Don Bradman averages 99.94. Nobody else is close. Cricket has been arguing about what that number really means ever since.
Fine Margins applies Bayesian modelling to Test cricket statistics, adjusting for the factors raw averages can't see.
The model accounts for conditions, opposition strength, era, and other factors that the raw figures don't capture. The result is a fuller picture than the plain stats.
It isn't to be trusted blindly. Cricket has more factors than any model can recreate, which is why every figure on the site comes with its uncertainty attached, and why we never claim 100% confidence in a conclusion.
Fine Margins is still in its early years. For now the scope is modern Test cricket, and the model itself is under construction. The plan is to go further: other formats, the franchise game, women's cricket alongside men's.
I created this project because I felt there was a gap in cricket analytics. The raw stats are everywhere, but the impact behind them takes hours of tedious research to surface.
My name is Ansh Bansal. I'm 18, an aspiring cricketer, and a maths student trying to make the numbers say what the game actually shows.