My problem is the TPG's want to point out what a tiny percentage of mistakes they make. But they use their entire population to generate that number. A much more valid statistic would be: Out of stuff that has been altered, forged or otherwise made ungradable what percentage has been graded?
If 95% of what is submitted is legit but you're getting 20% of the bogus crap wrong you can claim a 99% success rate.
|