Sirlin is preoccupied with including "Yomi" in his games. "Yomi" is his term for rock-paper-scissors predictions. He asserts that rock-paper-scissors is flawed and uninteresting because all three options have equal value, but believes that when players must choose between rock-paper-scissors or rock-paper-scissors-lizard-spock and there is a different weighting (Maybe you get bonus points for winning with paper, etc), suddenly it becomes a good gameplay mechanic.
It can be a fun mechanic in realtime games, like fighting games he comes from. Turnbased games tend to make the mechanic far less interesting, because you have time to figure out the Nash equilibrium, and then randomly select a strategy based on that, be that with a crude RNG in your brain, or one you have on your desktop, or just a pseudorandom pattern you get into a habit of following that will take a while to catch on to. The player who is weaker at getting into the other guys head in "rock-paper-scissor" is inherently motivated to use random number generation, or random number generation emulation, to select his strategy.
Note that if you design a computer to play one of Sirlin's wonderful games, it can just use a random number generator and play the nash equilibriums straight. A pattern we find amongst game mechanics that actually make games fun and beautiful is the inability for computers to follow along. It's a mechanic computers deal with rather easily, yet Sirlin thinks it improves things.
Chess 2 starts out being a pretty reasonable faerie chess thing. He's not the first guy to do faerie chess. But then he ruins his faerie chess by cramming the yomi he is preoccupied with down its throat.
What makes Yomi even worse than being devoid of real value is that Yomi has a variance cost. When good gameplay mechanics have a variance cost it can be worth it, Swindler adds things you didn't have to think about before, the possibility of Chapel-fizzle leads to innovations like the Chapel-Baron opening for hedging bets, etc. Yomi doesn't add value, and it still has the variance cost, whichever player invokes RNG for their strategy selections invites variance to the game, which can make it harder to get consistent results. This isn't even thoughtful variance like, Treasure Map is variant, maybe i should pick a more reliable strategy. The random part happens after the Nash equilibriums are already calculated, after you decide what risk are involved in each of the choices and set up a ratio for them. It's just a pure matter of what player 2's random number generation method spits out. (Of course, if player did his Nash equilibriums wrong, then there is a skill factor there, but you get the same kind of calculations in perfect information turn based games, that's not a unique component.)
One of Sirlin's big games is called "Yomi", and it is not just a game sprinkled with Yomi, it is a game packed to the bring and chock full of yomi. In factor, the only other mechanic in the game is card counting, pretty much. It is one of the most painful games I've ever played, and it is remarkably unenjoyabe regardless of whether you win or lose. The only game of his that I enjoyed at all was Puzzle Strike, which lacked much yomi probably because it was too hard to cram it in there. Of course, he managed to screw that game up too, and then I found out about the original Dominion and how it's 10 times more suitable for a tournament environment, even if you open up a box of Base and don't change a single rule, and that is a rant involving the story of how I ended up being so active on this here forum.