The Trouble With Football Statistics
Liverpool travel to the West Midlands on Sunday looking to avenge the 3-1 home defeat that Aston Villa inflicted on them in December. I wrote a preview of the match (which you can read here if you're so inclined) which looked at the stats for the two teams' campaigns so far. The numbers suggested a comfortable home win was likely.
The problem is, so did the actual stats from the match, yet the Reds were 0-3 down at home for the first time in eighteen years and were soundly beaten. Is there therefore really any value in looking at football stats?
In the above mentioned preview, I noted:
Brendan Rodgers’ men have averaged 5.3 shots on target per home game this season, whilst Villa only manage 2.5 per match on the road, which is the lowest figure in the Premier League at present... (Villa) have converted just 23% of their shots on target this season, whilst the Reds have managed to net 27% of theirs.
Whilst the shots on target figures weren't quite as far apart as this pro rata, Liverpool still had 57% of the SoTs in the match (with eight to Villa's six), and in the five previous seasons the Reds have only lost 19 of the 116 league matches when they've had more than half of the SoTs. Eight shots on target is a very decent tally by the way; since August 2008, Liverpool have won eight of the other ten matches where they've had that number of SoT.
The shots on target stats went to form, yet Villa won comfortably. Although I didn't mention them in the preview, the final third passing stats also suggested the Reds should have easily triumphed.
Liverpool had their season high figures for both final third passes attempted (254) and completed (204) against Villa, with an accuracy of 80%; to give that context, teams that finish in the Premier League top four in the last two seasons have averaged 140 final third passes per match at 73% accuracy, so in numerical terms the Reds put in a very decent performance in this area. The Villans were only accurate with 60% of their passes in Liverpool's defensive third (with 55 of 92 finding a teammate). The difference was clearly in what Villa did with the ball when they did have it in the final third.The defensive stats I highlighted in the preview were also pretty accurate:
Liverpool have conceded the second fewest shots per game (9.5) on their own patch during this campaign, whilst Villa give away nearly twice as many (17.1) per game when away from Villa Park... As Liverpool have allowed their opponents a divisional best of just 39% of their shots in the Reds’ box, then these numbers suggest Villa will only have three or four shots within the area, so the chances of a Liverpool clean sheet should be strong.
The Reds had 29 shots, so a whopping twelve more than Villa concede on average, but the Liverpool defensive stats were very close: eleven shots against in total, with five (45%) in the box. Statistically, the match went to form in the key areas; the main difference being that the team with four away league goals prior to the match bagged three in ninety minutes at Anfield.
This is not to say that I suddenly have no faith in the value of stats with regards to football, as over a long-term they will give very reliable indications of how a team is performing.
But the Liverpool-Villa match from earlier this season showed that, now and again, all the positive stats in the world won't determine a football match; who scores most goals will still always do that. Please check out my other articles, and follow me on Twitter or via Facebook. Thanks.