To know how we are performing testing on the current build, we can match the current number of found defects with the predicted data and then take decisions of - Are we in the right direction? Are we missing anything, why there is a difference? etc
Is there any estimation techniques available to do this? Any experiences on this please share.