and unbeaten) form since the New Year.

To reach 85 points – Liverpool's PPG since January 1 – Liverpool would need eight wins and two draws over the final ten matches. To reach 80 points, Liverpool would need either 7W-0D-3L or 6W-3D-1L. And to reach 77 points, Liverpool would need 5W-3D-2L or 6W-0D-4L or 4W-6D-0L.

Assuming Liverpool earn the same amount of points from this season's matches as they did in last season's fixtures will give this season's side 78 points. Five wins, four draws, one loss. 2012-13 Liverpool beat Sunderland (h), Tottenham (h), West Ham (a), Norwich (a), and the 3rd Promoted Side (a) [also West Ham, this season it's Palace]. They drew with City (h), Chelsea (h), the 1st Promoted Side (a) [last year Reading, this year Cardiff], and Newcastle (h). And they lost at Manchester United.

For thoroughness' sake, here are the remaining fixtures for the four sides nearest Liverpool in the table.

Nota Bene: City's games against Sunderland and Villa had to be postponed because of cup runs, and still haven't been rescheduled.

Unlike last season, it's not as if we're wanting for fourth-place models. Or maybe I've just noticed more of them. There is some absolutely brilliant statistical work being done this season, and a few excellent writers have been keeping track of each team's chances for first, fourth, etc. over the course of the season.

Bass Tuned to Red
Michael Caley of Cartilage Free Captain (@MC_of_A on Twitter)
• Constantinos Chappas (see here, here, and here)
• 11tegen11
Colin Trainor

And most assuredly a few others I've regretfully left out. Seriously, go read all of them. They're outstanding.


Anonymous said...

So, last time the forecast was the average of the top and the bottom of the range, is that your forecast this time? 81 points?

Anonymous said...

Who's got the graph that can tell us we're going to WIN THE LEAGUE?

We're gonna win the league!!

Or maybe not.

Let's give it a real good go. Eh?

Let's start by stuffing United. Fergie's back from his Oscar party. So let's show him what we're all about. What's that you say Fergs? Moysies needs more time, like you had? Yes, but he has single handedly turned a stable of Thoroughbreds and Quarter Horses into a bunch of, well, donkeys. What's that? You've had trouble with horses before? Keano wouldn't let you forget it? But then you straightened him out? No Fergs. We're not talking about Rock of Gibraltar stud fees for Moysies. We're talking about his Alchemist abilities! He's applying those skills backwards. Have a talk with him over a bottle of 30 year Rip Van Winkle. That'll fix things. Find out why he gave away the farm to that scouser Rooney, too. He'll start talking after you get him good and tanked-up.

nate said...

There was no actual, nailed-down-number forecast last time, and there isn't this time. I hate making predictions, because those predictions often turn out to be wrong.

There were four numbers based on different points-per-game averages and comparable results last season; for the 'nine games left' version, they were 58, 61, 61, and 63 points. And it just so happened Liverpool finished with 61.

This season, similar stats lead to 77, 78, 80, and 85 points. It's very likely that Liverpool's total will be somewhere in that range. But that doesn't mean that I'm willing to state it'll be exactly the median. Or make any declarative statements, for that matter.