I mentioned before that the NetLogo
models library has a number of

models related to the prisoner's dilemma.
this is one of them they're gonna look

at now

it's called PD n person iterated

and you can download it from our course
materials page. this model

unlike the one we've just been talking
about has

multiple agents cure all the different
agents

each one them play a particular strategy

for the prisoner's dilemma here the
strategies

random which means you either cooperator
defected random

each time always cooperate this one

defect always defact, tit-for-tat we
talked about

unforgiving this strategy

always cooperate until its opponent

defects at which time it will always
defect Trevor

never forgives the defection unknown
they're sure of these right now but

you can actually feel in your own
strategy

by going to the code tab and seen

the comment this is filling your
strategy here for

unknown word I can use that right now LP

part of the homework but we're going to
do is look at

we have 10 of each strategies and
we're gonna watch them

interact so I gonna go

speed them up a little bit you see
they're moving around when they come

close enough

the interact can play around with the
prisoner's dilemma

according to their strategy each time
there's a game

you see the payoffs appear next to the
agents

you can also see the average payoff for
each type the strategy here

see that defect strategy of always
defect

his the highest second is random

then tit-for-tat than in unforgiving and
cooperators

lowest wanna rock but if we speed up and
play for a while

we'll see if that actually changes
overtime

you can make it really fast and you see
that defect

actually starts to go down and

defect actually goes down and

to tit-for-tat goes above defect. Okay so
anyway

I'm going to give you some exercises to
do with the homework using this model

which will perhaps give you some insight
into why these kinds of dynamics are

happening

the conclusions that Axelrod drew from
all of these tournaments

where the following: strategy seated well

were nice they were never the first to
defect

they were forgiving that is they were
willing to cooperate if corporations

offered

they were retaliatory that is they were
willing to defect

if others defected against them and he
added one more condition to be clear

that is be transparent about what your
strategy is make it easy to infer

so that your partner knows how to react
to what you did

it can predict what your next move is
going to be it knows you're going to be

retaliatory or forgiving

and these are the conditions which at
least in the prisoner's dilemma context

for what induce reciprocal cooperation
in an iterated prisoner's dilemma game

and several people have proposed these
as general

attributes that strategies should have
in the real world for trying to induce

cooperation

in arms race like situations and of
course tit-for-tat

has all these attributes that's all I'm
gonna say about the prisoner's dilemma

it's a huge field actually with many
many variations

and I've put some references up on the
course

materials page that will give you a
feeling for

the different kinds are variations
Axelrod's book the complexity of

cooperation talks about many of them

and summer the optional homework
assignments will have you read about

some them