The figures for the annual number of finds removed from archaeological assemblages calculated by Robbins and reported in the PAS Guide for Researchers is ~260 000. That works out at a round 5000 a week. So how many records are made by the PAS a week of those reported finds? Like right now for example, prime detecting time, though a bit coldish. Try it yourself, like for example the two-week period 31st October 2014 (Friday) to 15th November 2014 (Saturday), two weeks of record making from those 10 000 finds.... Click on the link and you will see that the number of records is 3461 (about 5063 items) reported by 433 finders (which will include non-metal-detectorists as well as a reporting few of the 7125 metal detectorists Robbins postulates are actively finding material).
In the two weeks previous to that it was 2705 records (containing 3244 items), 450 finders. The two weeks previous to that were similar 3156 records (containing 3747 items), 458 finders.
The PAS is consistently failing to make a record of more than a fraction of what is being dug up by Robbins' "7125" detectorists". Instead of the 30 000 pieces of archaeological evidence removed from the archaeological resource from the beginning of October 2014 to now, the PAS have made just 9322 records. By the figures published by the PAS themselves, more than two thirds of what is being dug up at the moment is vanishing without record and the random one third that is recorded of selectively collected material cannot be treated as any kind of valid sample on which to build any kind of archaeological inference.
I think the situation is worse than the PAS admit. I think they underestimate the number of active detectorists by quite a considerable degree (see several posts on this last week). I think we are seeing considerably fewer than one in three hoiked artefacts on the PAS database. This is officially-sanctioned knowledge theft on a massive and scandalous scale.