they may even once say the first few blocks above 1mb havnt caused issues. increment it in 0.25mb increments. taking possibly a day or less to get to 2mb... but one thing is for sure they wont just jump to 2mb -straight off. they will see if there are unknown bugs/orphan risks etc
-snip-
Slow increments will not reveal all possible exploits. They will be revealed once someone attempts to use them, at which time damage has already started occuring.
lets say there was another Leveldb 2013 episode, but this time at 1.2mb
the pools dont know its 1.2 yet though
now would pools jump to 2mb where there might be hit by that bug AND say 5 other propagation issues aswell all in one swoop of mega F*ck-ups and then have to tear up and remake a 1mb block with half as many tx's to try again..
or wrongly guess that it must be an issue at 2mb so try 1.9mb and have that block too throwing a tantrum in the orphanage, then try 1.8 and so on all creating orphan tantrums trying to find a new 'sweetspot' that is acceptable
(to potential knitpicker: yes i know bitcoin doesnt have an actual 'orphanage' place, im just entertaining wordplay of 'orphan drama' to entertain the imaginations of people with short attention span)or logically
start in small increments. and when hitting 1.200250mb.. the block causes orphans. so they just take 1tx (250bytes) out of their list and try again. knowing to not pass 1.2mb while things get sorted.
then later when sorted they go again. safely knowing at most its only going to cause minimal disruption... rather then several orphans of going straight to 2mb and haviing to work its way down trying to find a sweet spot. without instantly knowing whats a good or bad size to run with(apart from 1mb again)