REVIEW 3 major objections 6 minor 19 references
Improving stabilizer approximation with quantum strategy
T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read Borrowing the CHSH game's quantum strategy gives a qubit-by-qubit gauging procedure that improves stabilizer approximation of ground-state energies.
desk verdict The discrete examples are correct, but the paper's advertised general improvement is false: for H = -X1X2 - Z1 - Z2 the gauging cannot beat the stabilizer energy -2. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the single-qubit rotation $R_y(\theta) = \cos(\theta/2)I - i\sin(\theta/2)Y$ and its action by conjugation on the Pauli pair $\{X,Z\}$. At $\theta = \pi/4$ it produces exactly the CHSH measurements $X' = (X+Z)/\sqrt{2}$ and $Z' = (Z-X)/\sqrt{2}$. The paper treats this rotation as a gauge choice: rewrite the Hamiltonian in the rotated basis, keep one rotated Pauli per qubit as a stabilizer, discard the other, and proceed. The sequential version repeats this qubit by qubit, reducing the Hamiltonian after each step, and the continuous version lets the angle vary. The machinery carries the argument because every energy improvement in the paper comes from choosing these rotated Pauli axes rather than the original $X$/$Z$ axes.
What would settle it
For the intermediate hydrogen Hamiltonian $H''_{H_2} = I_1\otimes Z_2 - Z_1\otimes I_2 + 2X_1\otimes X_2$, compute the minimal energy reachable by sequential gauging with arbitrary rotation angles $\theta_1,\theta_2$. If that minimum is above $-2.5$, the paper's continuous-gauging claim is false; a direct scan or analytic minimization over the two angles would settle it.
Extended reading notes
Core claim
On its own terms, the paper claims that the optimal quantum strategy for the CHSH game is not merely a game-theoretic curiosity but a template for improving stabilizer approximations. The strategy's measurements on the second player, $B_0 = (X+Z)/\sqrt{2}$ and $B_1 = (Z-X)/\sqrt{2}$, are exactly what you get by conjugating $X$ and $Z$ with $R_y(\pi/4)$. Using this as a gauge transformation, a Hamiltonian initially written in $X$ and $Z$ can be rewritten in the rotated Paulis $X'$ and $Z'$, and a commuting subset of the rotated terms then stabilizes a lower-energy state. Applying the rotation qubit by qubit yields the paper's sequential gauging procedure, which improves the hydrogen-molecule energy from $-2$ to $-1-\sqrt{2}$, and with continuous angles to $-2.5$, compared with the exact $-2\sqrt{2}$.
Load-bearing premise
The method rests on assuming that locally rotating Pauli axes qubit by qubit and discarding one rotated Pauli at each step leaves a stabilizer subspace that still contains a good approximation to the true ground state, and that suitable continuous gauge angles exist to reach the claimed minus 2.5; the paper demonstrates this on two small examples but gives no general proof or explicit angles.
Editorial extensions
If this is right
- Stabilizer approximation gains a cheap preprocessing step: rotate each qubit's Pauli axes before selecting the commuting stabilizer subset, with no change to the Hamiltonian's size.
- For the CHSH Hamiltonian and for the equal-field Ising chain $g_x = g_z, J = 0$, the gauged stabilizer state achieves the exact ground-state energy.
- For the intermediate hydrogen-molecule Hamiltonian, sequential gauging lowers the approximate energy from $-2$ to $-1-\sqrt{2}$, and continuous gauging to $-2.5$, bringing stabilizer-based initial states closer to the exact $-2\sqrt{2}$.
- The procedure generalizes a fixed $\pi/4$ rotation to continuous angles, so the same framework covers both discrete and continuous gauge choices.
Reading between the lines
- Beyond the paper, the same gauging idea should transfer to other nonlocal games whose optimal strategies use different local measurement axes, which could cover Hamiltonians with competing terms not aligned with $X$ and $Z$.
- The sequential procedure is effectively a classical preprocessing algorithm: at each step it fixes one stabilizer and reduces the qubit count, suggesting a deterministic cost that scales with the number of qubits; the paper does not state this complexity explicitly.
- A direct testable extension is a variational search over the continuous gauge angles on small molecules, comparing the resulting energies against standard stabilizer approximations and variational baselines; the paper leaves the angle-finding rule unspecified.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes a heuristic to improve stabilizer approximation by borrowing the quantum strategy of the CHSH game. It shows that for the CHSH Hamiltonian of Eq. (8), rotating X and Z by Ry(pi/4) rewrites the Hamiltonian as in Eq. (14) and lowers the two-qubit energy from -2 to -2*sqrt(2). The same idea is then applied to a non-interacting Ising model and to an intermediate H2 Hamiltonian, where sequential gauging lowers the energy from -2 to -1-sqrt(2), and a continuous gauge parameter is claimed to lower it further to -2.5. The abstract and Section V generalize the procedure to a qubit-by-qubit gauging method that 'significantly improves' stabilizer approximation whenever the original approximation deteriorates.
Significance. If the general claim were valid, the procedure would be a useful, inexpensive preprocessing step for stabilizer-based ground-state initialization. The CHSH two-qubit calculation is elegant and correct, and the Ising and H2 examples are explicit and reproducible. The paper is not, however, a proof or even a precise conjecture of a general improvement: it is a collection of examples plus an unproved and, as shown below, false generalization. The main value of the note lies in the observation that local rotations can convert certain two-qubit sign patterns into anti-commuting pairs that are better approximated by product states; that observation should be stated as a restricted variational claim rather than as a general stabilizer-improvement theorem.
major comments (3)
- [Abstract; §V; §I] The claim that the procedure 'significantly improves the performance when the original approximation deteriorates' is false as stated. Consider H = -X1X2 - Z1 - Z2, where the non-commuting terms have equal coefficients, exactly the regime motivating Section I. The best standard stabilizer choice, -Z1 and -Z2, gives energy -2, while the exact ground energy is -sqrt(5) ~ -2.236. For every product state |r>|s> the expectation value is -r_x s_x - r_z - s_z >= -2, so no local Y-rotation and no sequential or continuous gauging of the type described in Sections IV.A-C can lower the energy below -2. Since the procedure terminates in a product state, this counterexample directly invalidates the general improvement claim. Please either restrict the claim to Hamiltonians with the CHSH sign structure of Eq. (8) and characterize that structure, or present the method as a product-state variational ansatz with explicit conditions under which it improves on a given stabilizer choice.
- [§IV.C] The central quantitative claim of Section IV.C, that 'properly choosing the gauge parameters' lowers the H2 energy from -1-sqrt(2) to -2.5, is asserted without any derivation. Equations (25)-(26) define the rotated operators, but no values of theta_1 and theta_2, no intermediate expectation values, and no optimization procedure are given. This number is load-bearing because it is the only evidence that continuous gauging improves on the sequential discrete result. Please supply the full calculation for the Hamiltonian in Eq. (22), including the optimized angles and the resulting state and energy, or remove the claim.
- [§IV.B; §IV.A-C] The paper does not state precisely what the gauging procedure is approximating. After a general rotation (25)-(26), the operators X' and Z' are not Pauli operators, and the state obtained by choosing, say, -X'_i as a 'stabilizer' is not a standard stabilizer state. In addition, each sequential step discards one rotated operator per qubit, so the final state is a product state; the paper should acknowledge this explicitly and explain why the discarded terms can be set to zero without an error that could increase the energy. If the method is intended as a variational product-state ansatz with Y-rotations, that should be stated, and the relation to existing product-state or mean-field initialization methods should be discussed.
minor comments (6)
- [§I; Ref. [13]] The name 'Clause-Horne-Shimony-Holt' should be 'Clauser-Horne-Shimony-Holt'.
- [§III, Eq. (4)] The symbol H is used both for the Hadamard gate in Eq. (4) and for Hamiltonians throughout the paper; please disambiguate these two uses.
- [§III, Eq. (15)] The stabilizer group of the state in Eq. (9) is not stated; the reader must verify that X tensor X and -Z tensor Z stabilize the state, and this would be clearer if the generators were listed explicitly.
- [§IV.A] The notation H'_Ising in Eq. (17) is confusing because the prime elsewhere denotes rotated operators; please use a different label, for example H_I(0).
- [Acknowledgments] The acknowledgments contain typos and informal phrases ('homepape', 'I don't know', 'I guess') that are out of place in a journal submission; please polish the entire manuscript for style.
- [Ref. [3]] The arXiv identifier '2209.095643' appears to have an extra digit; please verify it.
Circularity Check
No circularity: the stabilizer-gauging examples are explicit variational energy computations anchored in the external CHSH strategy, not reductions to the paper's own inputs.
full rationale
The paper's derivation chain is self-contained. The CHSH quantum strategy in Eqs. (4)-(7) is a standard external result; Eq. (8) is an explicit Hamiltonian encoding of the game; Eqs. (10)-(13) define a rotation whose algebraic effect is verified in Eq. (14). Choosing the rotated stabilizers X⊗X' and -Z⊗Z' gives energy -2√2, which is the exact ground energy of H0, so this example is a direct computation rather than a fitted prediction. The Ising and H2 examples are likewise explicit energy evaluations. For the intermediate H2 Hamiltonian (22), sequential gauging yields Eqs. (23)-(24) and the stated energy -1-√2. The continuous-gauging statement that the energy can be lowered to -2.5 is not written out in the text, but it is a reproducible variational calculation: with θ1=θ2=π/6, the general gauged energy is -sinθ1 - sinθ2 - 2cosθ1cosθ2 = -2.5. Thus the missing derivation is an omission, not a circular step. The self-citations [3-5] identify the previously proposed stabilizer-approximation framework that is being improved, but they do not supply the load-bearing equations of this paper, and the numerical benchmarks are checked against exact ground-state energies. The broad summary claim that the procedure 'significantly improves the performance when the original approximation deteriorates' may overgeneralize, but that is a correctness or scope concern, not circularity.
Assumptions & free parameters
free parameters (2)
- continuous gauge parameters θ_i =
unspecified (claimed to achieve -2.5 for H2)
- discrete sign choices for stabilizers =
-X'_1, -X'_2 in H2 example
assumptions (5)
- standard math Local unitary rotations preserve the spectrum and tensor-product structure of the Hamiltonian terms.
- standard math The CHSH game's optimal bias is √2/2 and the specific measurements in Eq. (4) achieve it.
- domain assumption The stabilizer approximation, selecting a commuting subset of Pauli terms, is a valid starting point for approximating ground states.
- ad hoc to paper Sequential gauging, i.e., choosing one rotated Pauli per qubit and discarding the other, yields an improved state.
- ad hoc to paper There exist continuous gauge parameters giving energy -2.5 for H2.
Cite this review
Pith. "Pith review of Improving stabilizer approximation with quantum strategy." pith.science (2026). https://pith.science/paper/2LKLLYXC
@misc{pith2026241206320,
author = {Pith},
title = {Pith review of: Improving stabilizer approximation with quantum strategy},
year = {2026},
howpublished = {\url{https://pith.science/paper/2LKLLYXC}},
note = {Machine review of arXiv:2412.06320}
}
read the original abstract
We introduce a quantum strategy from nonlocal games to improve the stabilizer approximation we proposed previously. The resulting approach turns out to be a qubit-by-qubit gauging procedure for standard stabilizers, which could involve discrete or continuous gauge parameters. We take examples from many-body physics and quantum chemistry to show such a procedure leads to an improvement of the performance.
Reference graph
Works this paper leans on
-
[15]
John F. Clauser, Michael A. Horne, Abner Shimony, Richa rd A. Holt, 1969, P roposed Experiment to Test Local Hidden-V ariable Theories,Phys. Rev. Lett. 23, 880
work page 1969
-
[1]
A. Y u. Kitaev, 1997, Fault-tolerant quantum computatio n by anyons, Annals Phys. 303 (2003) 2-30, [arXiv: quant-ph/9707021]
arXiv 2003
-
[2]
The corresponding new stabilizer state is simply [15] |Ψ ′⟩ = I ⊗ R† y( π 4 )|Ψ ⟩
In fact, this is the lowest eigenvalue of H0. The corresponding new stabilizer state is simply [15] |Ψ ′⟩ = I ⊗ R† y( π 4 )|Ψ ⟩. (15) It is not difficult to check that X ⊗ X ′ and −Z ⊗ Z ′ indeed stabilize |Ψ ′⟩. Therefore, the quantum strategy used in the CHSH game helps us improve the stabilize r approximation for the ground state energy. The essential p...
-
[3]
A brief in- troduction to quantum PCP conjecture
(24) Choosing −X ′ 2 as the second stabilizer and discarding Z ′ 2, we obtain an energy −1− √ 2 ≈ −2. 414, which is significantly lower than the naive stabilizer value −2. Still, this is higher than the lowest eigenvalue of H ′′ H2, which is −2 √ 2 ≈ −2. 828. 6 C. Continuous gauging All the previous steps could be generalized by introducing a gauging param...
-
[4]
Daniel Gottesman, 1996, Class of quantum error-correct ing codes saturating the quantum Hamming bound, Phys. Rev. A54 (1996) 1862, [arXiv: quant-ph/9604038]
arXiv 1996
-
[5]
Xinying Li, Jianan Wang, Chuixiong Wu, Fen Zuo, 2022, Sta bilizer Approximation, [arXiv: 2209.095643]
work page Pith review arXiv 2022
-
[6]
Jianan Wang, Chuixiong Wu, Fen Zuo, 2023, Stabilizer App roximation II: From H 2O To C 6H6, [arXiv: 2302.11734]
work page Pith review arXiv 2023
-
[7]
Chuixiong Wu, Jianan Wang, Fen Zuo, 2023, Stabilizer App roximation III: Maximum Cut, [arXiv: 2303.17215]
work page Pith review arXiv 2023
Show all 19 references
-
[8]
Zhong-Xia Shang, Ming-Cheng Chen, Xiao Y uan, Chao-Yang Lu and Jian-Wei Pan, 2021, Schrödinger-Heisenberg V ariational Quantum Algorithms, Phys. Rev. Lett. 131, 060406, [arXiv: 2112.07881]
2021
-
[9]
Kirby, Kaitlin N
Gokul Subramanian Ravi, Pranav Gokhale, Yi Ding, Willia m M. Kirby, Kaitlin N. Smith, Jonathan M. Baker, Peter J. Love, Henry Hoffmann, Kenneth R. Brown, and F rederic T. Chong, 2022, CAFQA: A classical simulation bootstrap for variational quantum al gorithms, In Proceedings o...
2022
-
[10]
M. H. Cheng, K. E. Khosla, C. N. Self, M. Lin, B. X. Li, A. C. M edina, M. S. Kim, 2022, Clifford Circuit Initialisation for V ariational Quantum Algorithms, [arXiv: 2207.01539]
2022 arXiv
-
[11]
Jiace Sun, Lixue Cheng, Shi-Xin Zhang, 2024, Stabilizer ground states for simulating quantum many- body physics: theory, algorithms, and applications, [arXi v: 2403.08441]
2024 arXiv
-
[12]
Brown, 2024, Stabilizer confi guration interaction: Finding molecular subspaces with error detection properties, [arXiv: 2410.2 1125]
Abhinav Anand, Kenneth R. Brown, 2024, Stabilizer confi guration interaction: Finding molecular subspaces with error detection properties, [arXiv: 2410.2 1125]
2024
-
[13]
thesis , California Institute of Technology, Pasadena, CA, 1997, [arXiv: quant -ph/9705052]
Daniel Gottesman, 1997, Stabilizer Codes and Quantum E rror Correction, Ph.D. thesis , California Institute of Technology, Pasadena, CA, 1997, [arXiv: quant -ph/9705052]
1997
-
[14]
Daniel Gottesman, 1998, The Heisenberg Representatio n of Quantum Computers, [https://arxiv.org/abs/quant-ph/9807006]. 8
1998 arXiv
-
[16]
Thomas Vidick, 2018, Quantum multiplayer games, testi ng and rigidity, Lecture notes of Spring school on Quantum Computation , March 19-22, 2018, University of California, San Diego
2018
-
[17]
Zhengfeng ji, 2016, Classical verification of quantum p roofs, STOC ’16: Proceedings of the forty- eighth annual ACM symposium on Theory of Computing , 885-898
2016
-
[18]
Seeley, Martin J
Jacob T. Seeley, Martin J. Richard, Peter J. Love, 2012, The Bravyi-Kitaev transformation for quantum computation of electronic structure, J. Chem. Phys. 137, 22 4109 (2012). [arXiv: 1208.5986]
2012 arXiv
-
[19]
Bob Coecke and Aleks Kissinger, 2017, Picturing quantum processes: A first course in quantum theory and diagrammatic reasoning, Cambridge University Press. 9
2017
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.