Most toss statistics are quietly measuring which team was better

The win rate for sides batting first at a venue is one of the most contaminated numbers in cricket, and it gets quoted every week.

A broadcaster puts up a graphic. At this ground, teams batting first have won sixty two per cent of matches. The commentators nod. The captain who has just chosen to chase is now, in the narrative, already behind.

That number cannot bear the weight being put on it.

Start with the obvious contamination. The toss is random, but the decision is not, and the sample is full of matches where a stronger side chose the option that suited them. Strong teams win more often for reasons unrelated to the coin, and if strong teams at that venue have tended to bat first, the batting-first record inherits their quality. You are looking at a correlation between two things that were both caused by something else.

Then there is the venue drift. A ground's record accumulated over fifteen years includes surfaces prepared by two or three different curators, a drainage renovation, a change in ball manufacturer, an outfield relaid, and a shift in what a par score even means. Averaging that history into one percentage assumes the ground is a constant. It is not. The last eight matches are worth more than the previous forty and nobody presents it that way, because eight is not enough for a confident graphic.

Weather and scheduling contaminate it further. Day-night matches at a ground where dew arrives in the second innings are a different game from afternoon matches at the same ground, and yet they sit in the same pile. Rain-affected results with adjusted targets are usually in there too, and adjusted targets have their own bias.

So what should a captain use. Honestly, the surface in front of them, the forecast, the state of their own bowling attack, and whether their batting order copes better with a target than with an open-ended innings. That last one is a real and underrated consideration, because some middle orders are much better at arithmetic than at judgement.

The one place the historical record genuinely helps is where the effect is huge and mechanical, such as a venue where evening dew makes the ball impossible to grip. That is not a statistical finding. It is a physical one, and you would know it without the graphic.

Everywhere else, the percentage is a comfort object for people who have to fill four minutes before the coin goes up.