Hierarchies of Beliefs and the Belief-invariant. Bayesian Solution

Size: px
Start display at page:

Download "Hierarchies of Beliefs and the Belief-invariant. Bayesian Solution"

Transcription

1 Herarches of Belefs and the Belef-nvarant Bayesan Soluton Qanfeng Tang March 26, 2015 Abstract The belef-nvarant Bayesan soluton s a noton of correlated equlbrum n games wth ncomplete nformaton proposed by Forges (1993), and herarchy of belefs over condtonal belefs s ntroduced by Ely and Pesk (2006) n ther study of nterm ndependent ratonalzablty. We study the connecton between the two concepts. We partally characterze the correlatons embedded among type spaces wth the same set of herarches of belefs over condtonal belefs wth partally correlatng devces, whch send correlated sgnals to players n a way that preserves each player s belef about others types. Snce the belef-nvarant Bayesan soluton s also mplemented by such correlatng devces, we then establsh that t s nvarant on equvalent type space. The author thanks Marcn Pesk for hs nvaluable gudance and detaled comments, Syang Xong for very benefcal dscussons, and semnar partcpants at Tsnghua Unversty and conference partcpants at 2012 Game Theory Socety World Congress for helpful comments. The author s also grateful to an edtor and two anonymous referees for valuable suggestons that substantally mprove the exposton of the paper. All errors are my own. School of Economcs, Shangha Unversty of Fnance and Economcs. Emal address: tangqanfeng198@gmal.com. 1

2 JEL Classfcaton: C70; C72 Keywords: Games wth ncomplete nformaton, Correlated equlbrum, Herarches of belefs 1 Introducton Harsany ( ) proposes type spaces to model players belefs and hgher-order belefs n games wth ncomplete nformaton, and later Mertens and Zamr (1985) construct a unversal type space whch ncorporates all herarches of belefs. These works provde the foundatons for strategc analyss of games wth ncomplete nformaton. One phenomenon that recently attracts game theorsts attenton s that, for a gven soluton concept, type spaces and herarches of belefs are not always strategcally equvalent. To be more precse, multple type spaces can represent the same set of herarches of belefs and are hence equvalent n ths respect. However, equvalent type spaces may dffer n the amount of correlatons ncorporated among types, and consequently, for a gven soluton concept, they may nduce dfferent predctons on the actons played by types wth the same herarchy of belefs. In partcular, type spaces that represent the same set of Mertens-Zamr herarches of belefs always nduce the same nterm correlated ratonalzable outcomes (Dekel et al., 2006; Dekel et al., 2007), but may nduce dfferent nterm ndependently ratonalzable outcomes (Ely and Pesk, 2006) and Bayes Nash equlbrum outcomes. The mplct correlatons, or hdden nformaton, that herarches of belefs fal to ncorporate are exactly those embedded among equvalent type spaces. Lu (2015) shows how to dstll such correlatons and explctly characterzes them va state-dependent correlatng devces. Accordng to the characterzaton, a type space has the same set of 2

3 Mertens-Zamr herarches of belefs as a non-redundant type space f and only f t can be generated from the conjuncton of the non-redundant type space and a state-dependent correlatng devce. 1 Therefore, the characterzaton decomposes the nformaton n type space nto a correlatng devce and the prmtve nformaton the herarches of belefs. Herarchy of belefs over condtonal belefs, also called -herarchy of belefs, s ntroduced by Ely and Pesk (2006) and then extended by Tang (2015) n the study of nterm ratonalzablty. It ncorporates rcher nformaton than Mertens-Zamr herarchy of belefs. As a result, t dentfes a stronger noton of soluton concept Ely and Pesk show that two types wth the same herarchy of belefs always have the same set of nterm ndependently ratonalzable actons. In the same sprt of Lu (2015) s work, we are nterested n understandng the correlatons embedded among type spaces wth the same set of -herarches of belefs and ther connecton to correlated equlbrum n games wth ncomplete nformaton. We begn wth provdng a partal characterzaton of such correlatons. For our purpose, two type spaces are sad to be equvalent f they have the same set of -herarches of belefs. A partally correlatng devce sends correlated sgnals to players based on nterm stage nformaton and preserves each player s belef about others types. We show that any type space s equvalent to ts conjuncton wth a partally correlatng devce, and that f two type spaces are equvalent, then they can always generate the same type space va conjunctons wth partally correlatng devces. The correlatons ncorporated n partally correlatng devces and the belef-nvarant Bayesan soluton (Forges, 1993;Forges, 2006) take the same form. When the sgnals n the correlatng devce are smply recommendatons of actons to the players, the conjuncton of the type space and the correlatng devce generates exactly an epstemc model for the 1 In a closely related work, Lu (2009) shows that under mld assumptons, a redundant type space can be represented by a non-redundant type space defned on an extended space of basc uncertantes. 3

4 belef-nvarant Bayesan soluton. Based on the partal characterzaton of correlatons obtaned, we establsh that for any game-form, the set of belef-nvarant Bayesan soluton payoffs on a type space s the unon of ts Bayesan Nash equlbra payoffs across equvalent type spaces. And n an mmedate corollary, we show that the belef-nvarant Bayesan soluton s nvarant across equvalent type spaces. Therefore, we have obtaned a belef foundaton for the belef-nvarant Bayesan soluton. To dentfy the set of all such solutons, t s wthout loss of generalty for us to consder the unversal type space of -herarches of belefs. 2 Ths paper s closely related to and s n parallel wth Lu (2015). We can fnd counterparts for varous concepts and results of our paper n Lu s work. Snce -herarchy of belefs ncorporates rcher nformaton than Mertens-Zamr herarchy of belefs, t dentfes stronger notons of soluton concepts (nterm ndependent ratonalzablty and the belef-nvarant Bayesan soluton versus nterm correlated ratonalzablty and Lu s noton of correlated equlbrum), and leaves less nformaton embedded among equvalent type spaces (partally correlatng devces versus state-dependent correlatng devces). Lehrer et al. (2010) and Lehrer et al. (2013) study the relatonshp between garblngs and the equvalence of type spaces. The non-communcatng garblngs that they use have smlar features as partally correlatng devces. Bergemann and Morrs (2014) propose and study the Bayes correlated equlbrum. They focus on correlatng devces that are dependent on ex post stage nformaton but free from belef-nvarance restrctons. The rest of ths paper s organzed as follows. We present the basc notatons Secton 2, and formulate herarches of belefs n Secton 3. In Secton 4, we characterze the correlatons embedded n equvalent type spaces and then apply the result on the belef- 2 An applcaton s on the robustness of soluton concepts to ncomplete nformaton, a study ntated by Kaj and Morrs (1997). The belef foundaton of the belef-nvarant Bayesan soluton allows us to study ts robustness by perturbng the -herarchy of belefs of a type, usng the nterm approach taken by Wensten and Yldz (2007) n ther study of the robustness of nterm correlated ratonalzablty. 4

5 nvarant Bayesan soluton. Secton 5 concludes. 2 Notatons For any metrc space X, let X denote the space of probablty measures on the Borel σ-algebra of X endowed wth the weak -topology. Let the product of two metrc spaces be endowed wth the product Borel σ-algebra. For any probablty measure µ X, let supp µ denote the support of µ, and for any measure µ (X Y), let marg X µ denote the margnal of µ on X. We study games wth ncomplete nformaton wth n players. The set of players s N = {1, 2,..., n}. For each N, let denote the set of s opponents. Players play a game n whch the payoffs are uncertan and parameterzed by a fnte set Θ. Each element θ Θ s called a state of nature. For each N, let A be the set of actons for player and let A N A be the set of acton profles. A (strategc-form) game s a profle G = (g, A ) N. For each N, we assume the payoff functon s bounded: g : A Θ [ M, M], for some postve real number M. The set of fnte bounded games s denoted by G. A type space over Θ s defned as T = (T, π ) N, where for each, T s a fnte set of types for player and π : T (T Θ) s a belef mappng such that π (t )[(t, θ)] descrbes player s belef on the event that the others type profle s t and the state of nature s θ. 3 3 Throughout, gven arbtrary x X and y Y, we use the notaton π (x)[y] to denote player s belef about y condtonal on x. More precsely, the object n the round bracket always denotes the object player that condtons on, and the object n the square bracket always denotes the object that assgns probablty to. 5

6 3 Herarches of belefs 3.1 Formulaton We frst present Mertens and Zamr s standard formulaton of herarches of belefs (see also Brandenburger and Dekel, 1993), and based on that present Ely and Pesk s constructon of -herarches of belefs. Let X 0 = Θ, and for each k 1, let X k = X k 1 ( (X k 1 )) n 1. Let h 1 (t ) = marg Θ π (t ) be player s belef over Θ at type t. For each k 2, let h k (t )[S] = π (t )[{(θ, t ) : (θ, (h l (t )) 1 l k 1 ) S}], for any measurable subset S X k 1. In ths constructon, h k (t ) (X k 1 ) represents player s k-th order belef at t. The profle h(t ) = (h 1 (t ),..., h k (t ),...) k=0 X k s called player s Mertens-Zamr herarchy of belefs at type t. As extensons of Mertens-Zamr herarches of belefs, -herarches of belefs descrbe players belefs and hgher-order belefs about the condtonal belefs on states of nature. Ths concept s ntroduced by Ely and Pesk (2006) n ther study of nterm ndependent ratonalzablty. We begn wth defnng condtonal belefs. Gven belef π (t ) (T Θ), the condtonal belef of type t over Θ, condtonng on the others types beng t, s denoted by π (t )(t ) (Θ), also wrtten as π (t, t ). For any type space T, let B (t ) = {π (t, t ) (Θ) : t T } be the set of all possble condtonal belefs at t. Type t s belef over T then nduces a belef over B (t ) (Θ) : for any measurable subset S (Θ), π (t )[S] = π (t )[{t : π (t, t ) S}]. The defnton of -herarchy of belefs at t treats the set of possble condtonal belefs,.e., (Θ), as the set of basc uncertantes. In a -herarchy of belefs, the frst-order belef of a player s her belef over the set of condtonal belefs, the second-order belef s her belef over the others belefs over the set of condtonal belefs, and so on. 6

7 For any type space T = (T, π ) N on Θ, we can transform t nto a type space T = (T, π ) N on (Θ). In the new type space, players types are unchanged, and player s belef at type t s π (t ) (T (Θ)), such that π (t )[S] = π (t )[{t : (t, π (t, t )) S}], for any measurable subset S T (Θ). In the type space T on (Θ), let the Mertens-Zamr herarchy of belefs at t be denoted by h(t T ). Defnton 1. For any type space T and nteger k 1, let the k-th order -herarchy of belefs at t T be h k (t T ) and denote t by δ k (t ). Also, let δ(t ) = (δ 1 (t ),..., δ k (t ),...) denote the -herarchy of belefs at t. By defnton, δ(t ) = h(t T ). For player, we use δ to denote the profle of the others -herarches of belefs. 3.2 Equvalence of type spaces For any type space T, let the set of all -herarches of belefs of player be Λ (T) = {δ(t ) : t T }. Just lke Mertens-Zamr herarches of belefs, the set of -herarches of belefs does not unquely pn down a type space. Instead, multple type spaces may nduce the same set of -herarches of belefs. Defnton 2. Two type spaces T and T are equvalent, wrtten as T T, f they have the same set of -herarches of belefs for all players, that s, f Λ (T) = Λ (T ), N. 7

8 A type space n whch dfferent types of a player always have dfferent herarches of belefs s called a reduced type space (Aumann, 1998), or a non-redundant type space (Mertens and Zamr, 1985). For any Mertens-Zamr herarchy of belefs, we are able to construct such a type space that generates t. However, ths s not true for -herarches of belefs. We llustrate ths wth a smple type space taken from Ely and Pesk (2006). Example 1. Consder a type space T n whch Θ = T 1 = T 2 = {+1, 1}, and players belefs are updated from a common pror π (Θ T 1 T 2 ) such that π(t 1, t 2, θ) = 1 4 f t t 2 = θ, 0 otherwse. In ths type space, the set of condtonal belefs for each type contans pont mass on θ = +1 and pont mass on θ = 1, and both types of player 1 (or player 2) have the same -herarchy of belefs common certanty of equal probablty on the pont masses. However, type space T s the most compact one that supports ths -herarchy of belefs. 4 Man results 4.1 A partal characterzaton of correlatons Wthout dstngushng non-redundant and redundant type spaces, we can acheve a partal characterzaton of the correlaton embedded across equvalent type spaces. As we shall see, ths partal characterzaton s suffcent for the nvarance of the belef-nvarant soluton. Defnton 3. For any type space T, a partally correlatng devce on T s a profle Q = (q, S ) N, where for each N, S s a fnte set of sgnals and q : T S, 8

9 where S = N S, s a belef mappng such that 1. For any = j, t T, supp q (t) = supp q j (t). 2. belef nvarance s satsfed: at dfferent types t, t of the others, player receves s wth the same probablty,.e., q (t, t )[s ] = q (t, t )[s ],, t, s. {s S:s =s } {s S:s =s } From the defnton, player beleves that at (t, t ) T, the partally correlatng devce selects a sgnal profle (s, s ) S accordng to the dstrbuton q (t, t ) S, and for each j N, s j wll be reported by a medator to player j. Belef nvarance ensures that from the sgnals that the players receve, they cannot nfer any extra nformaton about the others types. Also note that the correlated sgnals depend only on the nterm stage nformaton players types not on states of nature. The partally correlatng devce here and the state-dependent correlatng devce of Lu (2015) take smlar forms, whle dfferng from each other manly n two ways: () n the former, correlated sgnals are dependent on the nterm stage nformaton (the type profle of players), whle n the latter, they are dependent on the ex post stage nformaton (the state of the world,.e., both the type profle of players and the true state of nature); and () the belef-nvarance restrcton of the former requres that the sgnals do not change each player s belef about the others types, whle that of the latter requres that they do not change each player s belef about the states of the world. Every partally correlatng devce s a state-dependent correlatng devce, but the converse s not true. When the partally correlatng devce uses actons as sgnals and therefore the sgnals are smply recommendatons of play, we say that the correlatng devce s canoncal. Defnton 4. If for all N, S = A, then Q s a canoncal partally correlatng devce. 9

10 Let q (t, t )[s s ] be player s belef on the others recevng the sgnal profle s, condtonal on her recevng s. Ths condtonal belef can be easly calculated gven q : T S. Defnton 5. For any type space T = (T, π ) N and any partally correlatng devce Q = (q, S ) N, let T Q be the enlarged type space generated from T through operatng Q on T. More precsely, T Q = (T Q, π Q ) N such that T Q = {(t, s ) : t T, q (t)[s ] > 0, for some t T }, and for all (t, s ) T Q, θ Θ and (t, s ) T Q, π Q ((t, s ))[((t, s ), θ)] = π (t )[(t, θ)] q (t, t )[s s ]. Intutvely, we can vew T Q as the type space faced by players at the ex ante stage, when n addton a correlatng devce s expected to send sgnals to players after the realzaton of players types. The followng theorem provdes a partal characterzaton of the correlaton embedded n equvalent type spaces. Proposton 1. Let T be any type space. Then 1. for any partally correlatng devce Q, T Q T; more specfcally, for any (t, s ) T Q, δ((t, s )) = δ(t ). 2. for any type space ˆT satsfyng T ˆT, there exst partally correlatng devces Q and ˆQ such that T Q = ˆT ˆQ. The frst part of the proposton states that the enlarged type space T Q generated by type space T and any partally correlatng devce Q always has the same set of - 10

11 herarches of belefs as T does. In partcular, after recevng any sgnal s at type t, player s -herarchy of belefs does not change. The second part of the proposton states that for any par of type spaces T and ˆT whch have the same set of -herarches of belefs, we can fnd two partally correlatng devce Q and ˆQ through whch T and ˆT generate the same enlarged type space. Our characterzaton of correlaton s partal because unlke Lu (2015), we do not show that any redundant type space can be enlarged from the nonredundant type space wth a correlatng devce. Instead, we establsh the connecton between two equvalent type spaces wthout dstngushng redundant and non-redundant type spaces. The proof of ths proposton s relegated nto Appendx A. The proof of part 1 s by nducton, and the key s the belef nvarance property of partally correlatng devces. Next, we brefly sketch the ntuton for the proof of part 2. We transform the type space ˆT nto a partally correlatng devce Q and transform the type space T nto a partally correlatng devce ˆQ. Then we show that T Q and ˆT ˆQ both represent some verson of the product type space T ˆT and hence are the same. Specfcally, when the correlatng devce Q s operated on type space T, from player s vew, at any type profle (t, t ), the set of sgnals she may receve from Q s exactly the set of types n the type space ˆT that have the same -herarchy of belefs as t, and when she receves any such sgnal ˆt, the set of profles of sgnals that her opponents may receve from Q s exactly the set of type profles ˆt n ˆT that nduce the same profle of -herarches as t and the same condtonal belef at ˆt as π (t, t ). 11

12 4.2 The belef-nvarant Bayesan soluton The belef-nvarant Bayesan soluton s a noton of correlated equlbrum for games wth ncomplete nformaton proposed by Forges (1993) 4 and later dscussed n Forges (2006) and Bergemann and Morrs (2014). A belef-nvarant Bayesan soluton on a type space T s defned as an epstemc model that embeds T n a belef-nvarant way and satsfes Bayesan ratonalty at each state of the world. As Forges (2006) ponts out, each such epstemc model s an enlarged type space T Q for some partally correlatng devce Q, and each belef-nvarant soluton corresponds to a Bayes Nash equlbrum (BNE) of the game (G, T Q ). 5 Therefore, the set of belef-nvarant Bayesan soluton payoffs on T s the unon of BNE payoffs across enlarged type spaces. Let Q denote the set of all partally correlatng devces. For each game (G, T) and a BNE σ of ths game, let g (t σ) be player s nterm payoff at type t, and let g(t σ) be the nterm payoff vector at type profle t T. Denote the set of nterm BNE payoffs of the game (G, T) as NE(G, T) = {g(t σ) : t T, σ s a BNE of (G, T)}. Let B I (G, T) denote the set of nterm belef-nvarant Bayesan soluton payoffs, then B I (G, T) = {Q:Q Q}NE(G, T Q ). The result below states that the set of players nterm payoffs from belef-nvarant Bayesan solutons at a type space s exactly the unon of nterm BNE payoffs on equva- 4 Forges defnton of the Bayesan soluton s restrcted to two-player games for type spaces wth common prors; what we present here s the n-player non-common pror analogue of her defnton. 5 From a pont of vew analogous to the "revelaton prncple" n the mechansm desgn lterature, Forges (2006) also characterzes belef-nvarant Bayesan soluton payoffs wth canoncal partally correlatng devces. 12

13 lent type spaces. Proposton 2. B I (G, T) = { ˆT: ˆT T} NE(G, ˆT). Proof. By Proposton 1, for all Q Q, T T Q. Therefore, t s straghtforward that {Q:Q Q}NE(G, T Q ) { ˆT: ˆT T} NE(G, ˆT). We also know that for any ˆT such that ˆT T, there exst ˆQ and Q such that ˆT ˆQ = T Q. Therefore, NE(G, ˆT) NE(G, ˆT ˆQ ) NE(G, ˆT) NE(G, T Q ). That s, we also have { ˆT: ˆT T} NE(G, ˆT) {Q:Q Q}NE(G, T Q ). Lehrer et al. (2010) and Lehrer et al. (2013) study the payoff equvalence of nformaton structures wth respect to belef-nvarant Bayesan solutons. They show that two nformaton structures are equvalent f and only f there exst non-communcatng garblng transformatons between them. Non-communcatng garblngs share smlar features as partally correlatng devces. The man dfference between them s that a garblng garbles orgnal sgnals, whle a correlatng devce ntroduces new sgnals nto the structure. It s mmedate from Proposton 2 that f two type spaces represent the same set of - herarches of belefs, they must nduce the same set of belef-nvarant Bayesan soluton payoffs n any game. In other words, the belef-nvarant Bayesan soluton s nvarant on the equvalent class of type spaces. Corollary 1. If ˆT T, then B I (G, T) = B I (G, ˆT). 13

14 5 Concluson We study the correlatons embedded n type spaces wth the same set of herarches of belefs over condtonal belefs, t turns out that such correlatons can be expressed explctly wth partally correlatng devces, whch operate n the nterm stage of the game. Wth these results, we see a complete pcture of the connectons among -herarches of belefs, nterm partally correlated ratonalzablty, and the belef-nvarant Bayesan soluton. We already know that Partally correlatng devces characterze correlatons embedded n type spaces wth the same set of -herarches of belefs, and mplement the Bayesan soluton. Furthermore, Tang (2015) shows that -herarches of belefs fully dentfy nterm partally correlated ratonalzablty. As a result, nterm partally correlated ratonalzablty and the belef-nvarant Bayesan soluton are payoff equvalent. A Appendx: Proof of Proposton 1 Part I. We use nducton to show that for any (t, s ) T Q, δ((t, s )) = δ(t ). Frst note that for any (t, s ) T Q, (t, s ) T Q, and θ Θ, π Q ((t, s ), (t, s ))[θ] = πq ((t, s ))[(t, s ), θ] π Q ((t, s ))[(t, s )] = π (t )[(t, θ)] q (t, t )[(s, s )] π (t )[t ] q (t, t )[(s, s )] = π (t, t )[θ]. Therefore, for any (t, s ) T Q, the set of condtonal belefs at (t, s ) s the same as 14

15 that at t. Furthermore, for any condtonal belef β B (t ), π Q ((t, s ))[β] = π Q ((t, s ))[{(t, s ) : π Q ((t, s ), (t, s )) = β}] = π Q ((t, s ))[{(t, s ) : π (t, t ) = β}] = π Q ((t, s ))[{t : π (t, t ) = β}] = π (t )[{t : π (t, t ) = β}] = π (t )[β]. The fourth equaton above comes from belef nvarance. We have proved that for all (t, s ) T Q, δ 1 ((t, s )) = δ 1 (t ). For hgher-order belefs, we prove by nducton. Now suppose for all 0 < l k and (t, s ) T Q, δ l ((t, s )) = δ l (t ), we show that for all (t, s ) T Q, δ k+1 ((t, s )) = δ k+1 (t ). Let the support of the l-th order belef at type t be B l (t ). As a result, the set of condtonal belefs s relabeled as B 1 (t ). By the premses of nducton, for all (t, s ) T Q and 0 < l k, B l ((t, s )) = B l (t ). Indeed, for any (β, δ 1,..., δ k ) 0<l k B l (t ), δ k+1 ((t, s ))[(β, δ 1,..., δ k )] = π Q ((t, s ))[{(t, s ) : (π Q ((t, s ), (t, s )), (δ l ((t, s ))) k l=1 ) = (β, δ1,... δ k )}] = π Q ((t, s ))[{(t, s ) : π (t, t ) = β, δ 1 (t ) = δ 1,..., δ k (t ) = δ k }] = π (t )[{t : π (t, t ) = β, δ 1 (t ) = δ 1,..., δ k (t ) = δ k }] = δ k+1 (t )[(β, δ 1,..., δ k )]. By nducton, for all (t, s ) T Q, δ((t, s )) = δ(t ). Naturally, T Q and T have the same set of -herarches of belefs, T Q T. Part II. Fx a par of type spaces T = (T, π ) N and ˆT = ( ˆT, ˆπ ) N. Suppose T ˆT, 15

16 we now construct Q and ˆQ such that T Q = ˆT ˆQ. To do that, we transform the type space ˆT nto a partally correlatng devce Q and transform T nto a partally correlatng devce ˆQ. We then show that the generated type spaces T Q and ˆT ˆQ are the same. Step 1. Before we start, we need a few ntermedate results. The lemma below says that f two types have the same -herarchy of belefs, then they must have the same belef over condtonal belefs and the others -herarches of belefs. Lemma 1. Fx type spaces T and T. If t T, t T and δ(t ) = δ(t ), then π (t )[(β, δ )] = π (t )[(β, δ )], β, δ. Proof. Wth the basc property of probablty measures, π (t )[(β, δ )] = π (t )[{t : π (t, t ) = β, δ 1 (t ) = δ 1,..., δn (t ) = δ n,...}] = π (t )[ n {t : π (t, t ) = β, δ 1 (t ) = δ 1,..., δn (t ) = δ n }] = lm π (t )[{t : π (t, t ) = β, δ 1 (t ) = δ 1 n,..., δn (t ) = δ n }] = lm δ n+1 (t )[(β, δ 1,..., δ n )] n = lm δ n+1 (t n )[(β, δ1,..., δ n )] = π (t )[(β, δ )]. Also, f two types t T, t T have the same -herarchy of belefs, and f t has a condtonal belef π (t, t ), then t must also have such a condtonal belef, condtonal on some type profle of the others t whch have the same -herarches of belefs as t. Lemma 2. Fx type spaces T and T. Suppose t T, t T, and δ(t ) = δ(t ). Then for any t T that satsfes π (t )[t ] > 0, there exsts t T that satsfes δ(t ) = δ(t ) and π (t )[t ] > 0, such that π (t, t ) = π (t, t ). 16

17 Proof. We prove by contradcton. Suppose t s not the case. Then there exsts a t that satsfes π (t )[t ] > 0 and π (t, t ) = β, such that for all t that satsfes π (t, t ) = β, π (t )[t ] > 0, t must be that δ(t ) = δ(t ). As a result, π (t )[(β, δ (t ))] = 0. However, π (t )[(β, δ (t ))] π (t )[t ] > 0. Gven Lemma 1, ths s n contradcton wth δ(t ) = δ(t ). Step 2. Usng nformaton n type space ˆT, we now construct a partally correlatng devce Q = (q, S ) N whch s to be operated on type space T. For each N, let the set of sgnals for player be S = ˆT, and defne S N S. Defne S (t ) {ˆt ˆT : δ(ˆt ) = δ(t )} and S (t, t ˆt )) {ˆt ˆT : δ(ˆt ) = δ(t ) and ˆπ (ˆt, ˆt ) = π (t, t )}. Intutvely, we are gong to construct q : T S n a way such that the set of sgnals that player could possbly receve when her type s t s restrcted to be S (t ), whch s the set of t s equvalent types n ˆT. Smlarly, S (t, t ˆt )) wll be the restrcted set of sgnals ˆt that players may receve at type profle t from player s vew, when her own type s t and she receves sgnal ˆt. We need the followng result, whch s mmedate from Lemma 1 and Lemma 2, n the constructon of q. It states that f two types are equvalent n ˆT, then ther belefs over condtonal belefs and the others -herarches of belefs must be the same. Lemma 3. If ˆt, û S (t ), then ˆπ (ˆt )[S (t, t ˆt ))] = ˆπ (û )[S (t, t û )]. 17

18 Defne on the type space ˆT a pror ˆp ( ˆT ˆT Θ) for player as follows: ˆp [(ˆt, ˆt, θ)] = 1 ˆT ˆπ (ˆt )[(ˆt, θ)], (ˆt, ˆt, θ) ˆT ˆT Θ. From player s vew, the partally correlatng devce operates only n states of the world (ˆt, ˆt, θ) such that ˆp (ˆt, ˆt, θ) > 0. For each N, we can construct the belef system q : T S as follows: q (t, t )[(ˆt, ˆt )] = ˆp [(ˆt,ˆt )] ˆp [S (t ) S (t,t ˆt )], f (ˆt, ˆt ) S (t ) S (t, t ˆt ); 0, otherwse. Wth Lemma 3, for any (ˆt, ˆt ) S (t ) S (t, t ˆt ), q (t, t )[(ˆt, ˆt )] = = ˆp [ˆt ] ˆπ (ˆt )[(ˆt, θ)] û S (t ) ˆp [û ] ˆπ (û )[S (t, t ) û ] 1/ ˆT 1/ ˆT S (t ) ˆπ (ˆt )[(ˆt, θ)] ˆπ (ˆt )[S (t, t ) ˆt ]. The expresson of q can be rewrtten as q (t, t )[(ˆt, ˆt )] = 1 S (t ) ˆπ (ˆt )[ˆt ] ˆπ (ˆt )[S (t,t ˆt )], f (ˆt, ˆt ) S (t ) S (t, t ˆt ); 0, otherwse. Now we prove that the Q defned above s ndeed a partally correlatng devce. Frst, for any = j, t T, supp q (t) = supp q j (t) = k N S k (t k ). Ths s because from player s vew, each ˆt S (t ) s sent to her wth probablty 1 S (t ), and that for each ˆt k N\{} S k (t k ), there must be ˆt S (t ) such that ˆt 18

19 S (t, t ˆt ) and ˆπ (ˆt )[ˆt ] > 0, due to Lemma 2. Second, belef nvarance s satsfed: for any (t, t ) T and any û S (t ), the probablty that player wll receve sgnal û s q (t, t )[(û, ˆt )] = {ˆt ˆT:ˆt =û } {ˆt :ˆt S (t,t û )} 1 S (t ) ˆπ (û )[ˆt ] ˆπ (û )[S (t, t û )] = 1 {ˆt :ˆt S (t,t û )} ˆπ (û )[ˆt ] S (t ) ˆπ (û )[S (t, t û )] = 1 S (t ) = 1 S (t ), ˆπ (û )[S (t, t û )] ˆπ (û )[S (t, t û )] whch s ndependent of t. Thus the sgnal does not provde extra nformaton on the others types. Step 3. Gven the partally correlatng devce Q constructed usng nformaton n ˆT, we can generate a new type space T Q = (T Q, π Q ) N from the type space T. In T Q, T Q = {(t, ˆt ) : t T, ˆt S (t )}, and for any (ˆt, ˆt ) S (t ) S (t, t ˆt ), π Q ((t, ˆt ))[((t, ˆt ), θ)] = π (t )[(t, θ)] ˆπ (ˆt )[ˆt ] ˆπ (ˆt )[S (t, t ˆt )]. Smlarly, we can construct another partally correlatng devce ˆQ usng nformaton n the type space T, and generate a new type space ˆT ˆQ from ˆT. In ˆT ˆQ, ˆT ˆQ ˆT, t S (ˆt )}, and for any (t, t ) S (ˆt ) S (ˆt, ˆt t ), = {(ˆt, t ) : ˆt π ˆQ π ((ˆt, t ))[((ˆt, t ), θ)] = ˆπ (ˆt )[(ˆt, θ)] (t )[t ] π (t )[S (ˆt, ˆt t )]. It s straghtforward that T Q = ˆT ˆQ, N. Now we show π Q ((t, ˆt )) = π ˆQ ((ˆt, t )). By the defnton, for any (t, t ) and (ˆt, ˆt ) S (t ) S (t, t ˆt ), we know that 19

20 π (t, t ) = ˆπ (ˆt, ˆt ) = β, δ(t ) = δ(ˆt ) = δ, for some β and δ. We can decompose the belef π Q as follows: π Q ((t, ˆt ))[((t, ˆt ), θ)] = π (t, t )[θ] π (t )[t ] = π (t, t )[θ] π(t )[t ] ˆπ (ˆt )[ˆt ]. π (t )[(β, δ )] ˆπ (ˆt )[ˆt ] ˆπ (ˆt )[{ˆt : δ(ˆt ) = δ(t ), ˆπ (ˆt, ˆt ) = π (t, t )}] Smlarly, π ˆQ ((ˆt, t ))[((ˆt, t ), θ)] can also be decomposed: π ˆQ ((ˆt, t ))[((ˆt, t ), θ)] = ˆπ (ˆt, ˆt )[θ] ˆπ (ˆt )[ˆt ] π (t )[t ]. ˆπ (ˆt )[(β, δ )] We compare π Q and π ˆQ term by term. Frst, π (t, t )[θ] = ˆπ (ˆt, ˆt )[θ]. Second, π (t )[t ] ˆπ (ˆt )[ˆt ] = ˆπ (ˆt )[ˆt ] π (t )[t ]. Thrd, from Lemma 1, π (t )[(β, δ )] = ˆπ (ˆt )[(β, δ )]. Snce for any N, (t, ˆt ) T Q = ˆT ˆQ, π Q ((t, ˆt )) = π ˆQ ((ˆt, t )), we have T Q = ˆT ˆQ. 20

21 References Aumann, Robert (1998), Common prors: a reply to Gul, Econometrca, 55, Bergemann, Drk and Stephen, Morrs (2014), Bayes correlated equlbrum and the comparson of nformaton structures n games, workng paper, Prnceton Unversty. Brandenburger, Adam and Edde Dekel (1993), Herarches of belefs and common knowledge, Journal of Economc Theory, 59, Dekel, Edde, Drew Fudenberg and Stephen Morrs (2006), Topologes on types, Theoretcal Economcs, 1, Dekel, Edde, Drew Fudenberg and Stephen Morrs (2007), Interm correlated ratonalzablty, Theoretcal Economcs, 2, Ely, Jeff and Marcn Pesk (2006), Herarches of belef and nterm ratonalzablty, Theoretcal Economcs, 1, Forges, Francose (1993), Fve legtmate defntons of correlated equlbrum n games wth ncomplete nformaton, Theory and decson, 35, Forges, Francose (2006), Correlated equlbrum n games wth ncomplete nformaton revsted, Theory and decson, 61, Harsany, John ( ), Games wth ncomplete nformaton played by "Bayesan" players, Parts I, II, III, Management Scence, 14, , , Kaj, Atsush and Stephen Morrs (1997), The robustness of equlbra to ncomplete nformaton, Econometrca, 65, Lehrer, Ehud, Dnah Rosenberg and Eran Shmaya (2010), Sgnalng and medaton n games wth common nterests, Games and Economc Behavor, 68,

22 Lehrer, Ehud, Dnah Rosenberg and Eran Shmaya (2013), Garblng of sgnals and outcome equvalence, Games and Economc Behavor, 81, Lu, Qngmn (2009), On redundant types and Bayesan formulaton of ncomplete nformaton, Journal of Economc Theory, 144, Lu, Qngmn (2015), Correlaton and common prors n games wth ncomplete nformaton, Journal of Economc Theory, 157, Mertens, Jean-Francos and Shmuel Zamr (1985), Formulaton of Bayesan analyss for games wth ncomplete nformaton, Internatonal Journal of Game Theory, 14, Tang, Qanfeng (2015), Interm partally correlated ratonalzablty, Games and Economc Behavor, accepted. Wensten, Jonathan and Muhamet Yldz (2007), A structure theorem for ratonalzablty wth applcaton to robust predctons of refnements, Econometrca, 75,

Genericity of Critical Types

Genericity of Critical Types Genercty of Crtcal Types Y-Chun Chen Alfredo D Tllo Eduardo Fangold Syang Xong September 2008 Abstract Ely and Pesk 2008 offers an nsghtful characterzaton of crtcal types: a type s crtcal f and only f

More information

Hierarchies of belief and interim rationalizability

Hierarchies of belief and interim rationalizability Theoretcal Economcs 1 (2006), 19 65 1555-7561/20060019 Herarches of belef and nterm ratonalzablty JEFFREY C. ELY Department of Economcs, Northwestern Unversty MARCIN PESKI Department of Economcs, Unversty

More information

Interim Correlated Rationalizability 1

Interim Correlated Rationalizability 1 Interm Correlated Ratonalzablty Edde Dekel Northwestern Unversty and Tel Avv Unversty Drew Fudenberg Harvard Unversty Stephen Morrs Prnceton Unversty Frst Draft: May 2003. Ths Draft: November 2006 Ths

More information

CMS-EMS Center for Mathematical Studies in Economics And Management Science. Discussion Paper #1586

CMS-EMS Center for Mathematical Studies in Economics And Management Science. Discussion Paper #1586 CMS-EMS Center for Mathematcal Studes n Economcs And Management Scence Dscusson Paper #1586 "When Do Types Induce the Same Belef Herarchy?" Andres Perea and Wllemen Kets Northwestern Unversty December

More information

HIERARCHIES OF BELIEF AND INTERIM RATIONALIZABILITY

HIERARCHIES OF BELIEF AND INTERIM RATIONALIZABILITY HIERARCHIES OF BELIEF AND INTERIM RATIONALIZABILITY JEFFREY C. ELY AND MARCIN PESKI Abstract. In games wth ncomplete nformaton, conventonal herarches of belef are ncomplete as descrptons of the players

More information

Critical Types. September 24, Abstract

Critical Types. September 24, Abstract Crtcal Types Jeffrey C. Ely Marcn Pesk September 24, 2010 Abstract How can we know n advance whether smplfyng assumptons about belefs wll make a dfference n the conclusons of game-theoretc models? We defne

More information

Correlated Equilibrium in Games with Incomplete Information

Correlated Equilibrium in Games with Incomplete Information Correlated Equlbrum n Games wth Incomplete Informaton Drk Bergemann y Stephen Morrs z Frst Verson: October Current Verson: May 4, Abstract We de ne a noton of correlated equlbrum for games wth ncomplete

More information

Games in Preference Form and Preference Rationalizability

Games in Preference Form and Preference Rationalizability Games n Preference Form and Preference Ratonalzablty Stephen Morrs Satoru Takahash September 8, 2012 Abstract We ntroduce a game n preference form, whch conssts of a game form and a preference structure,

More information

Subjective Uncertainty Over Behavior Strategies: A Correction

Subjective Uncertainty Over Behavior Strategies: A Correction Subjectve Uncertanty Over Behavor Strateges: A Correcton The Harvard communty has made ths artcle openly avalable. Please share how ths access benefts you. Your story matters. Ctaton Publshed Verson Accessed

More information

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal Inner Product Defnton 1 () A Eucldean space s a fnte-dmensonal vector space over the reals R, wth an nner product,. Defnton 2 (Inner Product) An nner product, on a real vector space X s a symmetrc, blnear,

More information

Interim Rationalizability. Eddie Dekel, Drew Fudenberg and Stephen Morris. Working Paper No September, 2005

Interim Rationalizability. Eddie Dekel, Drew Fudenberg and Stephen Morris. Working Paper No September, 2005 Interm Ratonalzablty By Edde Dekel, Drew Fudenberg and Stephen Morrs Workng Paper No. 6-2005 September, 2005 The Foerder Insttute for Economc Research and The Sackler Insttute of Economc Studes Interm

More information

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and Ths artcle appeared n a journal publshed by Elsever. The attached copy s furnshed to the author for nternal non-commercal research and educaton use, ncludng for nstructon at the authors nsttuton and sharng

More information

An Explicit Approach to Modeling Finite-Order Type Spaces and Applications

An Explicit Approach to Modeling Finite-Order Type Spaces and Applications An Explct Approach to Modelng Fnte-Order Type Spaces and Applcatons Cheng-Zhong Qn and Chun-Le Yang November 18, 2009 Abstract Every abstract type of a belef-closed type space corresponds to an nfnte belef

More information

A note on the one-deviation property in extensive form games

A note on the one-deviation property in extensive form games Games and Economc Behavor 40 (2002) 322 338 www.academcpress.com Note A note on the one-devaton property n extensve form games Andrés Perea Departamento de Economía, Unversdad Carlos III de Madrd, Calle

More information

Common Belief Foundations of Global Games

Common Belief Foundations of Global Games Common Belef Foundatons of Global Games Stephen Morrs Prnceton Unversty Hyun Song Shn Bank for Internatonal Settlements November 2015 Muhamet Yldz M.I.T. Abstract We study coordnaton games under general

More information

Games of Threats. Elon Kohlberg Abraham Neyman. Working Paper

Games of Threats. Elon Kohlberg Abraham Neyman. Working Paper Games of Threats Elon Kohlberg Abraham Neyman Workng Paper 18-023 Games of Threats Elon Kohlberg Harvard Busness School Abraham Neyman The Hebrew Unversty of Jerusalem Workng Paper 18-023 Copyrght 2017

More information

CS286r Assign One. Answer Key

CS286r Assign One. Answer Key CS286r Assgn One Answer Key 1 Game theory 1.1 1.1.1 Let off-equlbrum strateges also be that people contnue to play n Nash equlbrum. Devatng from any Nash equlbrum s a weakly domnated strategy. That s,

More information

Bounded Reasoning and Higher-Order Uncertainty

Bounded Reasoning and Higher-Order Uncertainty Bounded Reasonng and Hgher-Order Uncertanty Wllemen Kets September 4, 2014 Frst verson: November 2011 Abstract Expermental evdence suggests that ndvduals cannot reason about ther opponents belefs up to

More information

Appendix for Causal Interaction in Factorial Experiments: Application to Conjoint Analysis

Appendix for Causal Interaction in Factorial Experiments: Application to Conjoint Analysis A Appendx for Causal Interacton n Factoral Experments: Applcaton to Conjont Analyss Mathematcal Appendx: Proofs of Theorems A. Lemmas Below, we descrbe all the lemmas, whch are used to prove the man theorems

More information

Continuous Implementation

Continuous Implementation Contnuous Implementaton Maron Oury HEC Pars Olver Terceux Pars School of Economcs and CNRS Abstract It s well-known that mechansm desgn lterature makes many smplfyng nformatonal assumptons n partcular

More information

The Second Anti-Mathima on Game Theory

The Second Anti-Mathima on Game Theory The Second Ant-Mathma on Game Theory Ath. Kehagas December 1 2006 1 Introducton In ths note we wll examne the noton of game equlbrum for three types of games 1. 2-player 2-acton zero-sum games 2. 2-player

More information

The Order Relation and Trace Inequalities for. Hermitian Operators

The Order Relation and Trace Inequalities for. Hermitian Operators Internatonal Mathematcal Forum, Vol 3, 08, no, 507-57 HIKARI Ltd, wwwm-hkarcom https://doorg/0988/mf088055 The Order Relaton and Trace Inequaltes for Hermtan Operators Y Huang School of Informaton Scence

More information

Affine transformations and convexity

Affine transformations and convexity Affne transformatons and convexty The purpose of ths document s to prove some basc propertes of affne transformatons nvolvng convex sets. Here are a few onlne references for background nformaton: http://math.ucr.edu/

More information

Bounded Reasoning and Higher-Order Uncertainty

Bounded Reasoning and Higher-Order Uncertainty Bounded Reasonng and Hgher-Order Uncertanty Wllemen Kets February 1, 2012 Abstract Standard models of games wth ncomplete nformaton assume that players form belefs about ther opponents belefs about ther

More information

Topologies on Types: Connections

Topologies on Types: Connections Topologes on Types: Connectons Y-Chun Chen y Syang Xong z May 28, 2008 Abstract For d erent purposes, economsts may use d erent topologes on types. We characterze the relatonshp among these varous topologes.

More information

Tit-For-Tat Equilibria in Discounted Repeated Games with. Private Monitoring

Tit-For-Tat Equilibria in Discounted Repeated Games with. Private Monitoring 1 Tt-For-Tat Equlbra n Dscounted Repeated Games wth Prvate Montorng Htosh Matsushma 1 Department of Economcs, Unversty of Tokyo 2 Aprl 24, 2007 Abstract We nvestgate nfntely repeated games wth mperfect

More information

Perfect Competition and the Nash Bargaining Solution

Perfect Competition and the Nash Bargaining Solution Perfect Competton and the Nash Barganng Soluton Renhard John Department of Economcs Unversty of Bonn Adenauerallee 24-42 53113 Bonn, Germany emal: rohn@un-bonn.de May 2005 Abstract For a lnear exchange

More information

Epistemic Game Theory: Beliefs and Types

Epistemic Game Theory: Beliefs and Types Epstemc Game Theory: Belefs and Types Marcano Snscalch March 28, 2007 1 Introducton John Harsany [19] ntroduced the formalsm of type spaces to provde a smple and parsmonous representaton of belef herarches.

More information

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction ECONOMICS 5* -- NOTE (Summary) ECON 5* -- NOTE The Multple Classcal Lnear Regresson Model (CLRM): Specfcaton and Assumptons. Introducton CLRM stands for the Classcal Lnear Regresson Model. The CLRM s also

More information

Incomplete Information and Robustness in Strategic Environments

Incomplete Information and Robustness in Strategic Environments Unversty of Pennsylvana ScholarlyCommons Publcly Accessble Penn Dssertatons Sprng 5-17-2010 Incomplete Informaton and Robustness n Strategc Envronments Antono Penta Unversty of Pennsylvana, apenta@ssc.wsc.edu

More information

Journal of Mathematical Economics

Journal of Mathematical Economics Journal of Mathematcal Economcs 46 (2010) 393 399 Contents lsts avalable at ScenceDrect Journal of Mathematcal Economcs journal homepage: www.elsever.com/locate/jmateco A note on the evaluaton of nformaton

More information

Axiomatizations of Pareto Equilibria in Multicriteria Games

Axiomatizations of Pareto Equilibria in Multicriteria Games ames and Economc Behavor 28, 146154 1999. Artcle ID game.1998.0680, avalable onlne at http:www.dealbrary.com on Axomatzatons of Pareto Equlbra n Multcrtera ames Mark Voorneveld,* Dres Vermeulen, and Peter

More information

APPENDIX A Some Linear Algebra

APPENDIX A Some Linear Algebra APPENDIX A Some Lnear Algebra The collecton of m, n matrces A.1 Matrces a 1,1,..., a 1,n A = a m,1,..., a m,n wth real elements a,j s denoted by R m,n. If n = 1 then A s called a column vector. Smlarly,

More information

Lecture 3: Probability Distributions

Lecture 3: Probability Distributions Lecture 3: Probablty Dstrbutons Random Varables Let us begn by defnng a sample space as a set of outcomes from an experment. We denote ths by S. A random varable s a functon whch maps outcomes nto the

More information

Robust Implementation: The Role of Large Type Spaces

Robust Implementation: The Role of Large Type Spaces Robust Implementaton: The Role of Large Type Spaces Drk Bergemann y Stephen Morrs z Frst Verson: March 2003 Ths Verson: Aprl 2004 Abstract We analyze the problem of fully mplementng a socal choce functon

More information

Power law and dimension of the maximum value for belief distribution with the max Deng entropy

Power law and dimension of the maximum value for belief distribution with the max Deng entropy Power law and dmenson of the maxmum value for belef dstrbuton wth the max Deng entropy Bngy Kang a, a College of Informaton Engneerng, Northwest A&F Unversty, Yanglng, Shaanx, 712100, Chna. Abstract Deng

More information

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur Module 3 LOSSY IMAGE COMPRESSION SYSTEMS Verson ECE IIT, Kharagpur Lesson 6 Theory of Quantzaton Verson ECE IIT, Kharagpur Instructonal Objectves At the end of ths lesson, the students should be able to:

More information

Constant Best-Response Functions: Interpreting Cournot

Constant Best-Response Functions: Interpreting Cournot Internatonal Journal of Busness and Economcs, 009, Vol. 8, No., -6 Constant Best-Response Functons: Interpretng Cournot Zvan Forshner Department of Economcs, Unversty of Hafa, Israel Oz Shy * Research

More information

(1 ) (1 ) 0 (1 ) (1 ) 0

(1 ) (1 ) 0 (1 ) (1 ) 0 Appendx A Appendx A contans proofs for resubmsson "Contractng Informaton Securty n the Presence of Double oral Hazard" Proof of Lemma 1: Assume that, to the contrary, BS efforts are achevable under a blateral

More information

Maximizing the number of nonnegative subsets

Maximizing the number of nonnegative subsets Maxmzng the number of nonnegatve subsets Noga Alon Hao Huang December 1, 213 Abstract Gven a set of n real numbers, f the sum of elements of every subset of sze larger than k s negatve, what s the maxmum

More information

Payoff Information and. Self-Confirming Equilibrium 1

Payoff Information and. Self-Confirming Equilibrium 1 Payoff Informaton and Self-Confrmng Equlbrum 1 Frst verson: Aprl 25, 1995 Ths revson: July 12, 1999 Edde Deel Drew Fudenberg Davd K. Levne 2 1 We than Faru Gul, an assocate edtor, two referees, and semnar

More information

Cowles Foundation for Research in Economics at Yale University

Cowles Foundation for Research in Economics at Yale University Cowles Foundaton for Research n Economcs at Yale Unversty Cowles Foundaton Dscusson Paper No. 666 ROBUST IMPLEMENTATION IN GENERAL MECHANISMS Drk Bergemann and Stephen Morrs Month 28 An author ndex to

More information

12 MATH 101A: ALGEBRA I, PART C: MULTILINEAR ALGEBRA. 4. Tensor product

12 MATH 101A: ALGEBRA I, PART C: MULTILINEAR ALGEBRA. 4. Tensor product 12 MATH 101A: ALGEBRA I, PART C: MULTILINEAR ALGEBRA Here s an outlne of what I dd: (1) categorcal defnton (2) constructon (3) lst of basc propertes (4) dstrbutve property (5) rght exactness (6) localzaton

More information

Game Theory. Lecture Notes By Y. Narahari. Department of Computer Science and Automation Indian Institute of Science Bangalore, India February 2008

Game Theory. Lecture Notes By Y. Narahari. Department of Computer Science and Automation Indian Institute of Science Bangalore, India February 2008 Game Theory Lecture Notes By Y. Narahar Department of Computer Scence and Automaton Indan Insttute of Scence Bangalore, Inda February 2008 Chapter 10: Two Person Zero Sum Games Note: Ths s a only a draft

More information

C.V. STARR CENTER FOR APPLIED ECONOMICS

C.V. STARR CENTER FOR APPLIED ECONOMICS ECONOMIC RESEARCH REPORTS The Truth s n the Eye of the Beholder: or Equlbrum n Belefs and Ratonal Learnng n Games by Yaw Nyarko RR# 98-12 Aprl 1998 C.V. STARR CENTER FOR APPLIED ECONOMICS NEW YORK UNIVERSITY

More information

Edge Isoperimetric Inequalities

Edge Isoperimetric Inequalities November 7, 2005 Ross M. Rchardson Edge Isopermetrc Inequaltes 1 Four Questons Recall that n the last lecture we looked at the problem of sopermetrc nequaltes n the hypercube, Q n. Our noton of boundary

More information

Assortment Optimization under MNL

Assortment Optimization under MNL Assortment Optmzaton under MNL Haotan Song Aprl 30, 2017 1 Introducton The assortment optmzaton problem ams to fnd the revenue-maxmzng assortment of products to offer when the prces of products are fxed.

More information

2E Pattern Recognition Solutions to Introduction to Pattern Recognition, Chapter 2: Bayesian pattern classification

2E Pattern Recognition Solutions to Introduction to Pattern Recognition, Chapter 2: Bayesian pattern classification E395 - Pattern Recognton Solutons to Introducton to Pattern Recognton, Chapter : Bayesan pattern classfcaton Preface Ths document s a soluton manual for selected exercses from Introducton to Pattern Recognton

More information

Bayesian epistemology II: Arguments for Probabilism

Bayesian epistemology II: Arguments for Probabilism Bayesan epstemology II: Arguments for Probablsm Rchard Pettgrew May 9, 2012 1 The model Represent an agent s credal state at a gven tme t by a credence functon c t : F [0, 1]. where F s the algebra of

More information

Linear, affine, and convex sets and hulls In the sequel, unless otherwise specified, X will denote a real vector space.

Linear, affine, and convex sets and hulls In the sequel, unless otherwise specified, X will denote a real vector space. Lnear, affne, and convex sets and hulls In the sequel, unless otherwse specfed, X wll denote a real vector space. Lnes and segments. Gven two ponts x, y X, we defne xy = {x + t(y x) : t R} = {(1 t)x +

More information

Group Analysis of Ordinary Differential Equations of the Order n>2

Group Analysis of Ordinary Differential Equations of the Order n>2 Symmetry n Nonlnear Mathematcal Physcs 997, V., 64 7. Group Analyss of Ordnary Dfferental Equatons of the Order n> L.M. BERKOVICH and S.Y. POPOV Samara State Unversty, 4430, Samara, Russa E-mal: berk@nfo.ssu.samara.ru

More information

Higher-Order Expectations *

Higher-Order Expectations * Hgher-Order Expectatons * Benjamn Golub Stephen Morrs August 31, 2017 Abstract We study hgher-order expectatons parallelng the Harsany (1968) approach to hgher-order belefs takng a basc set of random varables

More information

FACTORIZATION IN KRULL MONOIDS WITH INFINITE CLASS GROUP

FACTORIZATION IN KRULL MONOIDS WITH INFINITE CLASS GROUP C O L L O Q U I U M M A T H E M A T I C U M VOL. 80 1999 NO. 1 FACTORIZATION IN KRULL MONOIDS WITH INFINITE CLASS GROUP BY FLORIAN K A I N R A T H (GRAZ) Abstract. Let H be a Krull monod wth nfnte class

More information

SL n (F ) Equals its Own Derived Group

SL n (F ) Equals its Own Derived Group Internatonal Journal of Algebra, Vol. 2, 2008, no. 12, 585-594 SL n (F ) Equals ts Own Derved Group Jorge Macel BMCC-The Cty Unversty of New York, CUNY 199 Chambers street, New York, NY 10007, USA macel@cms.nyu.edu

More information

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number

More information

Implementation and Detection

Implementation and Detection 1 December 18 2014 Implementaton and Detecton Htosh Matsushma Department of Economcs Unversty of Tokyo 2 Ths paper consders mplementaton of scf: Mechansm Desgn wth Unqueness CP attempts to mplement scf

More information

2.3 Nilpotent endomorphisms

2.3 Nilpotent endomorphisms s a block dagonal matrx, wth A Mat dm U (C) In fact, we can assume that B = B 1 B k, wth B an ordered bass of U, and that A = [f U ] B, where f U : U U s the restrcton of f to U 40 23 Nlpotent endomorphsms

More information

Mechanisms with Evidence: Commitment and Robustness 1

Mechanisms with Evidence: Commitment and Robustness 1 Mechansms wth Evdence: Commtment and Robustness 1 Elchanan Ben-Porath 2 Edde Dekel 3 Barton L. Lpman 4 Frst Draft January 2017 Current Draft July 2018 1 We thank numerous semnar audences and Joel Sobel

More information

Bayesian predictive Configural Frequency Analysis

Bayesian predictive Configural Frequency Analysis Psychologcal Test and Assessment Modelng, Volume 54, 2012 (3), 285-292 Bayesan predctve Confgural Frequency Analyss Eduardo Gutérrez-Peña 1 Abstract Confgural Frequency Analyss s a method for cell-wse

More information

Numerical Heat and Mass Transfer

Numerical Heat and Mass Transfer Master degree n Mechancal Engneerng Numercal Heat and Mass Transfer 06-Fnte-Dfference Method (One-dmensonal, steady state heat conducton) Fausto Arpno f.arpno@uncas.t Introducton Why we use models and

More information

Graph Reconstruction by Permutations

Graph Reconstruction by Permutations Graph Reconstructon by Permutatons Perre Ille and Wllam Kocay* Insttut de Mathémathques de Lumny CNRS UMR 6206 163 avenue de Lumny, Case 907 13288 Marselle Cedex 9, France e-mal: lle@ml.unv-mrs.fr Computer

More information

COWLES FOUNDATION FOR RESEARCH IN ECONOMICS YALE UNIVERSITY

COWLES FOUNDATION FOR RESEARCH IN ECONOMICS YALE UNIVERSITY Best Response Equvalence by Stephen Morrs and Takash U July 2002 COWLES FOUNDATION DISCUSSION PAPER NO. 1377 COWLES FOUNDATION FOR RESEARCH IN ECONOMICS YALE UNIVERSITY Box 208281 New Haven, Connectcut

More information

A note on almost sure behavior of randomly weighted sums of φ-mixing random variables with φ-mixing weights

A note on almost sure behavior of randomly weighted sums of φ-mixing random variables with φ-mixing weights ACTA ET COMMENTATIONES UNIVERSITATIS TARTUENSIS DE MATHEMATICA Volume 7, Number 2, December 203 Avalable onlne at http://acutm.math.ut.ee A note on almost sure behavor of randomly weghted sums of φ-mxng

More information

Canonical transformations

Canonical transformations Canoncal transformatons November 23, 2014 Recall that we have defned a symplectc transformaton to be any lnear transformaton M A B leavng the symplectc form nvarant, Ω AB M A CM B DΩ CD Coordnate transformatons,

More information

Understanding Reasoning Using Utility Proportional Beliefs

Understanding Reasoning Using Utility Proportional Beliefs Understandng Reasonng Usng Utlty Proportonal Belefs Chrstan Nauerz EpCenter, Maastrcht Unversty c.nauerz@maastrchtunversty.nl Abstract. Tradtonally very lttle attenton has been pad to the reasonng process

More information

MATH 241B FUNCTIONAL ANALYSIS - NOTES EXAMPLES OF C ALGEBRAS

MATH 241B FUNCTIONAL ANALYSIS - NOTES EXAMPLES OF C ALGEBRAS MATH 241B FUNCTIONAL ANALYSIS - NOTES EXAMPLES OF C ALGEBRAS These are nformal notes whch cover some of the materal whch s not n the course book. The man purpose s to gve a number of nontrval examples

More information

A new construction of 3-separable matrices via an improved decoding of Macula s construction

A new construction of 3-separable matrices via an improved decoding of Macula s construction Dscrete Optmzaton 5 008 700 704 Contents lsts avalable at ScenceDrect Dscrete Optmzaton journal homepage: wwwelsevercom/locate/dsopt A new constructon of 3-separable matrces va an mproved decodng of Macula

More information

Infinitely Split Nash Equilibrium Problems in Repeated Games

Infinitely Split Nash Equilibrium Problems in Repeated Games Infntely Splt ash Equlbrum Problems n Repeated Games Jnlu L Department of Mathematcs Shawnee State Unversty Portsmouth, Oho 4566 USA Abstract In ths paper, we ntroduce the concept of nfntely splt ash equlbrum

More information

Beliefs, Plans, and Perceived Intentions in Dynamic Games.

Beliefs, Plans, and Perceived Intentions in Dynamic Games. Belefs, Plans, and Perceved Intentons n Dynamc Games. Perpaolo Battgall Department of Decson Scences and IGIER, Boccon Unversty perpaolo.battgall@unboccon.t Ncodemo De Vto Department of Decson Scences,

More information

ON THE EQUIVALENCE OF ORDINAL BAYESIAN INCENTIVE COMPATIBILITY AND DOMINANT STRATEGY INCENTIVE COMPATIBILITY FOR RANDOM RULES

ON THE EQUIVALENCE OF ORDINAL BAYESIAN INCENTIVE COMPATIBILITY AND DOMINANT STRATEGY INCENTIVE COMPATIBILITY FOR RANDOM RULES ON THE EQUIVALENCE OF ORDINAL BAYESIAN INCENTIVE COMPATIBILITY AND DOMINANT STRATEGY INCENTIVE COMPATIBILITY FOR RANDOM RULES Madhuparna Karmokar 1 and Souvk Roy 1 1 Economc Research Unt, Indan Statstcal

More information

PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 125, Number 7, July 1997, Pages 2119{2125 S (97) THE STRONG OPEN SET CONDITION

PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 125, Number 7, July 1997, Pages 2119{2125 S (97) THE STRONG OPEN SET CONDITION PROCDINGS OF TH AMRICAN MATHMATICAL SOCITY Volume 125, Number 7, July 1997, Pages 2119{2125 S 0002-9939(97)03816-1 TH STRONG OPN ST CONDITION IN TH RANDOM CAS NORBRT PATZSCHK (Communcated by Palle. T.

More information

20. Mon, Oct. 13 What we have done so far corresponds roughly to Chapters 2 & 3 of Lee. Now we turn to Chapter 4. The first idea is connectedness.

20. Mon, Oct. 13 What we have done so far corresponds roughly to Chapters 2 & 3 of Lee. Now we turn to Chapter 4. The first idea is connectedness. 20. Mon, Oct. 13 What we have done so far corresponds roughly to Chapters 2 & 3 of Lee. Now we turn to Chapter 4. The frst dea s connectedness. Essentally, we want to say that a space cannot be decomposed

More information

Better Outcomes from More Rationality

Better Outcomes from More Rationality Better Outcomes from More Ratonalty Jng Chen CS Dept., Stony Brook Unversty Stony Brook, NY 794 jngchen@cs.stonybrook.edu Rafael Pass Dept. of Computer Scence Cornell Unversty rafael@cs.cornell.edu Slvo

More information

Computing Correlated Equilibria in Multi-Player Games

Computing Correlated Equilibria in Multi-Player Games Computng Correlated Equlbra n Mult-Player Games Chrstos H. Papadmtrou Presented by Zhanxang Huang December 7th, 2005 1 The Author Dr. Chrstos H. Papadmtrou CS professor at UC Berkley (taught at Harvard,

More information

Higher Order Beliefs in Dynamic Envinronments.

Higher Order Beliefs in Dynamic Envinronments. Hgher Order Belefs n Dynamc Envnronments. Antono Penta y Dept. of Economcs, Unversty of Pennsylvana May 2008 (prelmnary and ncomplete verson) Abstract Ths paper explores the role of hgher order belefs

More information

Robust Implementation: The Role of Large Type Spaces

Robust Implementation: The Role of Large Type Spaces Robust Implementaton: The Role of Large Type Spaces Drk Bergemann y Stephen Morrs z Frst Verson: March 23 Ths Verson: June 25 Abstract A socal choce functon s robustly mplemented f every equlbrum on every

More information

Backward Induction Reasoning in Games with Incomplete Information

Backward Induction Reasoning in Games with Incomplete Information Backward Inducton Reasonng n Games wth Incomplete Informaton Antono Penta y Unversty of Wsconsn - Madson, Dept. of Economcs Ths verson: September 26, 2011 Abstract Backward Inducton s one of the central

More information

Perron Vectors of an Irreducible Nonnegative Interval Matrix

Perron Vectors of an Irreducible Nonnegative Interval Matrix Perron Vectors of an Irreducble Nonnegatve Interval Matrx Jr Rohn August 4 2005 Abstract As s well known an rreducble nonnegatve matrx possesses a unquely determned Perron vector. As the man result of

More information

Foundations of Arithmetic

Foundations of Arithmetic Foundatons of Arthmetc Notaton We shall denote the sum and product of numbers n the usual notaton as a 2 + a 2 + a 3 + + a = a, a 1 a 2 a 3 a = a The notaton a b means a dvdes b,.e. ac = b where c s an

More information

Representation theory and quantum mechanics tutorial Representation theory and quantum conservation laws

Representation theory and quantum mechanics tutorial Representation theory and quantum conservation laws Representaton theory and quantum mechancs tutoral Representaton theory and quantum conservaton laws Justn Campbell August 1, 2017 1 Generaltes on representaton theory 1.1 Let G GL m (R) be a real algebrac

More information

Online Appendix. t=1 (p t w)q t. Then the first order condition shows that

Online Appendix. t=1 (p t w)q t. Then the first order condition shows that Artcle forthcomng to ; manuscrpt no (Please, provde the manuscrpt number!) 1 Onlne Appendx Appendx E: Proofs Proof of Proposton 1 Frst we derve the equlbrum when the manufacturer does not vertcally ntegrate

More information

A SURVEY OF PROPERTIES OF FINITE HORIZON DIFFERENTIAL GAMES UNDER ISAACS CONDITION. Contents

A SURVEY OF PROPERTIES OF FINITE HORIZON DIFFERENTIAL GAMES UNDER ISAACS CONDITION. Contents A SURVEY OF PROPERTIES OF FINITE HORIZON DIFFERENTIAL GAMES UNDER ISAACS CONDITION BOTAO WU Abstract. In ths paper, we attempt to answer the followng questons about dfferental games: 1) when does a two-player,

More information

Estimation: Part 2. Chapter GREG estimation

Estimation: Part 2. Chapter GREG estimation Chapter 9 Estmaton: Part 2 9. GREG estmaton In Chapter 8, we have seen that the regresson estmator s an effcent estmator when there s a lnear relatonshp between y and x. In ths chapter, we generalzed the

More information

princeton univ. F 17 cos 521: Advanced Algorithm Design Lecture 7: LP Duality Lecturer: Matt Weinberg

princeton univ. F 17 cos 521: Advanced Algorithm Design Lecture 7: LP Duality Lecturer: Matt Weinberg prnceton unv. F 17 cos 521: Advanced Algorthm Desgn Lecture 7: LP Dualty Lecturer: Matt Wenberg Scrbe: LP Dualty s an extremely useful tool for analyzng structural propertes of lnear programs. Whle there

More information

The Myerson value in terms of the link agent form: a technical note

The Myerson value in terms of the link agent form: a technical note The Myerson value n terms of the lnk agent form: a techncal note André Casajus (September 2008, ths verson: October 1, 2008, 18:16) Abstract We represent the Myerson (1977) value n terms of the value ntroduced

More information

Difference Equations

Difference Equations Dfference Equatons c Jan Vrbk 1 Bascs Suppose a sequence of numbers, say a 0,a 1,a,a 3,... s defned by a certan general relatonshp between, say, three consecutve values of the sequence, e.g. a + +3a +1

More information

Module 17: Mechanism Design & Optimal Auctions

Module 17: Mechanism Design & Optimal Auctions Module 7: Mechansm Desgn & Optmal Auctons Informaton Economcs (Ec 55) George Georgads Examples: Auctons Blateral trade Producton and dstrbuton n socety General Setup N agents Each agent has prvate nformaton

More information

Week 2. This week, we covered operations on sets and cardinality.

Week 2. This week, we covered operations on sets and cardinality. Week 2 Ths week, we covered operatons on sets and cardnalty. Defnton 0.1 (Correspondence). A correspondence between two sets A and B s a set S contaned n A B = {(a, b) a A, b B}. A correspondence from

More information

= z 20 z n. (k 20) + 4 z k = 4

= z 20 z n. (k 20) + 4 z k = 4 Problem Set #7 solutons 7.2.. (a Fnd the coeffcent of z k n (z + z 5 + z 6 + z 7 + 5, k 20. We use the known seres expanson ( n+l ( z l l z n below: (z + z 5 + z 6 + z 7 + 5 (z 5 ( + z + z 2 + z + 5 5

More information

Mechanisms with Evidence: Commitment and Robustness 1

Mechanisms with Evidence: Commitment and Robustness 1 Mechansms wth Evdence: Commtment and Robustness 1 Elchanan Ben-Porath 2 Edde Dekel 3 Barton L. Lpman 4 Frst Draft January 2017 1 We thank the Natonal Scence Foundaton, grant SES 0820333 (Dekel), and the

More information

Online Appendix: Reciprocity with Many Goods

Online Appendix: Reciprocity with Many Goods T D T A : O A Kyle Bagwell Stanford Unversty and NBER Robert W. Stager Dartmouth College and NBER March 2016 Abstract Ths onlne Appendx extends to a many-good settng the man features of recprocty emphaszed

More information

ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EQUATION

ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EQUATION Advanced Mathematcal Models & Applcatons Vol.3, No.3, 2018, pp.215-222 ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EUATION

More information

THE CHVÁTAL-ERDŐS CONDITION AND 2-FACTORS WITH A SPECIFIED NUMBER OF COMPONENTS

THE CHVÁTAL-ERDŐS CONDITION AND 2-FACTORS WITH A SPECIFIED NUMBER OF COMPONENTS Dscussones Mathematcae Graph Theory 27 (2007) 401 407 THE CHVÁTAL-ERDŐS CONDITION AND 2-FACTORS WITH A SPECIFIED NUMBER OF COMPONENTS Guantao Chen Department of Mathematcs and Statstcs Georga State Unversty,

More information

10. Canonical Transformations Michael Fowler

10. Canonical Transformations Michael Fowler 10. Canoncal Transformatons Mchael Fowler Pont Transformatons It s clear that Lagrange s equatons are correct for any reasonable choce of parameters labelng the system confguraton. Let s call our frst

More information

DIFFERENTIAL FORMS BRIAN OSSERMAN

DIFFERENTIAL FORMS BRIAN OSSERMAN DIFFERENTIAL FORMS BRIAN OSSERMAN Dfferentals are an mportant topc n algebrac geometry, allowng the use of some classcal geometrc arguments n the context of varetes over any feld. We wll use them to defne

More information

THE CHINESE REMAINDER THEOREM. We should thank the Chinese for their wonderful remainder theorem. Glenn Stevens

THE CHINESE REMAINDER THEOREM. We should thank the Chinese for their wonderful remainder theorem. Glenn Stevens THE CHINESE REMAINDER THEOREM KEITH CONRAD We should thank the Chnese for ther wonderful remander theorem. Glenn Stevens 1. Introducton The Chnese remander theorem says we can unquely solve any par of

More information

Continuous Time Markov Chain

Continuous Time Markov Chain Contnuous Tme Markov Chan Hu Jn Department of Electroncs and Communcaton Engneerng Hanyang Unversty ERICA Campus Contents Contnuous tme Markov Chan (CTMC) Propertes of sojourn tme Relatons Transton probablty

More information

Lecture 12: Discrete Laplacian

Lecture 12: Discrete Laplacian Lecture 12: Dscrete Laplacan Scrbe: Tanye Lu Our goal s to come up wth a dscrete verson of Laplacan operator for trangulated surfaces, so that we can use t n practce to solve related problems We are mostly

More information

COWLES FOUNDATION FOR RESEARCH IN ECONOMICS YALE UNIVERSITY

COWLES FOUNDATION FOR RESEARCH IN ECONOMICS YALE UNIVERSITY ROBUST MECHANISM DESIGN By Drk Bergemann and Stephen Morrs Aprl 2004 COWLES FOUNDATION DISCUSSION PAPER NO. 1421R COWLES FOUNDATION FOR RESEARCH IN ECONOMICS YALE UNIVERSITY Box 208281 New Haven, Connectcut

More information

The Expectation-Maximization Algorithm

The Expectation-Maximization Algorithm The Expectaton-Maxmaton Algorthm Charles Elan elan@cs.ucsd.edu November 16, 2007 Ths chapter explans the EM algorthm at multple levels of generalty. Secton 1 gves the standard hgh-level verson of the algorthm.

More information