Monte Carlo Chess

Question

I am trying to develop a Monte Carlo Chess game, to get rid of draw death in higher ranked matches, and give humans a chance against engines. I want it to still be a strategic game, just with determinism removed.

My Current idea: The objective of the game remains the same: Capture the opponent's King.

Each piece has an initial Power Ranking

Queen(Q) = 9
Rook(R) = 5
King(K) = 4
kNight(N) = 3
Bishop(B) = 3
Pawn(P) = 1

Rules:
(1) When a piece A moves to attack another Piece B, the probability of the attack succeeding is given by: P = A/(A + B/W) *

(2)If the attack fails, the attacking piece is removed and vice versa. The piece which wins the confrontation, has its power rankings updated as such. (Supposing A Wins) A += B/X

(3) Kings are special. A King's power ranking is 4 when attacking, but when the King is being attacked, it's power ranking is reduced to 4/Y Opponents do not have to respond to checks, but will lose the game if their King is captured. **

(4) If a pawn C gets promoted to a piece D, then C's power ranking changes as such; C += D/Z

(5) You are allowed to put your king in danger

I don't think a draw will occur, unless we have 50 repetitive moves, or both players agree to draw. Because a player lacking any valid move, will not occur since players are allowed to risk their kings, and may be forced to.

What values do you propose for W, X, Y, Z?
What rules do you think should be changed?

(*) This is to favour the attacking piece, and hence propel the game forward.
(**) This is to keep in line with modern chess, since the objective remains capturing the King and not eliminating the oponent's king.

Answer 1

This type of thing is frequently suggested by people who don't understand chess and don't understand what makes chess appealing to serious players. Adding any element of probability to chess is a serious no-no. The determinism is the whole point.

Answer 2

The idea seems interesting but at the same time it differs quite a lot from what chess is; it is not entirely clear to me what the basic rules should be for this to be somehow a strategic game. In particular my concerns are mainly the following: let W and B the two players and f_i(A,B) the evaluation of the i^th move, namely the sum of all possible pieces confrontations (and all combinations thereof) between the two players, at moment i^th of the game. Moreover, since you said

The piece which wins the confrontation, has its power rankings updated as such. (Supposing A Wins) A += B/X

it follows that if f_i(A,B) > 0 then one can already calculate exactly what f_k(A,B) is for k>i (given the weight given to each piece) and I suspect that in most cases one has f_k(A,B) > f_i(A,B) for k>i. In other words if winning a piece increases the attacking weight of a figure then this may cause a chain reaction, that is you just win the first piece and then automatically win them all (since the attacking weight of your side increased spontaneously after winning the first piece).

• are the players forced to move? If so, since the probability of attacking success is known a priori (as a function of A and B) why would a player choose to play at all, if the attack is inferior by definition?
• since the attacking success is known a priori I would see this favouring too much the player who starts first: this is because they can choose to start attacking maximising f(A,B) in a given position.
• positional games will have no meaning because of the above: it is not the position that counts, rather the weight assigned to each figure,no matter what the evolution of the game can be (since the evolution is frowned upon by the fact that pieces "gain" weight after winning attacks).
• the whole idea behind chess is that figure do not have a predefined value, rather their scope and attacking power depends ultimately on their position, which is what makes chess a strategic game.
• sacrifices do not make any sense, since everything is "forced" to happen (it is because you already know what the value of a piece is after winning the attack).

Other than that it is definitely worth it looking at some sort of game like you mentioned, only I fear it will differ very much from actual chess (plus it might be completely solvable after 2-3 moves).