[from]

People underestimate the extent to which a bunch of opaque "anti-abuse" algorithms control things. Everyone is given a risk score and if you exceed an internal threshold they will never respond to your support requests until your complaint gets on the HN frontpage. Then as justification to continue their pointless cat and mouse game the abuse department types will come in and say "well if we told you why we arbitrarily decided to <steal your money/delete your 20 year old email account/prevent you from logging in with a weird error message> then the real criminals would know how we detected them!"