clock menu more-arrow no yes mobile

Filed under:

Newcastle's turnaround was totally predictable

After two months of crisis suddenly it appears Newcastle might be sort of good. By the numbers, they were sort of good all along.

Richard Heathcote

The cries of "Sack Pardew" and "Pardew Out" have quieted down. With an impressive come-from-behind victory at White Hart Lane, Alan Pardew's Newcastle United have climbed out of the drop zone. They are carrying a three-match unbeaten  streak and have won two in a row. How did they do it?

On first glance, the causes of the turnaround are difficult to parse. That doubtful back line featuring Fabricio Coloccini, Paul Dummett and either Michael Williamson or Steven Taylor is still in place. New signings Rémy Cabella and Emmanuel Rivière have been dropped, but it is hard to see how their removal would help Newcastle to allow just one goal per match in their most recent fixtures. The team's tactics have at most been tweaked and hardly overhauled. Pardew has surely worked to motivate his players, but what would cause them to listen now if they didn't early in the season? No doubt some of these small tweaks have made a difference. But for a team to go from 0-3-3 with a minus-6 goal difference to two wins out of three requires a better explanation.

Perhaps the key is something that did not change. Newcastle have continued creating good scoring chances at roughly the same rate they did through the first six matches. They have continued conceding good scoring chances at roughly the same rate as well. All that has changed during their recent run of form is only the rate at which those chances were scored.

In those awful first six matches, Newcastle attempted 89 shots and scored just six, while their opponents attempted 63 shots and scored 12. Now, obviously not all shots are created equal. Maybe the problem was finishing, but maybe Newcastle were just letting speculative shots fly from distance while allowing their opponents easy chances in front of goal. To account for the estimated quality of each chance, I have created a measure called "expected goals" which takes into consideration a large variety of factors. The location of the shot, the type of assist, the speed of the attack leading up to the shot, whether the shot is headed, and several more. (You can read my full expected goals methodology here.)

It turns out that while Newcastle were attempting slightly lower-expectation shots than their opponents, still the balance of chances fell the Magpies' way. I estimate the expected goals value of Newcastle's shots in their first six matches at about nine xG and the expected goals value of their opponents' shots at about seven xGA. That means Newcastle's expected goals difference was plus-2 even as their real goal difference was minus-6. The last three matches, then, have simply seen Newcastle start scoring goals at the rate projected by their chances created.

This is a typical result. I looked in my database for other Premier League teams in recent seasons that underperformed their expected goals difference by at least five goals through the first six matches of the season. I found 11 teams. These teams averaged over twice as many goals conceded as goals scored, leading to a goals ratio of about 0.300. (Goals ratio is simply goals scored divided by goals scored plus goals conceded.) But they averaged roughly even expected goals ratios.

How did these underperformers do over the rest of the season? They scored about as many goals as they conceded, perfectly matching their expected goals numbers from the first six matches.

Expected Goals Ratios

The green column shows these teams' combined expected goals ratio in the first six matches of the season and then over the next 32. They created about as many good chances as they conceded consistently over the full season. The red columns show their goals ratios. And there it is, a huge jump from relegation-level 0.300 goals ratios to a perfectly respectable 0.500 average.

So Newcastle's turnaround is not unusual. Clubs that massively underperform their expected goals early in the season typically click back into form and start winning matches at a respectable rate. These teams started out the season with only about four points on average, but they averaged about 41 points over the remainder of the season. So if Newcastle follow the averages, they should end up easily safe with a mid-40s points haul.

I have collected the data on these eleven teams in the table below. There is a mix, from the truly awful (Portsmouth 2009-2010) to the really quite good (Arsenal 2011-2012) and a bunch in between.

First Six Matches Rest of Season
Club xGR GR Pts/M xGR GR Pts/M
Sunderland 2013 0.419 0.222 0.17 0.423 0.446 1.16
Newcastle 2013 0.551 0.389 1.17 0.521 0.429 1.31
Queens Park Rangers 2012 0.377 0.235 0.33 0.380 0.356 0.72
Norwich City 2012 0.479 0.235 0.50 0.440 0.451 1.28
Liverpool 2012 0.568 0.429 0.83 0.665 0.667 1.75
Arsenal 2011 0.532 0.391 1.17 0.685 0.650 1.97
Wigan Athletic 2010 0.382 0.133 0.83 0.412 0.442 1.16
West Ham United 2010 0.382 0.235 0.67 0.421 0.406 0.91
Everton 2010 0.596 0.364 0.50 0.542 0.553 1.59
Blackburn Rovers 2009 0.607 0.400 1.17 0.469 0.432 1.34
Portsmouth 2009 0.434 0.200 0.00 0.436 0.365 0.88

The most concerning row in this table for Newcastle fans may be the second one. Last season Newcastle struggled to convert their share of chances as well. It is possible there is a structural problem with the club that is leading to expected goals underperformance. But given that so many other clubs improved their results quickly after slow starts like Newcastle's, most likely nothing is terribly wrong. Newcastle are what we thought they were: a defensively-shaky mid-table club. A bad run of finishing form should not be mistaken for a fundamental problem with the club.