there has been talk recently (see the set of posts starting from there) about seed cycling. more precisely, concerning shrines.
as you may already know, upon game creation certain values are set, which determine the effect of some shrines, for example hidden and enchanted ones. however, those "seed" values change and get cycled as the game progresses, in such a way to render the default, fixed start value more-or-less random. leaving the seeds uncycled gives off certain advantages, such as ensuring shield- and chain-lightning-effecting results.
concerning the procedures that could be used to render shrine values more randomized than that, not much has yet been suggested, except a more-or-less-tested game-rejoining hypothesis. what is needed here might be light, to be shed on such procedures.
there are various ways of testing those, short of fiddling with the source code. one could, for example, hack it so as to render all shrines enchanted, then test from there, as it would be much more convenient to do so that way. another way to bring it to experiment would be to get it to single-player, permitting save-loading abusing.
in short, if you have any information concerning such seed-cycling procedures, be welcomed to share back.