Our stats guru looks at why sometimes statistics aren't all they're cracked up to be. It's all about context...
After watching Scott struggle through a quagmire of useless statistics related to defenders and goalkeepers, it made me realise more than ever that there are so many holes in supposedly accurate statistical models, and that without context you might as well just be making the numbers up. Defensive statistics are probably the most misleading of all the individual player stats, because so much depends on the style of play of their team. Chico being the 2nd best ‘statistical’ defender in the Premier League last season (link) and Curtis Davies being the best ‘statistical’ defender in the Premier League this season (link) tells you all you need to know. Neither of these players will be playing Champions League football next season and that tells you the top managers and defensive coaches see through the bullpoop[ed] of defensive stats. To quote my favourite player stats dude on Twitter @mixedknuts; “Stats and centre-backs are not a healthy match yet.”
There are many examples of these raw stats not being representative of the true ability of defenders, but the Swans are a great general example of how, for the moment, you can only really judge defenders by isolated moments rather than number crunching. Our fixation on controlling possession being the most important aspect to our game for several seasons has meant that our defenders are much less likely to be involved in ‘defensive actions’ in comparison with a Tony Pulis style centre back. If your team are set up to be ‘under the cosh’ for large parts of the game, your defenders will be a lot more active than teams that are set up to control the game.
If, and it seems very likely to me, we are mostly playing a diamond midfield with 2 strikers next season, we will find ourselves under a lot more defensive pressure than in recent seasons, and this style would be very unlikely to place Chico near the top of best statistical defenders in the league imo. The diamond midfield can also put the full backs under a lot more pressure because there is so much space in front of them, and I’m not so sure Rangel will deal with that too well next season, statistically.
From what I can see (and I’m no coach) ‘defensive actions’ happen all over the pitch, with every player, especially in recent seasons with a rising tendency for teams to play an uber pressing style (Atletico/Dortmund for example), and if this isn’t taken into account, the actions a ‘defender’ carries out are not really relevant. Imo the two best ‘defenders’ I’ve seen in European football are Lionel Messi and Diego Costa.
So, until there is an extremely complicated algorithm made to take into account all these extra considerations when evaluating a defender, I think the old school football brain is the best way to evaluate them, especially central defenders. Do they make the right decisions under pressure, do they get caught out of position, are they prepared to stick their head in when it counts, are all questions which will never be answered by statistics imo, and that is why we have the Monk.
Both Ashley Williams and Steven Caulker are tall, strong, fast and good in the air and I think the rest is left to the evaluation of professionals with experience of what’s needed, and again, that’s what the Monk is for.
However, I might as well give my opinion, and that is that there is very little difference between the two players except age and leadership qualities. Caulkers age means that he has the advantage in pace (I think), but if Ash were to go, we would really miss his presence on the field and although I think Leon is the natural replacement captain, the Liberty would be a much quieter place.