Transcript
Okay. Hello. Welcome to my presentation
about game theory basics and winning
strategies in life and business. And the
goal here really is to make you unlock
smarter decisions to outsmart your
friends in video games or ace in group
projects.
So let's move to a short introduction
about me. I'm Fri Kamisha currently in
Germany. I um attended a year at the
university programs in economics at the
University of Rosto. I'm currently in a
German high school with advanced courses
in mathematics, physics, and computer
science.
I did some internships at EY, PWC, and
Alian.
So, but let's talk about why game theory
matters. Game theory shows up how
teaming up often beats going solo. For
example, in video games, you often see
that when you team up with others, you
have a way higher chance of winning than
just playing alone. And this gives you
the tools to guess. Game theory gives
you the tools to guess what other
people's will do and to plan your moves
accordingly.
The key idea is that your choice um
affects others and their choices affect
you. Like in a group chat, deciding
where to hang out.
Fun fact, game theory was invented in
the 1940s by math genius Fenoyman and
Morgan Stern and is now used in video
games, apps, and even Netflix
recommendations.
Okay, so let's talk about what game
theory is. Game theory is a smart way to
study decisions where everyone's choice
affects the outcome like a massive
multiplayer game. The key parts here are
the players, the strategies, and the
payoffs. The players are you, your
friends, but also can be teams or
companies. The strategies are your
moves, for example, like passing or
shooting in soccer. And payoffs is for
example, what you get, the points, the
grades in school, for example, or money
which companies get.
So in game theory, you always have to
remember that decisions are not perfect
because we have to deal with incomplete
information with stress or biases. For
example, choosing how much to study for
a test while guessing if your classmates
will cream or slack. So let's talk about
time types of games first and then we
can move into more specific details. So
there are multiple types of games. For
example, the simultaneous move game
where everyone acts at once and there is
no peing. For example, rock paper
scissors would be such a game. Then we
have the sequential move game where you
take turns and observe the previous
moves. For example, in Uno or Monopoly.
Then we have the zero zum game. So where
one's gains, the other one's losses. For
example, in debate or sports because
there can only be one winner.
Then we have zero and nonzero sum games
where everyone can win or lose together.
For example, in a group project where
the whole group success is really and
where everyone can win or everyone can
lose.
Then then we have perfect information
games where all moves are visible. For
example, chess because you know which
moves the other player has done in the
past. And then we have imperfect
information where you don't know all the
information from the other players. For
example, in poker where you don't know
the enemy's cards.
And of the biggest contracts are between
zero sum versus non-zero sum games and
simultaneous and sequential games. And
to have the right decision to make the
right decision, you need to know which
game you are playing in.
In real life, most often it's you're
play you're dealing with imperfect
information
because we just don't know everything
which goes on. So, let's talk about
dominant strategies first. A dominant
strategy is a move that is your best
response regardless of what your
opponent chooses. For example, in life,
if you have a dominant strategy, they're
usually rare, but if you have one, you
should always use it. However, if your
move is 100% predictable, smart opponent
will exploit you and will exploit your
move. And therefore, you should
randomize your moves a bit so no one can
really read you.
So, let's talk about another term really
important for game theory. It's a Nash
equilibrium.
For example, um it's a state where no
player can improve the outcome by
changing the strategy alone. An example
for this would be the taxing back
standoff. So you don't want to tax
first. The other person also doesn't
want to tax first and so you both wait
for hours and neither of you wants to
change. And that's a Nash equilibrium.
So let's talk about a specific example
for game theory. A specific example or
the most common example is the prisoners
dilemma. And the core problem with the
prisoners dilemma is that two players
choose can choose to cooperate or to
defect.
Um and each player is a temptation to
betray for personal gain. But if both do
everyone loses
and the lesson therefore is selfishness
can sometimes backfire. For example,
when both players here are selfish and
defect, then they're worse off than if
they both would have cooperated.
And the problem really shows with the
price of anarchy. Basically, it's the
gap between how well a group could do if
they co coordinated well so poorly they
do when everyone is selfish.
You can calculate it really easy by
dividing the social cost of the selfish
behavior by the social cost of the
optimal coordination.
For example, if everyone rushes at the
same door at lunch hour, everyone gets
stuck. If everyone walked in a single
line completely coordinated, the whole
group would eat 5 minutes faster. And
the real world impacts is for example in
traffic. And that's why we have traffic
lights because without them the price of
anarchy would be a constant gridlocks
and many accidents.
So let's talk about social capital and
repeated games. How and that's how we
can avoid selfish behavior. It's about
the concept that in one time games
betrayal often pays out. As we saw in
the prisoners dilemma, if only one
person defects, it's really positive for
them.
However, in repeated games, your
reputation is your currency.
And so, if you betray one person once,
they might never cooperate with you
again. And that's a permanent game over.
As you see in many companies, it's not
just about making one deal, one good
decision, but about building trust and
building a history that makes other
people want to team up with you. And so
the key lesson is being nice is a
dominant strategy if you plan on playing
for a long time. If it's only one game,
often
um betraying others can pay out.
However, in real life as it's a long
game and if you because you play
repeated games
being nice it's the dominant strategy.
So let's talk about other things how you
can make good decisions and the main
point about game theory is always about
making the rational decision.
So rationality means choosing the path
that leads you to the best possible
future outcome. However,
there's often a trap with this and
that's why people do not always do the
rational thing and that trap it's called
the sunk cost fallacy. So throwing good
energy after bad energy just because
you've already spent the time or money.
An example for this would be watching a
boring 3-hour movie because you paid for
the ticket even if it would be way
better to just get out and leave.
Or another example would be staying in a
club you dislike because you already
bought the hoodie but don't really like
it. You could do something way better
somewhere else.
And so the rule for life, what everyone
should remember is forget the past. It's
a s cost. You can't change it anymore
and only focus on the future payoff
which can occur.
So let's talk about some other options
for game theory. How to make the games
more fair and how to force or how to um
increase cooperation.
An example
um would be reverse game theory. So
instead of playing the game, you just
build the game that everyone is forced
to play thing. This is done in video
games where matchmaking algorithms
prevent people's um from smurfing. So
pros playing against worse players or in
a school where teachers using random a
random app for teams so nobody feels
left out or the teams are completely
fair.
And the key insight about this that is
that if the game is broken, you don't
just play the game or you don't just
play better, but you change the rules so
so that the game becomes fair for
everyone.
An example of such a game would be 4K
cutting. What you want is you have want
to have a fair division division and so
the minmax principle applies here. You
want a game that um where you have to
act to minimize your maximum possible
loss.
And the algorithm for this in this
simple example with cake cutting would
be one person cuts the cake
or for example divides the jaws or does
something else and the other person
chooses their piece first. So the cutter
is forced to be 100% fair because if
they make one side better, the other
chooser will take it.
Another example for making games fair is
in auctions and especially the victory
auction. So it's used by eBay and Google
ads and the winner is the highest
bidder, but they only pay the price of
the second highest bid. And why is that
so important? It's because it removes
the game of trying to guess the others
bit. It encourages you to bid your true
value
and it avoids the winner's curse that
you just overpay to win. And actually
the person who invented it um William
Vickery won in 1969 um 1996 the Nobel
Prize for improving efficiency in
auctions.
Okay, let's talk about one of the most
important strategies
which there are in game theory. This is
a really simple strategy but it won many
tournaments
and it's a tit for t strategy. So it's
basically you start being nice. You
start by cooperating all the time.
However, if someone betrays you, you
retaliate immediately and you copy their
move. Basically, they betray you. So,
you betray them as well. However, then
you are forgiving it again. If they
cooperate again, you go back to being
nice.
And why this wins it is because it's
simple and it's transparent and prevents
others from exploiting you. For example,
if someone ghosts you on your DMs, pull
back. However, if they reply, be cool
again.
So let's talk about some real world
applications.
And basically in summary, life is a
series of interconnected games.
Whether it's in physics or everyone
else, you want need to understand the
players, the payoffs, and the rules.
Especially if you work somewhere and if
you study something, you always need to
think about the broader picture.
And the takeaway is don't just play the
game, but analyze it and design your own
incentives and always look for the
win-win because most of the time
cooperation is the best thing for
everyone.
So, let's um talk about the conclusion.
So, the conclusion is really cooperate
long-term and aligned rewards, designs,
and smarter rules. Thank you for