The Setting
In parallel repetition, each player receives a string of questions and returns a string of answers. The upper tail measures the chance of winning a fraction of coordinates above the single-round entangled value.
A finite probability law on a finite set is a nonnegative weight function of total mass one.✦ For a function on with values in a real vector space, the expectation under is given by (1); for a finite index set , the same equation gives the product law of finite probability laws on the common finite set , with and for every . (1)
A finite two-player one-round game has finite nonempty question alphabets and answer alphabets , a question law on , and a Boolean acceptance predicate .✦ Write ; all logarithms are natural.
A positive operator-valued measure (POVM) on a finite-dimensional Hilbert space with outcome alphabet is a family of operators satisfying (2): (2) An entangled strategy consists of local Hilbert spaces of positive finite dimensions, a unit vector , and local POVMs and for each local question. In fixed orthonormal bases, the state coefficient matrix of an entangled strategy and the corresponding state vector are given by (3), where is the squared Hilbert-Schmidt norm: (3)
For local operators , the bipartite operator expectation and its real part are defined in (4). The same equation defines the Born probabilities , the joint measurement distribution for the local POVMs: (4) A single-round strategy for the game is an entangled strategy with question alphabets and answer alphabets . The winning probability of a single-round strategy for and the finite-dimensional entangled value of are defined by (5), with the supremum over all finite-dimensional single-round strategies for : (5)
Repeated Experiments
For an integer , a repeated strategy for coordinates of the game is an entangled strategy with question alphabets and answer alphabets .
A repeated quantum experiment on a finite nonempty coordinate set consists of a question law on , a normalized state coefficient matrix , string POVMs and , and a Boolean predicate on the single-coordinate questions and answers.✦ A question profile of a repeated quantum experiment is a function assigning a question pair to each coordinate; write for its two local strings.✦ A transcript of a repeated quantum experiment is a tuple consisting of a question profile and answer strings and .✦ The conditional answer law , the joint transcript law , and the coordinate winning predicate are defined in (6): (6) The repetition model associated to a repeated strategy for takes and uses ’s question law and predicate together with ’s coefficient matrix and local string POVMs.✦
The uniform law on assigns weight to each coordinate.✦ For a finite probability law on outcomes and Boolean winning predicates on , the number of wins , the winning fraction , and the event probability at a real threshold are defined by (7); here denotes an indicator: (7) For , write . For a real excess , the repeated threshold probability is the probability of the event in (8) when question pairs are sampled independently from and answers follow .✦ The entangled threshold value of the game is the supremum of over all finite-dimensional repeated strategies for coordinates: (8)
The Claim
Theorem 1. There is a universal constant , with admissible choice , such that the following holds. Let be a finite game as above with , let , and let be an integer. The bounds for every finite-dimensional repeated strategy and for the threshold value are (9): (9)✦ For a repeated quantum experiment of on , let . Under the hypotheses and , its real win-fraction threshold satisfies (10): (10)✦ Let be a repeated strategy of positive integer length , and let . Its associated repetition model satisfies (11): (11)✦
In the paper, the result reads: there is a universal such that, for every finite two-player one-round game with Boolean acceptance predicate and entangled value , every , every integer and every finite-dimensional joint strategy for repetitions, , with the same inequality for the supremum over those strategies; the question law may be correlated, have zero entries and have disconnected support.
What Is Checked
The claims (9), (10) and (11) are stated over finite games with nonempty alphabets and finite-dimensional normalized bipartite states, with local POVMs on complete question and answer strings and product question laws across coordinates.