Messaggioda fields » 11/07/2007, 15:17

@wedge

Testo nascosto, fai click qui per vederlo
@wedge

I think that your strategy would work if agreements were allowed between players. But if the rational agents cannot interact, this is a single player game. IMHO :wink:
[i]La Realtà non si capisce, alla Realtà ci si abitua[/i]
fields
Senior Member
Senior Member
 
Messaggio: 767 di 1717
Iscritto il: 20/07/2006, 15:32
Località: Wien

Messaggioda Fioravante Patrone » 14/07/2007, 22:43

Very interesting answers.
I think that there is enough to debate for all of the summer!

I will start answering first the choices to the first round. Then I will move to comments, and eventually to the "second round".

Just to fix some coordinates, if we assume that preferences of the players:
- depend only on the money that they get
- they like more money than less,
then we are facing a "free riding" problem, that arises typically when contributions to some public project are considered.
Our "camallo" (alias cheguevilla) has correctly detected the category of problems to which the example belongs.

To make more "concrete" the problem we have discussed, you can think that you are asked to make a contribution for, e.g., building a bridge to cross a river that divides your city into two parts.
The bridge is a typical "public good", with the characteristic that its "consumption" (better, the "consumption" of the services that it gives) is neither rival nor excludable. Is not rival since the fact that you use the bridge does not prevent other people from using it. It is not excludable, since you will have the right to use the bridge, independently on the fat that you contributed to it or not.
[As the example of the bridge shows, the rivarly/non rivaly and excludability/non excludability are rarely 0/1 variables. Congestion could bring rivarly, while you could use a toll system for the bridge, anyway].

As a last word (for today) I subscribe also the connection that cheguevilla has made with the prisoner's dilemma.
Avatar utente
Fioravante Patrone
Cannot live without
Cannot live without
 
Messaggio: 1539 di 10811
Iscritto il: 09/06/2006, 19:18
Località: Temporaneamente a Novi Ligure ;-)

Messaggioda Admin » 15/07/2007, 07:23

Maybe you have to open a new topic
"Dai diamanti non nasce niente, dal letame nascono i fiori" F. De Andrè
Avatar utente
Admin
Amministratore
Amministratore
 
Messaggio: 1057 di 3581
Iscritto il: 25/07/2001, 00:00

Messaggioda Fioravante Patrone » 15/07/2007, 18:47

Let's move to the formal analysis of the game.


First of all, I will describe the game form (see: https://www.matematicamente.it/f/viewtopic.php?t=20428, sorry for the Italian...)


We have six players. The strategy space for each one is ${0,1,2,\ldots,20}$.
For sake of convenience, I will denote by $X_i$ the strategy space of player $i$. Of course, all of the $X_i$ coincide with ${0,1,2,\ldots,20}$.
As $E$ we can take $RR^6$, even if we could make more parsimonious choices.


The function $h$ is defined as follows. Since $h$ take values in $RR^6$, it will be convenient to describe it as $(h_1,\ldots,h_6)$, where $h_i$ are real-valued functions.
We have:
$h_i(x_1,\ldots,x_6) = 20 - x_i + \frac{1}{2} \cdot \sum_{j=1}^6 x_j$.


Notice that the game form is completely independent of any assumption about the preferences of the players. So, in particular, it is independent of what I said in my previous post about the "meaning" of the game we are discussing.


In the next post I will discuss the preferences of the players, so that we shall be able to say something more interesting, perhaps.
Avatar utente
Fioravante Patrone
Cannot live without
Cannot live without
 
Messaggio: 1545 di 10811
Iscritto il: 09/06/2006, 19:18
Località: Temporaneamente a Novi Ligure ;-)

Messaggioda Fioravante Patrone » 16/07/2007, 16:53

In the last post I have described the game form.

Of course, I should say "a game form", since it is my choice as a modeler to identify what I consider the correct mathematical objects to use.
However, in this case it can be considered that the game form is more or less(*) "obliged", apart from obvious and trivial arbitrarynesses. Of course, considering $E = QQ^6$ is just a sample of what I mean.

So, let's move to preferences.
If we assume, as I said in a couple of post before, that the preferences of each player:
- depend only on the money that they get
and that players:
- they like more money than less,
Than it is immediate to show that, given whatever choice of the remaining players, the choice of putting $0$ in the pot is the best choice; i.e. is a strongly dominant choice.

That is, $(0,\ldots,0)$ not only is a Nash equilibrium, but each of its component strategies are strongly dominant.
Exactly as in the prisoner's dilemma, we are facing a very compelling "solution".

And, exactly as for the prisoner's dilemma, the result is inefficient.
This is the source of the enormous interest of such "games". Please notice that:
- these games are "toy games", true. But they represent situations that we face quite often!
- we have not assumed that players are stupid, crazy, or have strange beliefs abot their opponents (or co-players, better...)


I wait for comments. Please don't be too nasty


(*)
I feel obliged to discuss a little bit this "triviality"...
"more or less", since there is always a nontrivial arbitrary choice: which are the truly consequences of the "game" that is being played? It is the nightmare that will wake up in the middle of the night Cozza Taddeo, who has seen half of his money disappear? Or should we consider one of the straight men that put zero in the pot and will not be able to survive the humiliation of having shown to be so greedy to all of the community of this forum? Or it will be the fact that I will be banned forever from this forum after a burning discussion with one of them?

This is a issue that should be well known to economists (I bet it isn't!). An economic agent is assumed to have preferences on his possible consumption acts. But the preferences are on the consequences of consumption not on the consumptive act "per se". So, which are the truly relevant consequences of eating one apple instead of a pear?
Avatar utente
Fioravante Patrone
Cannot live without
Cannot live without
 
Messaggio: 1553 di 10811
Iscritto il: 09/06/2006, 19:18
Località: Temporaneamente a Novi Ligure ;-)

Messaggioda Cheguevilla » 16/07/2007, 17:41

Our "camallo" (alias cheguevilla) has correctly detected the category of problems to which the example belongs.
Ohh, surprise surprise, maybe sometimes longshoremen are not that bad.
Immagine

Rischiavano la strada e per un uomo
ci vuole pure un senso a sopportare
di poter sanguinare
e il senso non dev'essere rischiare
ma forse non voler più sopportare.
Avatar utente
Cheguevilla
Cannot live without
Cannot live without
 
Messaggio: 1802 di 3869
Iscritto il: 12/02/2003, 13:24
Località: København

Messaggioda Cheguevilla » 16/07/2007, 18:05

I will keep myself away from the discussion, at least for this moment.
I only suggest to pay attention in making general theories from abstract examples; this theory ALWAYS works, but we have to consider some particulars:
- players informations level.
- players personal utility function.
Thus, keeping in mind that the second is strictly depending from the former, don't be surprised if, sometimes, a choice may appear "illogical".
In this sense, somebody could trust on the good faith of other players and put 20.
Somebody other could have his's own satisfaction in making the interest of the community before than the personal one.
This is a strong demonstration of how, in many occasions, a cooperating community is better than a community based on individualism and competition.
Obviously, this last sentence, is a purely personal consideration.
Immagine

Rischiavano la strada e per un uomo
ci vuole pure un senso a sopportare
di poter sanguinare
e il senso non dev'essere rischiare
ma forse non voler più sopportare.
Avatar utente
Cheguevilla
Cannot live without
Cannot live without
 
Messaggio: 1803 di 3869
Iscritto il: 12/02/2003, 13:24
Località: København

Messaggioda marco vicari » 21/08/2007, 20:45

I should like to propose a variation in rules of game:
- At the beginning each player receives 20 € and expresses his choice in a reserved area with his keyboard.
- The money collected in the pot will be multiplied by $3$.
- Each player who will have put an inferior amount to the average, he will have to play the difference and one penalty equal to the average (supposing $S_i i=1,2,...,n$ the chosen sum by the player $i$ , $AA S_i<M; P_i=M-S_i+M$ $P_i$ Pi equal to the penalty paid by the player i).
- Only voluntarily inserted sums by the players will be multiplied by $3$, the penalties ( equal to $M$ ) will be inserted in the original value.
How would you play in this case?
marco vicari
Junior Member
Junior Member
 
Messaggio: 10 di 121
Iscritto il: 06/08/2007, 09:59

Precedente

Torna a Matematica per l'Economia e per le Scienze Naturali

Chi c’è in linea

Visitano il forum: Nessuno e 1 ospite