Actually, the more I think about it, the more I favor making the split fair and optimizing it as two separate steps. That is, first divide the treasure in a way that’s definitely fair, and then follow with both sides buying, selling, or trading specific items. Another point in favor of this is that at least one side, and probably both, already has other assets that weren’t part of the “even split” arrangement, and which we could profitably trade.
Though, yet another factor, here: The Guild probably has a fair bit of practice at making transactions lacking in mutual trust, and so likely already has some standard procedure for such things. If their proposal is close to reasonable, it might be worthwhile to agree to it, just for the sake of future relations.
I would favor the auction approach, but I’d see it this way: Heroes and Guild each have a 50% share in every item. Since gems are interchangeable and equally valued, they are simply paid out to each side 50/50 at the beginning.
Once the auction begins, you can offer a combination of gems and/or your share of magic items to bid. A simple bid might settle on 500 gems for a magic scroll. A more complicated bid might be a 50% interest in a magic wand plus 500 gems for a 50% interest in a magic sword and magic shield. By exchanging shares in the magic items directly, it doesn’t really matter whether enough gems are available to equal the magic items in value.
I’d split the jewels down the middle. Then both groups can bid for first pick at the magic loot, paying the other for the right to take one magic item. Then the other group gets the second pick, and so forth. This allows both groups to maximise the perceived worth of their magic crap relative to their idiosyncrasies & needs.
If, afterwards, the Guild wants to sell some magic doodad back to the Heroes for gems (or vice versa), they can negociate.
As Nava says, the real problem is when you have a huge pile of loot that nobody really cares about… and one (1) Arkenstone on top.
The lack of information probably works in our favor, here. If there is an Arkenstone in the treasure, then it’s likely that at most one side knows about it, so the other will peacefully agree to letting them have it. Which, yes, might mean that we miss out on that item, but that’s probably better than going to war over it.
I’m going to assume there are more than just a couple and a not altogether huge number of discrete items whose utility only depends on who is valuing it, not on whether one has certain other magic items, and that there is a fungible currency with which to deal with. In this case, I would propose that a first picker would be chosen by a method I will prescribe later, and the first picker will take one item, then each side will alternate taking two items until they are all distributed. The method of determining the first picker would be as follows:
Each team writes down their choice of first or second pick (going first is not always best: if there are three equally valued items and a bunch of worthless trinkets, going second is optimal). They reveal, and if they are different, then no problem. If they are the same, then they have a live auction in the aforementioned fungible currency for the right to decide who will pick first, with the one side bidding higher paying the other that amount.
If there are a huge number of items such that the above is not feasible to be completed, then there will likely be some way of aggregating large numbers of items together in a logically coherent way that can be mutually agreed upon beforehand. Say, if there are 9 million different sapphires of varying qualities, they can be divided into 9 lots of a million (or whatever) random sapphires, with the ability of each side to inspect the lot as they see fit before choosing it as one of their items.
The scenario somewhat precludes there being only 2 or 3 items, but I think that it works in those cases equally well.
If the two sides do not have a usable fungible currency but can agree on the values of the gems, then the gems can first be divided evenly (with perhaps a note of the amount by which the division is unequal that will applied to the bid) and used as currency with the values that each places on them. If this does not lead to enough fungibility to allow flexible enough bidding amounts, then my plan will not work.
Oh, right, I should have bumped this thread to say how it turned out. We first agreed to collaboratively identify all of the items, so everyone had (mostly) full information (there were some details the Identify spell didn’t reveal). Then, a third party managed to convince all of us that a certain portion of the treasure should go elsewhere. Then we flipped a coin and took turns selecting one item at a time, and at the end we were able to negotiate one exchange. This may or may not have been completely fair, but it had the virtue of simplicity, and I’m apparently the only one present who was geeking out about this sort of thing.
In the end, we ended up with none of the jewelry, but all but one of the magic items we were interested in, plus a few magic items that we’re not really sure we have use for, and one that we thought was good but turned out to have unfortunate drawbacks.
Oh, and a Wand of Wonder, which I’m not really sure if it’s something we actually wanted or have use for, but it sure is fun.