A note from the desk: a tighter, stricter card

This one is worth reading because you will see the difference straight away.

Over the last fortnight we tested a set of stricter rules for which systems are allowed onto the card. Each rule was checked against a version of itself with the real information swapped for random noise of the same size. A rule only went live if the real version clearly beat the random one. It sounds fussy, but it is the only way to know a change is real and not just luck in the numbers.

Four things changed. Only a short list of the most reliable form filters can now fire a bet. Systems that hardly ever produce a selection are dropped, because a handful of lucky results is not a record. Systems are now ranked on how their particular handicap line and filter have done, not just their overall numbers. And deeper handicap lines, three quarters of a goal and more, get a small lift in the ranking.

The effect is big. The pool of systems allowed to bet has gone from around two and a half thousand to around three hundred. On the first test card ten of the twelve picks changed.

You should see fewer picks on most days and a few more of the deeper lines. Stakes stay the same.

I will be honest that the last two weeks have been poor, and this did not start because of them. It was already in testing. But the results have made me keen to get it live.

David