Observer on Steve Mclaren: Lies, Damn Lies, Statistics

The Observer sports subs no doubt readied their pages before the youngsters of Israel not only escaped a late defeat but only turned round and whacked Mother Russia. The McClaren spread, presumably prepared as a good riddance send off, was led by this piece from Paul Wilson.

As he concludes:

Almost unbelievably, England are still in the land of the living.

So that's a big Hallelujah then? Hosana in the highest!

But running right across the bottom of the page are a series of tables, facts and figures. The one that leaps out as having serious Libdemologist statistics mangling tendancies is above, and headed "How McClaren Compares".

Talk about starting with a ruling theory and providing the facts to fit. There are a number of ways one might rank such a table. On notional points per fixture perhaps? On ratio of wins per game? Or, as the Observer sports subs choose, on the ratio of losses per game.

What's my point? Surely any which way will do? Six of one and half a dozen of the other? Well it's actually not that simple. Choosing to rank on losses rather than either of the other measures is actually perverse and antagonistic.

McClaren, like Don Revie, has 24% losses and the pair are a couple of percentage points clear of the field. Terry Venables on the other hand has a marvellous 4% and is far and away "best".

But rank eleven managers on wins and while not turned on its head McClaren (53%) is sixth. Firmly mid-table. Three places ABOVE Venables.

Similarly on points per game, three for a win one for a draw, McClaren is mid-table with Venables and Winterbottom. A cluster of four better 2.0. But ahead of Keegan, Robson, Taylor and Revie.

In competitive games his loss ratio drops to 18% and his win ratio is up to 64%, with a points haul of 2.09 - better than any of the scores in the all games exercise. Though of course most of the records do improve on this basis.

I'm not saying the boy McClaren's the best. Or even that he's done good. But The Observer are being perverse, snide and unfair in contorting their statistics as if they are Libdemologists.