The Evolution of Cooperation // Iterated Prisoner's Dilemma

[@DrTrefor]

This was my favourite Game Theory episode yet, hope you all enjoyed! Thanks again to Brilliant for sponsoring, make sure to check them out ► brilliant.org/TreforBazett/

[@naman4067]

You basically gave me life lesson

[@WetwareRenderEngine]

There's an error on the slide around ~16:00. Selfish against Cooperate scores 18 over 6 rounds, not 15. This makes your point even stronger because it shows that Selfish initially beats Tit For Tat (32 to 29), but then loses against Tit For Tat (14 to 17) after Selfish drives Cooperate into extinction.

[@pst659]

this channel is pure gold.

[@DrTrefor]

Thank you!

[@fabiovargasbr]

Is more of this coming up or this is it for the series? It has been an great experience so far. Ty

[@DrTrefor]

It’s temporarily on pause, want to do a bunch more but got a bit sidelined:/

[@Darkev77]

This was such a cool concept, well put

[@DrTrefor]

Glad you enjoyed!

[@juanchetumare]

It is also interesting that cooperation maximizes the sum of both (all) players' points, which is called happiness in my head canon haha

[@AT-27182]

This is very clear and it is extremely useful for me. Thank you so much for making this available.

[@joyanbhathena7251]

Great playlist so far, hoping to see more videos!

[@DrTrefor]

Awesome, thank you!

[@abdulazizhawsah9884]

Selfish = Narcissistic people (deserving). Cooperative = Nice people (doormat). Tit for Tat = Strong character people (Fair).

[@mathflipped]

Great video, Trefor. It is amazing how such simple toy models can distill the essense of the mechanisms of the evolution of cooperation. Too bad you didn't have time to talk about the win-stay, lose-shift strategy that outperforms tit-for-tat. If I remember correctly, it was first introduced in a 1993 Nature paper by Nowak and Sigmund. BTW, Sigmund gave an amazing talk during our workshop in Turin in 2018 about modeling corruption. You could feel how his model resonated with common sense and what we see happening in reality. It actually fit perfectly my major life decisions some 25 years ago.

[@DrTrefor]

Cool! Yes I wanted to talk about the Pavlov as it is sometimes called but it didn’t make the cut. It sometimes wins, but also depends on the pool of entrants iirc.

[@mathflipped]

@@DrTrefor You are right, it depends on the pool of strategies in the tournament. I think a major reason that tit-for-tat won in the first two Axelrod tournaments was the presense of the always cooperate strategy. It also didn't let the always defect strategy take advantage of itself.

[@Lukav1]

Great video! I love your channel!

[@kuanxD]

Thanks! this video gave me a big insight for my research

[@vincentrockel1149]

From a psychological point of view, the bonnie and Clyde varient the game is a better allegory for acting cooperatively in that if neither confess, it leads to the only non-negitive outcome. Nothing negative results. If you confess you are just lessening the duration of the consequences.

[@Κωνσταντίνος-ε7ω]

Mind blowing video!

[@fabiovargasbr]

Excellent

[@interest21stcentury74]

Hey dr, I hope you are doing well, it is an epic series thank you dr very much!!! But I practiced a few exercises and there is something called the reactive functions or something like that, I just would like to ask you if anything like that will be covered and know how many episodes you have left. Since Im taking Saturdays and Sundays 2 hours for game theory. Thank You Very Much!!!

[@invinciblemaverick22]

@Dr. Trefor Bazett at 10:34, what if the prober is SELFISH on the last move....tit for tat wont get a chance to replicate....isn't that a possible outcome

[@sanjaykrish8719]

Fantastic.. this tells we shdnt be very nice or very bad to others. Gt has some really cool applications

[@DrTrefor]

It really does!

[@ugestacoolie5998]

I like to all them copycats over tit for tat, but yeah usually that strategy wins out overall, it seems the same in real life as well. Like start out treating people nicely, if they harmed you you retaliate, then if they went back to being on ok terms you go back on being ok terms with them as well. Very intuitive video! I found some extra resources as well online about this and it's really fasinating.

[@paramvir_singh]

Hey Doc Can you make a video on tensors please? Tried hard but couldn't learn from textbooks

[@JohnSmall314]

kzbin.info/www/bejne/laGuqZWErs-Bm5o is pretty good.

[@andrewharrison8436]

Key takeaway from this video: If we don't start liking these videos and commenting nicely then Dr Bazett will stop cooperating.

[@stefenleung]

you missed to mention the most successful and in reality the only solution, mafia - snitch get stitches. like in prisoner dilemma, you either deny or confess and get murdered. it's also why society evolve into policing, law/court system. If you break the contrast, you not only get nothing, you need to pay too.

[@sreekesh5854]

Nice💗

[@jjolla6391]

you forgot to mention that t4t relies on the probability of meeting again in future interactions is high. If it is low, cheating increases in value. That's why used-car sales, or real-estate sales, have a reputation: never trust them. That's bc they don't rely on repeat business.

[@lamalamalex]

It can be completely selfish to cooperate. Selfishness means concern with one’s own interest.

[@maddison0]

If you enjoyed this, I highly recommend The Selfish Gene by Richard Dawkins. A large section of the book covers this exact topic and applies it to the evolution of life.

[@DrTrefor]

I read that book years ago and had forgotten that!

[@perekman3570]

I was coming here to make exactly this recommendation! But note that the chapter on game theory was added in the second edition of the book.

[@lennykoss8777]

Can you map Doughnut Economics and the GOLDEN Rule into Game Theory? Greed Obstinance Lies Deceit Envy Nervousness Rule - things to avoid that hurt the game of life