The software was tested and we were aware that a normalization process would have to occur when the swith was flipped to HF20... However we grossly underestimated how long the normalization would take and ultimately didn't realize how vastly the RC implementation would slow things like posting and voting initially..

As a former top 20 witness and a current backup witness (who is in the "secret slack" as some put it) I'd have to state that we weren't exactly meaning to bring in these issues to the network and as stated above didn't foresee the implications that slowed the network down to a crawl shortly after the HF.