I didn't even
know, it was the fact that there
was a foul, a foul that they
called a club for him. We're talking about the
reffing last night? Interesting. Good to get the W. Great, so in the past
couple of classes we've started to refine
our understanding of game theory and
Nash Equilibria. So we started off
with the basic idea of one-shot games and
the different types of games. we talked about
what it is to solve the game using Nash
Equilibria and then we talked about the
situations where there might not be pure
strategy but they're a mixed strategy Nash
Equilibria that exists. Today we're going to
go on and talk about the one-shot games
when they're repeated. So a lot of interactions
in real life don't happen once they
happen multiple times and the structure of
how these happen has a large effect on on
how people plan what the outcome will be.
So I thought the best way to get us in this
headspace is to do it ourselves. So
we're going to play a game where you're going
to be randomly matched up with another
person in this class. One player will be
our road player, one player will be our
column player, same as all the games we've
played. The payoffs are going to be on the
next page and there's going to be two phases.
In the first phase you're going to play
eight rounds of the game with your partner
and after eight rounds the game ends and
then you're going to be rematched with a
new partner and in this new phase the
second phase you'll play an unknown amount of
rounds of this game with your partner so
you won't know who each of these people are
that you're matched up with but after
each round you'll see what the other person
did that's obviously needed to know your
own payoff and you'll be able to see history
of the game as well so this is what the
the payoff table is row player column
player up down left right and the
payoffs are here. You'll be able to
see them all the time on the VEconLab
screen as well. So remember the first
phase you play exactly eight rounds of this
game with your partner and in the second
phase you'll play an unknown amount of
rounds with the partner. So if you can go to
VEconLab.com and go to BLGR28 and you can
get started at your own leisure if I've
done this correctly. Yeah log into
the session. Go to participant. Yeah so BLGR28. Let's read the
instructions and make your
decisions and we'll pair you up with
some of them. Blue is the row
players play-offs, red is the column
players play-offs. So it's the same thing,
the first play -off is always the
row player and the second play-off is
the column player. Some people are flying
through it now. Stakes are pretty
low here, in fact there are zero
stakes, so it doesn't matter if you get it
wrong. It is wrong. It is a thing, in
experiments where you're partnering
with someone, they can take off.
Yeah, what's up? I'm not waiting for it. Yeah, let's have a look. Yeah, so, just waiting for
the person you're matching with to
make their decision. See? Oh, yeah. Brooke, how are we
doing? I'm waiting. How many rounds? Just about five.
Oh, so, okay. You're partnering
with someone. Okay. I'm looking at the data. and it's beautiful. I'll show you all. How are we doing? Alright. Has anyone
finished up yet? I didn't. Finished up? How did you go, Ivan? It was alright. You happy with
how you went? No. No? What went wrong? I was too kind. You were too kind. You got taken advantage of? Yeah, I got
taken advantage of. In in both parts We'll see how it goes. Hopefully people are
starting to finish up. Please keep going
at your own pace. I'm gonna start
going on a little bit with this. So, this
is the, where are we at? Is anyone
still on part one? Or is everyone on
part two at least? Great. So, this is
the game we play. What game is this? Yeah? The Prisoner's
Dilemma. It's exactly the Prisoner's Dilemma.
It's not like the Prisoner's Dilemma.
It's exactly it. Both players have a dominant
strategy. so our row player can play up
or down if they play down when the other
player plays left they get 5 instead of 3 prefer
5 and if they play right when they play
down when the other player plays right
they get 1 instead of 0 they've got a dominant
strategy to play down and player 2 has a
dominant strategy to play right instead of
left so the playoffs are symmetrical in that
regard which means what is here what's our
down right symbolise sorry i heard the right answer i'm pretty sure what can we say about this quadrant here yeah it's our national
equilibrium of our one -shot game of our one
-shot game so we talked about this in in the
one-shot environment but let's talk about
it in the repeated um environment as
well so a finitely repeated game is a game
in which you just play the game not once but
multiple times and there are variations
of repeated games in which players either
do not know where the game will end that's
what you experience for the second part
here and when you know how many rounds
there are and when the game will end and
this is common knowledge and that was the
first part here and there are different
implications to both. So let's say we play
the Prisoner's Lemma for 10 rounds and
this is finite and no. What strategies are people going
to play here? How is it going to
differ from the one -shot game? And what is
the national equilibrium here? So firstly,
Ivan, we spoke about what happened in
the in the first half of the experiment
I tried to go with the I was the left
or right yeah yeah so you try to cooperate
yeah the defection option yeah they chose
it straight away and then you just were
like screw this yeah Did anyone manage
to cooperate at all with their partner
in the first bit? Were you able to
cooperate the entire time? No, it was
the second part. Oh, in the second
part. In the first part,
was anyone able to cooperate for
a few rounds? Nobody. That's very interesting. So based on this
information, can anyone tell
me what they think the Nash
Equilibrium is here? Across ten rounds,
what would happen? Ivan, going back, what would you do
in hindsight? Going back, I would
have just chose the bottom one. Defect,
yeah, the whole time. Great. Daniel? Wouldn't the
notch delivery be one and then one, two? Just
because you're assuming that the other
person's already like picking your like dot.
Yeah. Your intuition is on point here. And
the answer is actually defect every single
round in all ten rounds. Why is this the
case? we use something called
backwards induction of the rollback argument. Let's say you played
nine rounds with this person and now it's
the tenth and final round. It doesn't
matter what you've done in the past, the
payoffs for the tenth round are now the only
thing that matters. This is essentially
a one-shot game. So you know in our
one-shot prisoner's dilemma both
players have a dominant strategy to
defect, so the Nash equilibria is both
players defect. Now we can roll back
to the ninth round, and use the same
type of logic here. So you know that if
you're in the ninth round, people's decisions in
the tenth round are already locked in. So
you can't affect future decisions based on
what you choose now. So now the ninth round
is essentially the last round where you can
make a decision, and like before, this is
treated as a one-shot game. It doesn't
matter what you've done previously, all that
matters is this payoff in the ninth round. As a
result, both players have this dominant strategy
to defect and as a result we will play the
Nash Equilibria here and you can probably
see where this logic is going we can roll
back to the eighth round the same thing will
happen all the way back to the first round
where this is the only decision that matters
because all your other decisions are locked
in once again this is a one-shot game now so
you should defect so if you know the game
is finite and you know the number of rounds
there are this is common knowledge, then the
Nash Equilibria is just affecting the prisoners
of the whole time. Okay, now what if
the game is either let's say infinite,
so it lasts forever, or the amount of
rounds are unknown to both players,
and this is common knowledge that you
don't know, this is what we experienced
here, or let's say that there's a
stochastic endpoint. There's a 10%
chance that the game ends after
each round you play and a 90% chance
it continues. So what happens
here? Is there a different
Nash Equilibria? Is it the same? How
do people's strategies change? Sorry, what
was your name up there? You're not looking
at me. You're on your phone
at the moment. What was your name?
Sorry? Justin. You were saying you
cooperated in the second part. Yeah?
So why did you manage to cooperate
in the second part but not the first?
What do you think changed, at least
in your mindset? Sorry? Well, you didn't
know how many rounds there were. Is that
the reason why you chose it or were
you just trying to feel out what the
other person would do as well and you
managed to cooperate? Yeah, I think it
would be good. Nice, nice. Did anyone else
manage to cooperate? Ivan? Yeah, I just
assumed, like, since there was less
rounds, people were going to try to, like,
the first round I played with, was going
to try to take, like, more points, but once
they chose, like, to cooperate, like,
one time, then I just could be, like,
continue to cooperate at the end. Correct. I
think we didn't know how many rounds really
left, so we were just trying to maximize
our points. Yeah. So yeah, both those
points are really good. And the idea here
is this uncertainty of if the game will
end or if the fact that it doesn't end
actually creates different Nash equilibria.
In fact, this is a very famous theorem
called the folk theorem. And the technical
definition is any feasible and individually
rational payoff can be sustained as
a sub-game perfect Nash equilibria.
we're going to define that next week, don't
worry about it for now, if players are
sufficiently patient, i.e. their discount
factor is close to one, or the way they value
their time value of money is a certain
case. But what this actually means, in
other words, is anything goes. There are infinite
different strategies that can be had in
equilibrium when we have these types of games.
And I'll show you a bit more about that in
a second, but before i dive into it i just
wanted to show our data let's see if i can
get it here great so you can see here this
is the frequency of people cooperating
treatment one the first eight rounds so you
can see that we started off with around 40 30
percent cooperation around there and it
no and it nosedive to zero in the last round
so was there anyone around like four or five
who started thinking In a couple of
rounds, my opponent's probably got a defect
because it's the last round. I should defect
now and get an advantage. Was anyone thinking
that by any chance? Because that's
usually what happens in these types of
games. There's some cooperation at the
start, but eventually it falls off a
cliff like this. Whereas you can
see, when we had an unknown amount of
rounds, people didn't know when the end
point was, cooperation was actually
increasing. so you can actually sustain
cooperation in this unknown environment and
we're going to look a little bit more
into this now so all right so this is
essentially our prisoners dilemma this is us
and this is the other player and we can
cooperate or we can cheat or such defect so the
payoffs are what they usually are so if they
cheat and we cooperate we get minus one and
they get plus three if we both cheat we
both get nothing if we both cooperate we both
get plus two and if we cheat when they cooperate
we get plus three and they get minus one
so this is just our prisoner's dilemma
both players have this dominant strategy to
defect in the one-shot game but both players
do better if they both cooperate so let's see
what happens when we play more than once
so what we're to do is we're going to play
against five different players each of them
will have a slightly different hat and each of
these players represent a slightly different
strategy so i'm going to call one person in
the class to be our surrogate and make
all the decisions for that round and then um
when the next person with the hat comes up
i'll call on someone else remember the aim is to
just get the highest payoff we possibly can
so does anyone want to volunteer up first
ivan are we cheating or cooperate in okay
so yeah and there's anywhere between three
to seven rounds of each player so we're choosing
to cooperate more time definitely cheat it's in between
three to seven and you've done three now right okay thanks
We have a new player up. Does anyone want
to have a crack? Daniel, why didn't
you have a go? You can cheat. You can try one
more time cheating. Yeah Okay, yeah What do you think this
guy's strategy is? Yeah, okay. Okay. Anyone want
to have a go? I call someone
yeah cooperate on the first cooperate
on the first one Okay, well done.
Okay. This is our penultimate one so
second last chance of all of you What
do you want to do You want to cheat
you can cheat does anyone have any idea
what that strategy was from our yellow
-hatted friend? Yeah Okay, keep that
in mind. We'll see if that guess is
correct and our final round anyone want
to have a crack Ryan, how about you? Co-operate interesting. Okay, great. So we got 27 The lowest you
can score I think is 7 and the highest
is 49 and here are our five
strategies explained Don't worry about
the detective. Let's start off
here. Our bowler hat person just always
cheats simple strategy and our pink-hatted
person always cooperates. So these
are simple strategies. The two ones of, I
think, real interest are our yellow hat,
which Ivan correctly pointed out the
strategy. They call this the grudger, but
what we call it in economics is the grim
trigger strategy. This is actually
a nice strategy. They start off by
cooperating, and they'll cooperate every
round as long as you cooperated in the previous
round. But as soon as you defect once, they
defect forever. So Tess defected in the
first round and they'll be nice, and then
it's defected forever. Finally, we have Copycat. Copycat is also known as tit-for-tat colloquially, and this is essentially
a nice strategy that starts off by
cooperating, and then literally just
copies what the other player did in
the previous round. So if you defect
in the first round when they cooperate,
then they're going to defect in the
next round, and then just mirror
what it is you do. So those are the four
main strategies. Detective is a bit
of a weird one. But essentially it
ends up playing like a copycat if you
cheat. But if you never cheat, then it'll act
like always cheat. So that's the basics
of it. But let's see which strategy actually
does best. So what we're going to do is
each player is going to play against every other
player in 10 rounds of this game. game
whichever player has the highest payoff at the
end is going to win so the the the always
cheat will play against tit for tat then
they'll play against the detective then they'll
play against the grudger and play against the
always cooperate and that will hold for
everyone so let's get your predictions in who
do you think is going to win hands up if you
think it's going to be the always cooperate
strategy hands up if you think it's going
to be copycat hands up if you think is always
cheat okay a few votes our grudger a few votes
detective detective is most people and
hands like you've got no idea didn't want to
vote and just want to see how it plays out
i'll put my hand up there okay so most
people did i think always cheat or detective let's
go always cheat okay let's look at the
first match when the copycat plays always
cheat in the first round the copycat will cooperate
always cheat will defect so always cheat
gets three points copycat gets minus one
but then the copycat will always copy what
the other played in the previous round so both
just cheat forever and get nothing so the
final scores for the first for that match
is minus one and three then copycat versus
always cooperate i mean yeah they just both
cooperate forever always cooperate always
cooperates and the copycat just copies what they
do and this goes on so when the copycat and
the grudger play against each other remember
the grudger is cooperating the first
round and keep cooperating unless the other
person defects and then defect forever but
this just leads to both cooperating and we can
keep going with this and the important one i
want to show you is this one here always cheat
versus grudger has the same situation as
the copycat versus the always cheat so the
always cheat gains three in the first round
but then the grudger never cooperates again
and as a result they only get a plus three
instead of a plus 20 if they'd always cooperated
so as we keep going through this i'll rush
through the rest so the winner here is the
copycat the tip for tap strategy and
this is um uh really interesting and it goes
back to like the early 1970s when robert
axelrod ran a prisoner's dilemma tournament and
he had academics from all over the world in
different disciplines economics mathematics,
engineering, submit strategies. And a lot
of these strategies were extremely complex.
They relied on what players did like seven
rounds ago, et cetera, in deciding what to
do next. And the idea here was whichever
strategy had the highest points at the end of
the game, when all strategies played
against each other, would win. And this is a
strategy that won. Copycat, tit for tat. A very
simple strategy. And the reason it won
is twofold. the first thing is it can cooperate
with fellow nice strategies so remember
when it played against both the grudger and
the always cooperate they'll able to score
20 points in both those matches they'll
able to take advantage of that however they can
punish nasty strategies so nasty strategies
can't take full advantage of them so
when they played against the always cheat yeah
the always got three points in the first round
but then the copycat strategy made sure
they never got a positive payoff again so
they're able to diminish the payoffs of the
nasty strategies but also take full advantage
of the nice strategies and that's why it
was able to do well so we're going to go
a little bit beyond this course right now
because i think it's really interesting
and look into a field called evolutionary game
theory um and richard dawkins actually wrote
a book in 1990 called the selfish gene and
this evolutionary game theory or
evolutionary biology comes from this it's like
which strategies win out in evolution so the
way this works is whichever strategy has
the best fitness or does the best for humans
are the ones that will win out in the end
and that depends on what other strategies
are in society so the way this game is going
to work is we're going to have some population
of players in society they're all
going to play this game against each other
then the lowest five scoring players are going
to die out from society the survival of the
fittest and they're going to be replaced
by the same strategies of the five highest
scoring players this is one way to model
the way this society changes over time and
we're going to see which winners um are there
in the long run so with evolutionary um
Game theory is if we play this game infinitely
which strategy is going to be the last
one standing kind of Okay, so let's see
this in action. So here we have a population
of 20 out of 25 people 15 always cooperates
five always cheats and five Copycats or tit for
taps. Does anyone want to take a
guess at which one's gonna end out winning
in society? Ivan? Always cheats.
Always cheat? Any reason for that?
I think if you look at their
individual game, like, they can
literally never lose. They can
only tie or win. Yeah. So I think
in the long run, like, everyone will
just go to always cheating. That's
good intuition. Did you sneak the
same thing? Does everyone here think
always cheat? Does anyone have any
other perspectives? Okay, great. Let's
see how this plays out. So remember,
each player is playing against the
other 24 players. So we're going
to always cheat. So this is the
score everyone gets. So 465 for the
always cheaters. 375 for tit for tat. 330 for always
cooperate. So let's get rid of the bottom 5.
we got rid of five always cooperates from
society and reproduced the top five so now
we have 10 always cheats same thing
again and we can get rid of five of the
lowest it's looking good for you at the
moment ivan let's play the tournament again
now you can see 30 so let's get rid
of these and let's replace them with five
more always cheats so you're almost there
ivan you've nearly one 102 for our tip
for tats and 21 for our always cheats
so let's get rid of the bottom five
reproduce top five let's play again same result
let's get rid of our always cheats
reproduce our top five um you can see where
this is going so the key thing you
forgot to take into account Ivan is, and
everyone I guess, is you're right, each
individual game the cheater always does best, but
when they play each other as well they
get zero, so they're missing out on certain
welfare each time, whereas when a tit for
tat player plays against each other, they score
20 points, and when they play against a
teacher, they score minus one, but the cheater
only gets three so that tit for tat overall
can take advantage of other nice people
in society and by take advantage I mean make
sure they're cooperated with each other and
get the full benefits but they can ensure
that the cheaters aren't taking full
advantage of them so once the always
cooperates are done out of society that's when
it changes and the tit for tat players
have this evolutionary stable strategy as
you can see here and we can do the same
thing as we add in our um grim trigger
strategy players so you can see the same
evolution happening um yeah so sometimes
you have some grudges that still exist as
well um because there's no difference between
grudges and copycats when they're the only
two left in society okay so there's a
problem look around the world is full of
total jerk wads if copycat is a strategy
in this repeated game of trust that's so
powerful and what happened in world war one
was that um uh the two armies in the
trenches i believe at um uh christmas decided
to stop fighting just gather together
and this is like a situation of complete
mistrust but they're still able to cooperate
in that sense so given tip for tat
is so strong what's actually causing so much
untrust so we looked at what happens when
they play ten matches against each other
but let's see what happens when they only
play three against each other do we get
the same result no in this case we see
always cheat wins out and this is the thing
in the first round always cheat always
does better than tip attack it's just that
in the long run tip attack can take advantage
of the fact that they're cooperating
more with the nice strategies and don't get
punished that much by the nasty strategies.
But if there are only three rounds, then
they can only score six points with a
fellow tick for tatter or a fellow cooperator.
Whereas if they're playing against the
always cheat, the always cheat will get
three and they get minus one. So the less
rounds there are, the more reward there is
for nasty strategies. Also, we can change
the payoffs to make cheating more
profitable so if it's more profitable to
cheat we see the same outcome as well so
depending on the dynamics of the situation
if there are less you know um
interactions or there's a larger payoff for
cheating we can see that there's going
to be more cheaters one last thing i want
to show you here is um there's a reason why
game theory has gone on for for like you
know 80 years now this is just a simple setup
so we have our two tip for taps here
they're both nice players their strategies
are to both always cooperate with each other
and this makes complete sense if you're
able to execute your intentional strategy
100% of the time but sometimes there are
errors for example you might want to rock up
to class on time so you don't disturb your
fellow classmates than me then you could get
stuck in traffic you could be halfway to
class and realize you forgot your laptop at
home people make errors so you rock up late
even though that wasn't your intention
and what happens when errors occur so these
are both nice players they both want to
cooperate they're tit for tat
so now they're going to copy each
other's action so now it looks
like this player cheated because they
tripped on the way to putting the coin
in and now does anyone know what's
going to happen if these two are
tip for tat players yeah yeah exactly exactly that they can
never cooperate again so how do you deal
with mistakes and the last thing I'll say is
there are strategies like copy kitten
which is the same as copycat but they'll
only switch to defection if someone defects
twice in a row against them so they
take into account that some people make
mistakes and they're more lenient as a strategy
yes that opens them up to being more exploited
by always cheaters but this is what
happens when we add more complexity into
the prisoners level there's a bit more
here but i'm not going to go through that i
want to get back to actually what we're
talking about which is looking at these
infinitely repeated games so an infinitely
repeated game is a game that is played over
and over again forever and the players receive
payoffs during each play of the game so
yes people don't live forever this is a
mathematical modelling decision arguably firms
could live forever with different people
operating them etc so because this goes
forever and you receive the payoff of the
first round today but the payoff of the
billionth round a long time in the future then
the same payoffs but they're going to be
worth different amounts to you because the
fact some of it occurs now and some of it
occurs in the future. So the time
value of money or how much you
discount the future is going to
matter a lot here. So just as a reminder,
we can put all of this into one value
to compare against each other using
net present value. So this is
essentially the future value of when you
get each payoff divided by one plus
whatever the interest rate is to the
power of that time period where you
get that payoff. Then minus your costs. so this is a prisoner's
dilemma with two firms and this is called
supporting collusion with trigger strategies
and something to note cooperation is a
good thing but when firms cooperate this is
collusion this is you know when there are bad
results for consumers and this is what we're
going to talk about actually in the in
the next topic as well so you can see they can
both set a low price or a high price if
both firms at a high price they both get an
economic profit of ten dollars if they both
set a low price they both don't get any
economic profit and if one undercuts the other
one so if one sets a higher price and you
undercut them with a low price you get 50
and they get minus 40. So this is an indefinitely
an infinite repeated game and we know in
the one shot game zero zero will be the
Nash Equilibria because they both have a
dominant strategy to set the low price this is
our prisoner's dilemma. However when this game
is repeated infinitely it's possible for
firms to cooperate without fear of being
cheated on using trigger strategies our
grim trigger so the idea is you tell the
other firm I'll cooperate as long as you do
but as soon as you defect I'll defect
forever as well so the idea here is depending
on your discount factor or the interest rate
there are situations where you can support
this cooperation. So let's look at the
general case here. So the interest rate
is I, the payoff you get from being cooperative
is PI Co-op, the payoff you get from
cheating is PI Cheat, and the one-shot Nash
Equilibria payoff is PI Nash. And this formula
here looks a bit complicated, I want to
make it a little bit easier so what this is here
is the gain you get from cheating in the
first round instead of cooperating so that
number here is going to be 50 minus 10 which
is 40 that's what you get if you defect on
the other person the first round and this is
divided by what you'd get if you cooperated
in every round instead of what you'd get in
the Nash Equilibria. So 10 minus 0, you
get 10 every round. So let's take this
to the other side. So now we have the
profit of cheating minus the profit of
cooperating has to be less than or equal
to the profit of cooperating minus
the profit of Nash divided by the interest
rate. And that's because the higher
the interest rate is, the more you could do
with the extra money you had in the first
round. With this $40 you have in the first
round, when you defect, you can essentially
put that in the bank and earn interest
on it. So the higher the interest is, the
less likely you'll be to cooperate, because the
benefit of cooperation happens in the future
rounds, whereas the benefit from
cheating happens today. So we can look at a
couple of examples to help us out here. so
let's say the interest rate is 40 percent
40 percent what is going to be the situation
here and the firm's going to collude or
defect against each other so we can
calculate the net present value of um uh of
cooperation both playing a higher price
forever so you get ten dollars in the first
round um in round zero you get ten dollars
in the first round in the second and the
third so on forever and when you have this
infinite sequence it just equals 10 times
or the payoff I should say times one plus the
interest rate divided by the interest rate
which equals $35 and the net present
value of defecting today is you get $50 in
round zero in the first round so since $50
is more than $35 then collusion is not
possible both firms get a higher payoff by ignoring
this um grim trigger strategy and just
defecting straight away however with a lower
interest rate collusion is possible so for
example with an interest rate of 20
percent we can see the net present value of always
cooperating is 10 times 1 plus the interest
rate divided by the interest rate which
equals 60 and now this is higher than the
50 from defecting in the first round so you
can actually sustain cooperation forever
based on this both firms know it's not in their
interest to defend because they'll make
themselves worse off. So we're speaking purely
from self-interest here that you can
sustain cooperation when you have a lower
interest rate. Because with a low interest
rate, when you put money in the bank,
it's going to be worth less in the future
than a high interest rate. So there will
be some cut-off point when you're indifferent
between the two. Finally, games with an uncertain
final period. We sometimes say that
the game is finite, but there's some
probability that the game continues, and
probability theta that the game ends between
0 and 1. So you play around, and if
there's a 10% chance that the game ends,
you spin the wheel, it lands on the
90, you continue, if it lands on
the 10, you end, and you do it
again and again. And an uncertain final
period mirrors the exact same analysis
as an infinitely repeated game. You can
use the same trigger strategies. So if you
cheat in the first round you'll get 50
but using the trigger strategy then you get
zero forever whereas if you cooperate
forever your payoff is going to be what you
get from cooperation divided by the
probability of the game ending and the probability
of the game ending is is as we you know
said theta so if theta was let's say
a 10 chance that the game ends then this
would be 100 which is greater than than 50
and you're able to sustain cooperation
but if Theta was larger like let's say it was
50% then this would be 20 which is lower
than 50 and then you wouldn't be able to
sustain cooperation. So what we've talked
about today is when a gain is finite and
the rounds are known we're going to have
this backward induction this rollback equilibrium
and there's no way for self-interested
players to sustain cooperation but as soon
as you introduce this uncertainty, whether
it's the infinite game, the stochastic game,
that's what this is called, or just an
unknown final period, it allows for any sort
of strategy to take place, including
strategies that cooperate. We also looked at
a little bit of the dynamics and
how complex this can actually get. This
is touching the surface of the
prisoner's dilemma. And finally, I just
want to say that all this is assuming that
everyone only cares about their own payoff.
There are certain preferences people may
hold that they care about the payoffs of
others as well and that changes the
strategies that people play. We can model
this, this is what we do in behavioral economics,
this is what we do in behavioral game
theory, way beyond this course but I
don't want you walking out of here thinking
everyone's selfish. As you can see we can
sustain cooperation even if everyone is
selfish but once you account for social
preferences, some people are nice, it can
change the composition of types of players
as well in the future. That's the layup for
you. If you've got any questions, you
can come ask me after. Otherwise, enjoy the
weekend, and hopefully we kick our Arizona's
ass on Saturday. I have a question.
I just wanted to tell you a
story I thought you might find
interesting. Of course. So, my
family's from Croatia. I go every summer
to work there. My grandmother has a
distillery. Yeah. Where in Croatia? It's
called Rovin. It's in the East Ria region.
it's like right by the sea it's really
beautiful but uh what I think it's a problem oh
yeah so that's that's more so that's a yeah
it's very tourist yeah yeah but um whenever
whenever I work so she has a distillery she
has like she makes a liqueur and spirits and
sells all over Croatia so when I work I do
I do a lot of like the selling and the
sales merchandising so their markets work like
very different compared to ours here yeah so
they're like when you bring the products
you have to restock it yourself oh really like
on the on the shelves like for most of the
stores some of the bigger ones like they
have workers to do it for you but since it's more
like a smaller like mom and pop yeah yeah
so you got to go to each store yeah and
restock it yourself and pricing yourself or so
that we have assigned prices yeah depending
on the regions and basically like how the
product selling in that certain spot but um
you like put in the own like every single
store has like a plastic thing and you like
print your own price tags and put it yeah
but uh it reminds me a lot of like the prisoner
dilemma especially like when I was working
this summer because there's a lot you know
you're competing against all these other like
yeah brands yeah and there um you can like
there's some like respectful brands that
like you cooperate like you don't touch their
stuff they don't touch yours but um then
there's also others who are like some like kind
of nasty uh companies they're like a lot
of these stores they go to they'll like hide
our bottles oh really our stuff oh interesting
a lot because our so do you like cooperate
with the other like companies that are
nice to try and like well yeah so like the
way i like because i always talked about it
with my dad because it's very different from
here yeah um like uh a lot of the companies
that you cooperate and you know like because
it's always the same people yeah and um
like you obviously you never touch their stuff
or you know they never touch ours and even
like because we have like uh our products
like like a lot cheaper compared to some of
the other ones and like some of these companies
will like just take our price tags and
hide them so they don't have any clue how much
that's crazy I know but um yeah he always
says like you can't like don't don't touch
other people's stuff because like then it's
just gonna start a war yeah like I'm just like
we can do more yeah today these are hated
scenarios if you've got a mix of play the
different intentions he sees like um play nice
with other people but if someone's like you
ask if you like kind of gang up on them
exactly yeah so yes I thought you would think
oh yeah it's brilliant you should know everyone
that it's a great example of how this
stuff works yeah that's awesome yeah of course
yeah but sometimes like they'll go to
some crazy extents like like completely just
hide our but like the whole joke they like
make our job harder and their job harder just
to like mess with people so yeah it's very
short-sighted yeah but it's it's for it's for
all the products my uh my friend works um it's
called hell it's an energy drink brand yeah
so he does the same job merchandising and
selling but for the this energy company and
he said it's the same thing same thing like
they have like they compete against like
Red Bull and monster or like kind of the
bigger ones there and like they'll like take like
the spacing on the shelving and they'll
like take the additional boxes and go back
and move it into the storage just making their
job harder and making you know it's crazy
some of the stuff it's more specifically
one because there's a there's like four main
supermarkets there's like plodine, konzum,
kaufland and um lidl and like lidl uh plodine
those are croatian or lidl lidl and kaufland
are german and um so you can't i i assume
you can't get away with yeah they they like
that's how it should be they restock it
for you so you never have to worry with
these problems you keep a super easy job but
consome is like my dad calls it the wild
west because it's like