This is among my stranger experiments. The idea is that, rather than choosing targets like a normal human being, you bring along a stack of cards that have various conditions on them. Then, you use those to determine who you're targeting with attacks and spell/ability targets. You aren't allowed to target anyone else or their permanents, no matter how much better that would be for you.
There's a few ways you can use the criteria cards. The simple way is, on your upkeep you draw another one (revealing it is optional) and use that until your next upkeep. The complicated way is to draw each one secretly, and just keep using it until someone guesses what the criteria is. Then, the correct guesser becomes immune for a turn cycle or two, and you move on to the next criteria. I should add that no one has been willing to do the complicated way yet.
Anyway, the deck itself is pretty simple but it's a pretty effective midrange aggro. I decided that I could only use cards with a multicolor color identity (no fetches, no sol ring, etc), mostly because this idea is stupid to begin with and it takes a very long time to make a 4 color deck if I have to look through all 4 sets of monocolor boxes (I've got roughly 1K cards per color). Also I decided to use a bunch of cycles for no good reason. Don't take this as any kind of serious deck, basically, but it works alright for the intended purpose and it's aggro as hell.
And the conditions deck:
Controls most enchantments
Highest life total
Controls most tapped lands
Controls toughest creature
Has the most counters among permanents they control
Has most cards in the graveyard
Controls highest cmc permanent
Controls most creatures
Controls first permanent alphabetically
Controls most permanents
Controls most tapped creatures
Controls most lands
Controls most untapped creatures
Most recently attacked me
Controls most artifacts
Most recently targeted me or my permanents
Most recently cast an instant
Has most cards in hand
Controls last permanent alphabetically
Controls most powerful creature
So far the experiment went...decently. As I said no one was willing to do the guessing game, which I thought was sad, but it also would have been hard to keep a secret while trying to figure out something as complicated as the alphabetical ones, or to count cards in graves for example without giving that one away. As a functional semi-NPC it was still amusing, though, most people thought it was pretty funny and resulted in some surprising twists on occasion. It also seemed to draw perhaps some fire away from me, at least people were less salty when I targeted them because it was just the will of the cards. I also think I took it less seriously, which is generally probably a good thing.
As far as the deck itself, super functional and had some pretty fast kills. Got out a fairly-fast saskia + aurelia + iroas game 1, which killed people very, very fast. Game 2 I kept a 2-lander with a karoo and never saw another land. Game 3 was pretty grindy, looked like I might lose the 1v1 to a topdeck gisela, but I kolaghan's commanded my olivia and stole it, saving myself just as I was at 3 life and turning the tide quickly to my victory. The manabase is a bit annoying without fetches, command tower, etc though.
So I've played this some more, and had a few other thoughts.
Thought #1 is that this is less of a disadvantage than it seems, largely because this deck is pretty powerful compared to my local meta, despite the restriction of all-multicolor, having a pretty slow curve, and being aggro. Saskia is pretty nasty as a commander and the deck can kill people pretty quickly. That means I'm usually the threat, and it doesn't matter too much who I kill first. Plus the cards are kind of designed to pick out someone who is likely in the lead anyway, so usually it's not a terrible metric. Rarely am I attacking someone I don't want to be.
Thought #2 is that I don't love the card design. It was designed for the guessing game, which is fun in theory but no one wanted to put in that much work so we've always just done one card per turn. As-is, I usually keep attacking the same person, the person who's ahead. So I think there are 2 ways to go with it. The boring way is to just go for a random attack. The more interesting way is to reduce the number of criteria by which I attack - maybe have 6 different things I look at and pick one randomly each turn (this also means I could get rid of the cards and just roll a dice, with a guide explaining what each roll does). The semi-random aspect was a bit fun, but I want my opponent to try to play around me, and one way to do that is to reduce the complexity so that they only have to track a few things. So my opponents maybe think twice before playing that last creature, because it'll put them over the top on # of creatures which will make them the target if I roll a 4 or whatever. That gives an advantage to my opponents who are paying attention to their board state and trying not to run afoul of my potential targeting criteria. Whereas right now there are too many things to worry about, so people just don't bother.
Top 6 are probably...
-last person to attack/target me
-controls the most permanents
-controls the first permanent alphabetically
-most cards in hand
-most recently cast an instant
-controls the most tapped lands
I think those are diverse enough that people will have to manage each one individually, and some are at odds with others.