This I never understood ever since Dogecoin forked in mid-February 2014: If the cause of the fork was, apparently, the same as the one that affected Bitcoin in March 2013 (namely: a very long block being accepted by a new wallet using a different database but rejected by the earlier wallets) then...
How come there were up to three forks of he blockchain at a time? I would've expect at most 2 competing chains, one kept forth by the miners using the old version (then, 1.4) and another by those already on the new one (1.5, then). But, how was it possible to have three forks?
And finally (rant here), what was the point of increasing the size of the potential block size? In retrospect it's obvious something like this was going to happen. Why not treat the rules of a coin as set in stone once it's released, non-critical bugs and all? I had to delete the blockchain and resync twice, each time getting into a different, invalid forked chain - at the end I simply downloaded the files a volunteer shibe uploaded, with the potential risks that has - and hated the experience.