Question - at this time - ruclips.net/video/qcLZMYPdpH4/видео.html - shouldn't the last columns be 123, 132, 213, 231, 132 and 321 ? I am not sure why Prof. Jackson mentions 123 consistently. Could someone explain
@@和平和平-c4i This is an introduction and and a gross simplification. Players might be accepted as members of a coalition (and in conditions of high social pressure often are) for the sake of their power to otherwise inflict high losses on the coalition from outside. Due to the same evolutionary rule the order of joining the coalition matters (Spencer's rule also known as the right to negative beneficence). That is, neophytes will very often (almost always?) be ostracised, leading to their low or even negative perceived value to a team, whereas in other scenarios their contribution might be highly positive. Further on, there is development potential of an individual, so some dynamical approach should follow. Result of rigid implementation of this sort of evaluation will be (and often is) leading to a gridlock. I would say - always remember your assumptions. We have no other means of gaining understanding, but by simplified models, but care should be taken in advance to expose their possible pitfalls. Last, but not least, let me wax biblical. If as a result of your model implementation somebody dies, it is you who picked up the first stone.
+Choi Hak Hey, have you found any videos or documents where I could learn how to calculate the Shapley Value when the tricky part you mentioned is a true condition? An example of this situation is a football game. It is a co-operative game and there are 11 players in a team. Say our objective is to find the teammate who contributes the most to the team. We cannot exactly form all the required coalitions though because removing even 1 player (or any combination of 2,3,4,5...,9,10 players) from the team would mean that that match will not start. So your tricky part of v(N)=1 and v(S) = 0 if N=/=S is true.
So Shapley value is simply the simple average of some marginal gain or loss! But, people do fight to have a higher share of gain, or a lower share of loss. Shapley value is not sophisticated enough for real game players.
12:32: The examples are totally unclear. How are the v(1) assigned? Where does it go from? Why 1st two lines have the same v(1) value? This is the worst explanation in the Universe, I'm frustrated.
Oh, he just gives an example for the 1st player only: in each row the contribution of the 1st player is where he appears - if it appears immediately, then we give v(1), at the 3rd row the 1st player appears after the 2nd and it's marginal contribution is v(12) - v(1) and so on.
![(AGT4E11) [Game Theory] Calculating Shapley Values: An Example](/img/1.gif)
Simple, clear and concise presentation - you are an outlier amongst academics. Thanks!
One of the best explanation out there. Really articulate professor, Thank you!
Cheers Matt. You just saved me a bunch of time figuring out my grad student's paper!
I'm not able to understand how were the weights assigned to v(1), v(12) - v(23). Can anyone please explain?
I think I got it, Thanks.
@@nature_through_my_lens I need your help about this!
@@Luckilydog999 May be you can share your LinkedIn or something.
@@nature_through_my_lens at here
@@Luckilydog999 yes.
Question - at this time - ruclips.net/video/qcLZMYPdpH4/видео.html - shouldn't the last columns be 123, 132, 213, 231, 132 and 321 ? I am not sure why Prof. Jackson mentions 123 consistently. Could someone explain
thank you!! for us students, this is of so much value!
...-shapley- value!
You saved me from my professor's poorly written slides. THANKS.
Did anyone crack up at his explanation of "dummy players"?
It simply means: if you bring no worth to the coalition, you deserve nothing .
i.e.: If your marginal contribution is 0, you receive 0 payoff.
@@和平和平-c4i crack up = laugh
@@和平和平-c4i
This is an introduction and and a gross simplification. Players might be accepted as members of a coalition (and in conditions of high social pressure often are) for the sake of their power to otherwise inflict high losses on the coalition from outside. Due to the same evolutionary rule the order of joining the coalition matters (Spencer's rule also known as the right to negative beneficence). That is, neophytes will very often (almost always?) be ostracised, leading to their low or even negative perceived value to a team, whereas in other scenarios their contribution might be highly positive. Further on, there is development potential of an individual, so some dynamical approach should follow. Result of rigid implementation of this sort of evaluation will be (and often is) leading to a gridlock.
I would say - always remember your assumptions. We have no other means of gaining understanding, but by simplified models, but care should be taken in advance to expose their possible pitfalls.
Last, but not least, let me wax biblical. If as a result of your model implementation somebody dies, it is you who picked up the first stone.
really like your video, very helpful for my study. Thank you very much for that. The examples are very illustrative, and very easy to understand.
This video should have started with the example at the end. But great video!
Thanks for the explanation!!! Very helpful!
Thank you so much~ And I have a question, when the number of players are very large ,how to deal with this situation? Thank you!
In the definition of the Shapley value, the sum is divided by |N|! and not by N!.
Thank you very much Matt. This is very useful.
Thank you for uploading this
thank u for this video, great explanation!
Thank you so much and I found I'm dummy player but I still receive your knowledge :(
Nice video, thanks so much
Great vid!
Thanks!!
You haven't covered the tricky part: v(N)=1, but v(S)=0 if N .not. = S.
+Choi Hak Hey, have you found any videos or documents where I could learn how to calculate the Shapley Value when the tricky part you mentioned is a true condition?
An example of this situation is a football game. It is a co-operative game and there are 11 players in a team. Say our objective is to find the teammate who contributes the most to the team. We cannot exactly form all the required coalitions though because removing even 1 player (or any combination of 2,3,4,5...,9,10 players) from the team would mean that that match will not start.
So your tricky part of v(N)=1 and v(S) = 0 if N=/=S is true.
Thanks, I get it!
thank you so much
amazing
Been feeling like a dummy player this pandemic
So Shapley value is simply the simple average of some marginal gain or loss! But, people do fight to have a higher share of gain, or a lower share of loss. Shapley value is not sophisticated enough for real game players.
world tune them
12:32: The examples are totally unclear. How are the v(1) assigned? Where does it go from? Why 1st two lines have the same v(1) value? This is the worst explanation in the Universe, I'm frustrated.
Oh, he just gives an example for the 1st player only: in each row the contribution of the 1st player is where he appears - if it appears immediately, then we give v(1), at the 3rd row the 1st player appears after the 2nd and it's marginal contribution is v(12) - v(1) and so on.
Very weak
did you check this one
/watch?v=w9O0fkfMkx0