“Bobby Fischer once said that the problem with chess is that you get the same exact starting position over and over. These days, there are millions of games in the database already, so it’s very hard to create original play, while chess960 is really your brain against mine. After the first or second move, you’re already thinking.” -Wesley So

The pie rule is a way to balance games that have a first-player advantage. Hex uses the pie rule, and gomoku uses a more intricate setup called Swap2. In go, we use the pie rule instead of komi. Since a move on the first line is much worse than komi while a move on the second is much better than komi, we also need a more intricate setup. Let P1 be the first player and P2 the second player:

1. P1 plays the first 3 moves
2. P2 plays the next 2 moves
3. P1 plays the next 2 moves
4. P2 choose a color
5. The game begins with white playing next

Komi is 0, and we recommend Chinese rules to prevent most ties. Additionally, each stone placed in the setup must have four liberties; this prevents first-line plays and contact plays.

Note: Pie rule only works for games between evenly-ranked players.

Our tentative name for this setup is 3-2-2 pie rule. Because there is no komi, black’s first few moves must be a bit worse than usual. In each round of placements, players want to create a fair board, or punish the other player for creating an unfair board. We let each player place stones to prevent one player studying a particular opening beforehand. This setup could increase to 5-2-2 or 5-4-4, but placing 7 moves on the board before the regular game starts feels like the right minimalist approach.

---

In delayed komi, players place black and white stones like they are playing a normal game. After so many moves, they pause and use an AI to determine what the komi should be.

Deferred bidding is similar to delayed komi, although after so many moves, the two players bid for komi instead of having the AI tell them what it should be.