Pith. sign in

REVIEW 3 major objections 4 minor 26 references

Baldereschi mean value points for three-dimensional Bravais lattices

T0 review · 3 major / 4 minor · reviewed 2026-08-04 · deepseek-v4-flash

Pith's one-line read This paper provides analytical formulas for the Baldereschi mean-value point of every three-dimensional Bravais lattice, giving quantum Monte Carlo simulations an optimal single twist to reduce finite-size errors.

desk verdict Useful and mostly correct reference table; the rhombohedral entries need independent checking before the 'all fourteen' claim is fully trusted. read the letter →

arxiv 2607.16974 v2 pith:QSWASEKD submitted 2026-07-18 cond-mat.mtrl-sci

classification cond-mat.mtrl-sci
keywords Baldereschipointmean-valueBravaislatticeBrillouinzonequantumMonteCarlotwistaveragingstarfunctionsfinite-sizeeffects
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper aims to supply the Baldereschi mean-value point—the single wavevector at which any smooth, symmetric function of wavevector stays closest to its Brillouin-zone average—for all fourteen three-dimensional Bravais lattices. For the orthorhombic, monoclinic, and triclinic families these formulas had not previously been reported. Because quantum Monte Carlo simulations can apply only one Bloch wavevector per simulation, this point is the natural choice for a twist, and the paper demonstrates that it reproduces the accuracy of full twist averaging in insulators while remaining more representative than the zone center in metals. The paper also corrects numerical results from a recent implementation of the same idea.

What carries the argument

The star functions A_n(k) = Σ_{R∈⋆n} exp(i k·R) form an orthogonal basis for symmetric functions of wavevector and carry the whole argument: the mean squared deviation of any smooth function from its average is a weighted sum of these star functions. The algorithm selects 'target stars'—the lowest-radius stars whose union spans R^3 and whose star functions can be varied independently—using equation (4), a sum-of-squares radius condition that detects whether a star carries a reducible representation of the lattice point group. The Baldereschi point is then the wavevector that zeroes the first two target star functions and either zeroes or minimizes the third. This reduces a problem about arbi

What would settle it

Compute the target stars for a rhombohedral lattice at a crossover angle (such as c/a = √6) using the paper's algorithm and compare the predicted Baldereschi point with an exhaustive numerical minimization of the mean squared deviation over an ensemble of rapidly convergent symmetric functions; any disagreement would show the radius test picks the wrong target stars.

Watch

Extended reading notes

Core claim

The paper's central claim is that the Baldereschi point is fixed entirely by lattice geometry and can be computed for every Bravais lattice by a systematic star-function procedure. The procedure identifies the lowest-radius 'target' stars whose union spans 3D space, tests each larger star for independent variability using the radius condition (2r_n)^2 = n_i^2 r_i^2 + n_j^2 r_j^2 + n_k^2 r_k^2, and then zeroes or minimizes the target star functions. The result is a complete table of Baldereschi points, with piecewise expressions where the optimal point jumps at certain lattice-parameter ratios (notably rhombohedral and body-centred tetragonal). Because the point depends only on the lattice, i

Load-bearing premise

Everything rests on the rule that a star of lattice points can be treated as independently variable unless its radius squared equals a sum of squared radii of smaller stars; the paper checks this rule numerically but does not prove it in full generality.

Editorial extensions

If this is right

  • Quantum Monte Carlo simulations on any of the fourteen Bravais lattices can use the exact Baldereschi twist, giving single-twist results that match twist-averaged accuracy in insulators while using a fraction of the computational cost.
  • The analytical formulas remove the need for numerical searches or Monte Carlo sampling of the Brillouin zone just to pick a twist; the choice is now a direct formula from the lattice constants.
  • For metallic systems, where twist averaging remains necessary, the Baldereschi point is the most representative single twist for wave-function optimization, reducing the sensitivity of the final twist-averaged result to the twist used in optimization.
  • The corrections to earlier numerical results establish a benchmark: any future implementation of mean-value-point algorithms should reproduce the table entries for all fourteen lattices.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the demonstrated effectiveness holds broadly, the Baldereschi point could allow QMC studies of insulators to skip twist averaging entirely, provided systematic finite-size corrections (such as the known Coulomb and correlation corrections) are applied separately.
  • The star-function selection algorithm could be transferred to two-dimensional and one-dimensional Bravais lattices, or extended to lattices with a basis, where the point group of the lattice is replaced by the space group; the same independent-variability test would apply.
  • The piecewise structure of the rhombohedral result implies that a continuous lattice distortion can cause the optimal twist to jump discontinuously; near these boundaries, twist choice should be examined with the actual lattice parameters rather than a generic formula.
  • A direct test of the paper's core claim would be to measure, for a range of materials across all lattice families, the difference between single-twist QMC energies at the Baldereschi point and the full twist-averaged energy; if the difference grows with Fourier-space slowness of the observable, the guarantee weakens for slowly convergent functions.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper defines and tabulates Baldereschi mean-value points for all fourteen three-dimensional Bravais lattices. The Baldereschi point is defined through an ensemble argument in which a symmetric function's Fourier coefficients decay exponentially with star index; the point is obtained by successively zeroing or minimizing the first few star functions A_n(k). The paper gives analytical results for the cubic, hexagonal, tetragonal, rhombohedral, orthorhombic, monoclinic, and triclinic families, extends earlier partial results, describes a target-star identification algorithm, and reports three applications (hBN DFT energies, graphene QMC, and the free-electron gas) supporting the practical usefulness of a single mean-value twist.

Significance. If correct, the tables provide a useful reference for QMC twist selection and single-point BZ sampling on all three-dimensional Bravais lattices. The paper is clearly written, and the star-function analysis for the high-symmetry rows is convincing; the inclusion of the numerical implementation in the casino distribution and the explicit comparisons with previous work are strengths. However, the central claim rests on the target-star selection algorithm, whose correctness is currently asserted rather than proved, and the rhombohedral crossover conditions contain an algebraic inconsistency. These issues are fixable but should be resolved before the tables are relied upon.

major comments (3)
  1. [Section 4, rhombohedral paragraph] The stated crossover conditions are inconsistent with the vectors in Table 1. For the obverse hexagonal cell a1=(a,0,0), a2=(-a/2,sqrt(3)a/2,0), a3=(a/2,sqrt(3)a/6,c/3), we have |a3|^2 = a^2/3 + c^2/9, so |a3|=|a1| indeed gives c/a=sqrt(6). But |2a3|^2 = 4a^2/3+4c^2/9, so |2a3|=|a1| has no solution with c>0; and |3a3-a1|^2 = a^2+c^2, so |3a3-a1|=|a1| would require c=0, not c/a=1. The same issue occurs for the reverse setting. Since this passage is the justification for the piecewise intervals in the rhombohedral rows of Table 1, the thresholds and the corresponding rhombohedral-cell entries need to be re-derived or corrected.
  2. [Appendix A and Section 3.2, Eq. (4)] The table entries depend on the target-star list, particularly in the rhombohedral case where the paper explicitly attributes disagreement with Ref. [8] to target-star selection. The test in Eq. (4) is only a test of squared radii; it is a necessary, not sufficient, condition for a star to be reducible into the target stars, and accidental radius coincidences are acknowledged. The paper's argument that wrong targets would produce nonunique solutions is empirical ('not observed in practice'), and the numerical check in Section 3.4 uses the same target-star algorithm, so it does not independently confirm the tables. Please add a proof or an independent verification of the target-star lists for at least the rhombohedral and low-symmetry lattices, for example by explicit irrep decomposition of low-radius stars or by direct numerical minimization of Eq. (2) without preselecting target stars.
  3. [Section 5] The validation examples test only the simple-cubic and hexagonal rows of the tables; the disputed rhombohedral entries and the new orthorhombic/monoclinic/triclinic tables are not exercised. Given that the central contribution is the full set of fourteen lattices, an independent check of the non-cubic, non-hexagonal rows would substantially strengthen confidence in the completeness of the tabulation.
minor comments (4)
  1. [Abstract] The abstract says 'every smooth periodic function' lies close to its mean value at the Baldereschi point. The derivation in Section 3.1 applies to an ensemble with exponentially decaying Fourier coefficients; the universal phrasing is stronger than what is proved. Please soften the wording.
  2. [Table 1] The values of A_t3(k_b) for the fcc, bcc, hexagonal, and rhombohedral rows are given as numbers such as 4.404, -3, and -1.608. Please state whether these are exact closed forms or numerical evaluations, and give the precision used.
  3. [Throughout] There are several typographical/encoding artifacts: 'T able 1' in the text, 'na¨ıve' for 'naive', and inconsistent italicization of 'casino'. These should be cleaned up.
  4. [Section 3.1] The statement that the star functions 'form an orthogonal basis' should specify the inner product (average over the Brillouin zone) and the domain of functions considered, for clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the table entries follow from the star-function definition and lattice geometry, with no fitted parameter renamed as a prediction.

full rationale

The derivation is self-contained: the Baldereschi point is defined by the star-function expansion (Eq. 1) and the exponentially weighted mean-square minimization (Eq. 2), and the target-star selection is a geometric test on lattice radii (Eq. 4). The tabulated points are analytical solutions of those equations; they are not obtained by fitting to the quantities used in the validation sections. In the graphene example the only fitted parameter (b in Eq. 5) is used for a twist-averaged control variate and is not used to set or adjust the Baldereschi point; the hBN and free-electron examples use the point as a fixed predictor. Self-citations ([12], [18], [19], [25]) relate to finite-size corrections, software, and Jastrow terms; none is load-bearing for the central lattice-geometry claim. The paper itself flags the main limitation in Sec. 3.2: Eq. (4) is relied on as a valid target-star test and the authors state that accidental satisfaction 'does not appear to be a problem in our numerical tests' and that incorrect target selection would give nonunique solutions, 'which are not observed in practice.' This is an unproved-lemma/verification-gap risk for the rhombohedral crossovers, particularly because the in-house numerical check uses the same target-star algorithm (Appendices A-B). That is a correctness caveat, not circularity: the manuscript does not define the Baldereschi point in terms of the table entries, nor does any predicted quantity reduce by construction to an input. Where prior results exist (cubic, hexagonal, tetragonal, bct), the paper reports agreement; its disagreements with [8], [9], and [15] are explicitly attributed to target-star selection or to a different (maximizing rather than minimizing) stationary point, not to circular fitting.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No free parameters are fitted. The central derivation rests on standard Fourier analysis and the Baldereschi definitional ensemble model, plus a heuristic target-star identification algorithm that is numerically verified.

assumptions (5)
  • standard math Every smooth periodic function f(k) with reciprocal-lattice symmetry has a Fourier expansion in star functions An(k), which form an orthogonal basis.
    Used in equation (1) to express f(k); this is standard Fourier analysis on the Brillouin zone.
  • domain assumption Physical observables are totally symmetric functions of k; only the totally symmetric part contributes to the mean value.
    Section 3.1; standard in BZ integration; justifies restricting to Γ1 symmetric functions.
  • domain assumption In insulators, expectation values are smooth functions of the twist k_s; in metals, they are discontinuous due to occupancy changes.
    Section 2; motivates the use of Baldereschi points for insulators and in metallic twist averaging.
  • domain assumption The ensemble of random functions with Fourier coefficients decaying exponentially with star index λ defines the optimality criterion; the Baldereschi point minimizes the mean-squared error in the large-λ limit.
    Section 3.1; this is Baldereschi's original definition, not proven from first principles; the paper acknowledges that for slowly convergent series the point's effectiveness must be checked empirically.
  • domain assumption Equation (4) correctly identifies whether a star carries an irrep or is a reducible combination of smaller stars, up to accidental degeneracies.
    Section 3.2 and Appendix A; the algorithm's correctness is verified numerically for the 14 lattices, not proven in general.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Baldereschi mean value points for three-dimensional Bravais lattices." pith.science (2026). https://pith.science/paper/QSWASEKD

@misc{pith2026260716974,
  author       = {Pith},
  title        = {Pith review of: Baldereschi mean value points for three-dimensional Bravais lattices},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QSWASEKD}},
  note         = {Machine review of arXiv:2607.16974}
}
read the original abstract

The Baldereschi point of a crystal is a wavevector in the Brillouin zone at which every smooth periodic function of wavevector lies close to its mean value. Although originally introduced in the context of one-electron methods, mean-value points are ideal for explicitly correlated many-electron methods such as quantum Monte Carlo simulations, which can only use a single Bloch wavevector in the Brillouin zone of a simulation supercell. We have therefore evaluated and tabulated the Baldereschi mean-value points of all fourteen three-dimensional Bravais lattices.

Figures

Figures reproduced from arXiv: 2607.16974 by the authors.

Figure 1
Figure 1. Nodes of the first two nonzero star functions A1(k) and A2(k) of a hexagonal lattice with c > √ 3a within the first BZ (the cyan hexagon). The Baldereschi points are shown by the filled circles. 5 [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Difference between the DFT-PBE energy of bulk hBN with a n × n × n k-point grid and the DFT-PBE energy with a large 123 × 123 × 40 k-point grid. Results are shown for different twists ks (i.e., offsets to the grid of k points). The labels “Γ”, “(1/4, 1/4, 1/4)”, and “Baldereschi” indicate the position of ks within the Brillouin zone of the unfolded n × n × n supercell. The twist-averaged results are averaged over ra… view at source ↗
Figure 3
Figure 3. Kinetic energy per electron T(N) of a two-component free electron gas in a simple cubic simulation cell against the reciprocal of the system size N. Results are shown for finite simulation cells subject to pure periodic boundary conditions, twisted periodic boundary condition with the twist at the simulation cell Baldereschi point, and twist-averaging. The kinetic energies with pure periodic boundary conditions and … view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: confirms that finite-size errors in the twist-averaged kinetic energy per electron fall off as N −4/3 [6], whereas the typical finite-size errors with periodic (Γ) and twisted periodic (Baldereschi) points fall off as N −1 , despite the fact that the latter has a much …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 1 linked inside Pith

  1. [8]

    Status Solidi B80K89–K93 URL https://onlinelibrary.wiley.com/doi/abs/10.1002/pssb.2220800239

    Bashenov V K, Bardashova M and Mutal A M 1977Phys. Status Solidi B80K89–K93 URL https://onlinelibrary.wiley.com/doi/abs/10.1002/pssb.2220800239

  2. [9]

    Stevanovic V 2024 A Python code for calculating the mean-value (Baldereschi’s) point for any crystal structure URLhttps://arxiv.org/abs/2405.00925

  3. [1]

    Baldereschi A 1973Phys. Rev. B7(12) 5212–5215 URL https://link.aps.org/doi/10.1103/PhysRevB.7.5212

  4. [2]

    Status Solidi B1199–40 URL https://onlinelibrary.wiley.com/doi/abs/10.1002/pssb.2221190102

    Evarestov R A and Smirnov V P 1983Phys. Status Solidi B1199–40 URL https://onlinelibrary.wiley.com/doi/abs/10.1002/pssb.2221190102

  5. [3]

    Monkhorst H J and Pack J D 1976Phys. Rev. B13(12) 5188–5192 URL https://link.aps.org/doi/10.1103/PhysRevB.13.5188

  6. [4]

    Rajagopal G, Needs R J, Kenny S, Foulkes W M C and James A 1994Phys. Rev. Lett.73(14) 1959–1962 URLhttps://link.aps.org/doi/10.1103/PhysRevLett.73.1959

  7. [5]

    Rajagopal G, Needs R J, James A, Kenny S D and Foulkes W M C 1995Phys. Rev. B51(16) 10591–10600 URLhttps://link.aps.org/doi/10.1103/PhysRevB.51.10591

  8. [6]

    Lin C, Zong F H and Ceperley D M 2001Phys. Rev. E64(1) 016702 URL https://link.aps.org/doi/10.1103/PhysRevE.64.016702

Show all 26 references
  1. [7]

    Cunningham S L 1974Phys. Rev. B10(12) 4988–4994 URL https://link.aps.org/doi/10.1103/PhysRevB.10.4988

  2. [10]

    Fraser L M, Foulkes W M C, Rajagopal G, Needs R J, Kenny S D and Williamson A J 1996 Phys. Rev. B53(4) 1814–1832 URL https://link.aps.org/doi/10.1103/PhysRevB.53.1814

  3. [11]

    Chiesa S, Ceperley D M, Martin R M and Holzmann M 2006Phys. Rev. Lett.97(7) 076404 URLhttps://link.aps.org/doi/10.1103/PhysRevLett.97.076404

  4. [12]

    Drummond N D, Needs R J, Sorouri A and Foulkes W M C 2008Phys. Rev. B78(12) 125106 URLhttps://link.aps.org/doi/10.1103/PhysRevB.78.125106

  5. [13]

    Azadi S and Foulkes W M C 2015J. Chem. Phys.143102807 ISSN 0021-9606 URL https://doi.org/10.1063/1.4922619

  6. [14]

    Helfrich B 1985Theor. Comp. Sci.41125–139 ISSN 0304-3975 URL https://www.sciencedirect.com/science/article/pii/0304397585900672

  7. [15]

    Chadi D J and Cohen M L 1973Phys. Rev. B8(12) 5747–5753 URL https://link.aps.org/doi/10.1103/PhysRevB.8.5747

  8. [16]

    Sobol I 1976USSR Comp. Math. Math. Phys.16236–242 ISSN 0041-5553 URL https://www.sciencedirect.com/science/article/pii/0041555376901543

  9. [17]

    Bratley P and Fox B L 1988ACM Trans. Math. Softw.1488–100 ISSN 0098-3500 URL https://doi.org/10.1145/42288.214372 13 IOP PublishingJournalvv(yyyy) aaaaaa Authoret al

  10. [18]

    Needs R J, Towler M D, Drummond N D, L´ opez R ´ ıos P and Trail J R 2020J. Chem. Phys. 152154106 ISSN 0021-9606 URLhttps://doi.org/10.1063/1.5144288

  11. [19]

    Thecasinoprogram is available to download for academic use, worldwide, free of cost

    Needs R J, Towler M D, Drummond N D and L´ opez R ´ ıos P 2026Casinoquantum Monte Carlo software. Thecasinoprogram is available to download for academic use, worldwide, free of cost. For commercial use of the software, please discuss with the authors. Access to casino’s Git re...

  12. [20]

    Chulkov E V and Psakh’e S G 1979Sov. Phys. J.221237–1238 ISSN 1573-9228 URL https://doi.org/10.1007/BF00894985

  13. [21]

    Lynch R W and Drickamer H G 1966J. Chem. Phys.44181–184 ISSN 0021-9606 URL https://doi.org/10.1063/1.1726442

  14. [22]

    Kristallogr.220567–570 URLhttp://www.degruyter.com/view/j/zkri.2005.220

    Clark S J, Segall M D, Pickard C J, Hasnip P J, Probert M I J, Refson K and Payne M C 2005Z. Kristallogr.220567–570 URLhttp://www.degruyter.com/view/j/zkri.2005.220. issue-5-6-2005/zkri.220.5.567.65075/zkri.220.5.567.65075.xml

  15. [23]

    Trail J R and Needs R J 2005J. Chem. Phys.122014112 URL http://scitation.aip.org/content/aip/journal/jcp/122/1/10.1063/1.1829049

  16. [24]

    Trail J R and Needs R J 2005J. Chem. Phys.122174109 URL http://scitation.aip.org/content/aip/journal/jcp/122/17/10.1063/1.1888569

  17. [25]

    Drummond N D, Towler M D and Needs R J 2004Phys. Rev. B70(23) 235119 URL https://link.aps.org/doi/10.1103/PhysRevB.70.235119

  18. [26]

    Umrigar C J, Toulouse J, Filippi C, Sorella S and Hennig R G 2007Phys. Rev. Lett.98(11) 110201 URLhttp://link.aps.org/doi/10.1103/PhysRevLett.98.110201 14

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.