THE EVOLUTION OF ZERMELO'S THEOREM

WRITTEN BY

HAFSA JAHAN
 WRITTEN BY

HAFSA JAHAN
The Zermelo's theorem for the example game of chess
was established in 1913 by the German mathematician
and logician Ernst Zermelo. It was subsequently given
his name. In every finite two-player game with perfect
knowledge in which players move alternately and
where chance has no influence on decision-making,
according to Zermelo's theorem. According to
Zermelo's theorem, a player in a winning position may
always force a win in two-player, perfect information,
zero-sum games regardless of the strategies used by
the other player.

Additionally, and as a result, if there are spots open, it

will never require a player to make more movements
to attain a winning position (with a position defined
as position of pieces as well as the player next to
move). Zermelo avoided talking about chance-based
or secret-information games. In 1913, the German
translation of Zermelo's original paper stating the
theorem, Über eine Anwendung der Mengenlehre auf
die Theorie des Schachspiels, appeared.

In 1997, Paul Walker and Ulrich Schwalbe

translated Zermelo's article into English
and included it as an appendix to Zermelo
and the Early History of Game Theory. In
chess, both White and Black can force a
win, or at the absolute least, both sides
can force a draw, according to Zermelo's
theorem, a well-known game theory
conclusion.

 We must emphasize that in

this theorem we speak
about any finite-stage, two-
player game with alternating
moves, with any action sets
and any real-valued reward
functions in order to give
the formulation of the result
the highest level of
generality.

1) In a combinatorial game, there are two players

who take turns making moves.
2) There are only so many moves that can be
made.
3) The procedures controlling changes in posts.
4) The game will only end if a certain player is
unable to go any farther.
5) The game ends right here at some point.

The outcome of the game is not necessarily

determined by these five steps.

 Two factors determine winning. As follows:

1) Normal play
2) Misère play

In a normal play, the last player to make a move

wins.
In a Misère play, the last player to make a move
loses.

The positions in the game must now be taken into

account.

 The present position is referred to as the N position if

the player who moves from it will win the combinatorial
game. On the other side, the winner will be in the P
position and will occupy the place previously held by
the person who made the move. In this case, only these
two locations, N and P, will be covered. These are
mainly used for games that have a win or loss outcome.

 One of the most important things

we need to concentrate on is the
fact that we won't be taking the
possibility of a draw into account in
this particular game that will be
played.

You can also think about choosing

to sketch if you wish to.

However, we won't take the

possibility of a draw into account
for the case that we'll be describing
following.

In the two-player strategy game from Chomp,

the rectangular grid is made up of smaller
square cells that resemble the chunks of a
chocolate bar. Chomp is a two-player perfect
information game that is impartial. We have
chosen this particular game in order to better
understand the guidelines that a combinatorial
game follows. Each player takes turns choosing a
block to "eat," or remove off the board, along
with any blocks beneath it and to its right, as
shown in the diagram we produced below.

In the two-player strategy game from Chomp, the

rectangular grid is made up of smaller square cells that
resemble the chunks of a chocolate bar. Chomp is a two-
player perfect information game that is impartial. We
have chosen this particular game in order to better
understand the guidelines that a combinatorial game
follows. Each player takes turns choosing a block to "eat,"
or remove off the board, along with any blocks beneath it
and to its right, as shown in the diagram we produced
below. Assume that the "poisoned" block in the bottom
left of this picture, the one with the "X" sign, causes the
person who eats it to lose. Let's look at the game tree
right now:

 The game tree for the next chomp game is

shown here. Let's assume that the first player
has chosen to grab the complete chocolate's
upper right block. It is the first action that our
game tree shows. You can see that the arrow
points to the middle chocolate provided in the
second stage if the first player chooses to
remove the upper right block of the whole
chocolate. There are now three distinct
chocolate forms in the second stage.

 If you look closely, you'll see that none of the three

shapes include the upper right block of the entire
chocolate bar.

The arrow would point to the chocolate that is in the

left side of step 2 if the player removes both the bottom
right block and the upper right block of the chocolate
from step 1.

The result of taking out those blocks is this. The arrow

would point to the chocolate on the right side of the
second step if the upper right and upper left blocks of
the chocolate in step one were both removed. There are
two possible arrangements for the center block in the
second step.

The arrows indicate one that is on the right side and the
other that is on the left side. Step 2's arrow would point
to the chocolate on the right side of the step if the
upper left block of the middle chocolate were removed.
Now it plainly misses the upper left block, as you can
see.

Step 2's arrow would point to the chocolate on

the step's left side if the lower right block of
the middle chocolate were to be removed. The
arrow would point to the final block of
chocolate that is "poisoned" in step 3 if the
upper block of the left chocolate is removed in
step 2. Finally, removing the right block of the
rightward chocolate in step two would direct it
to the final block of chocolate that is
"poisoned" in step three. How do we know if a
player has won a match?

What considerations do we need to make in order to

show that only one player will likely win? The
Zermelo's Theorem provides the solution. Let's
reiterate Zermelo's theorem. Zermelo's Theorem
allows for the solution of any two-player finite-stage
game with complete information and alternating
moves. The game must have two players, flawless
information be used, the board game must have a
certain number of pieces, the two players must be
able to take turns, and there cannot be any element
of chance.

 His theorem has mostly been applied to the game of

chess, despite the fact that he accepted the existence
of other games of a similar character. Typically, chess
seems to be a game with a conclusion. But is it really
so? In the game of chess, you can repeat some
moves. When you repeat particular moves, chess will
no longer be a game with a finite number of moves.
As a result, rules are frequently established to ensure
it. For instance, after three repetitions of a position, a
draw would result.

 When such constraints are

imposed, chess finally
becomes a game with finitely
many outcomes. Additionally,
the Zermelo's theorem states
that in chess, the game must
end in a draw if neither the
white player nor the black
player possesses a winning
strategy.

 PROOF OF ZERMELO’S THEOREM:

Zermelo's theorem's base is the induction.

Furthermore, it depends on how complex the game
tree is. We can determine if a player won, lost, or
drew the game with the help of the terminal nodes.
Let's take another look at the game tree now.

We must keep in mind that this is a typical game that

we are playing. Consider that blue makes the initial
decisions, and red follows with the final choice. Blue
ultimately made the choice. We can quickly and easily
show three of the winners from the diagram. It has
been assumed that Zermelo's method of proof was
backward induction. That is, either player 1 has a
method to force a win, either player 1 has a way to
force a tie, or player 2 has a way to force a win.
Induction provides the proof.

In order to solve finite extensive form and sequential

games and infer a sequence of optimal actions,
backward induction in game theory is an iterative
technique of reasoning backward in time from the
conclusion of a problem or circumstance. Forward
induction is the notion that players in a game think
their opponents would make sensible judgments in
the future, even when faced with an unexpected
event, based on the logical decisions they made in
the past.

By working backwards in time from the resolution

of a scenario or problem, reverse induction is the
process of determining the best path of action.
According to Zermelo's research, a player in a
winning position can always force a win in two-
player zero-sum games with perfect information,
regardless of the opposing player's tactics.
Furthermore, and as a result, it will never take
more moves to get into a winning position than
there are available positions.

 References:

Schwalbe, Ulrich; Walker, Paul. "Zermelo and the Early

History of Game Theory".

Lecture 3: Combinatorial Games: Zermelo’s Theorem,

IIT Bombay July 2018