WTMAD-4: A Fair Weighting Scheme for GMTKN55

Erin R. Johnson; Kyle R. Bryenton

arxiv: 2509.23498 · v2 · pith:PNJJGE33new · submitted 2025-09-27 · ⚛️ physics.chem-ph

WTMAD-4: A Fair Weighting Scheme for GMTKN55

Kyle R. Bryenton , Erin R. Johnson This is my paper

Pith reviewed 2026-05-22 13:12 UTC · model grok-4.3

classification ⚛️ physics.chem-ph

keywords GMTKN55WTMADdensity functional approximationsdispersion correctionweighted mean absolute deviationbenchmark metricsquantum chemistry

0 comments

The pith

A new WTMAD-4 weighting scheme for the GMTKN55 benchmark set fixes under-weighting of some tests by orders of magnitude.

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

The GMTKN55 data set gathers many standard tests of molecular energies, barriers, and interactions that quantum chemistry methods must pass. Common ways of combining the errors from these tests into a single score, called WTMAD, assign extremely small influence to some of the component benchmarks. The authors trace this imbalance to the way the weights were originally chosen. They define a replacement metric, WTMAD-4, whose weights come from the typical error sizes seen with ten simple dispersion-corrected density functionals. When the new metric is applied to 115 functionals, the overall rankings shift and an earlier functional tuned to the old weights performs noticeably worse on the previously marginalised tests.

Core claim

The paper establishes that existing WTMAD definitions under-weight certain GMTKN55 benchmarks by orders of magnitude. It introduces WTMAD-4, whose weights are set from the typical errors of ten minimally empirical dispersion-corrected density-functional approximations, to produce fair treatment across all benchmarks. Reassessment of 115 DFAs with WTMAD-4 then shows that a functional previously optimised by minimising WTMAD-2 underperforms on the benchmarks that had been marginalised by the older metric.

What carries the argument

WTMAD-4, a weighted mean absolute deviation whose component weights are chosen inversely to the typical errors observed for a fixed set of ten minimally empirical dispersion-corrected density functionals.

If this is right

DFAs that were optimised by minimising older WTMAD variants will show weaker results on the benchmarks that had received negligible weight.
Overall performance rankings of density functionals shift when the fairer weighting is used instead of WTMAD-2 or WTMAD-3.
The new metric gives comparable influence to small-molecule thermochemistry, reaction barriers, and non-covalent interaction tests.
Literature functionals tuned to the old weighting scheme require re-evaluation for balanced accuracy across the full GMTKN55 collection.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same error-based weighting idea could be tested on other large benchmark collections to check whether similar imbalances exist.
Method developers could adopt WTMAD-4 as a default composite score when designing or comparing new density functionals.
Composite metrics in computational chemistry may need periodic recalibration whenever the typical error profile of standard methods changes.

Load-bearing premise

The typical errors observed for the specific set of ten minimally empirical dispersion-corrected DFAs provide an appropriate and representative basis for defining weights that achieve fair treatment of all GMTKN55 component benchmarks.

What would settle it

A substantially different set of reference DFAs yields weights that change the relative importance of the GMTKN55 benchmarks by more than a factor of two, or re-ranking the 115 functionals with WTMAD-4 leaves the literature example unchanged in its relative performance.

read the original abstract

The GMTKN55 data set is a collection of standard benchmarks used in molecular quantum chemistry that spans small- and large-molecule thermochemistry, reaction barriers, and non-covalent interactions. Herein, we identify a flaw in the weighted mean absolute deviation (WTMAD) definitions commonly used to quantify performance of various electronic-structure methods for the GMTKN55 set, which under-weight some of its component benchmarks by orders of magnitude. A new WTMAD-4 metric is proposed, based on typical errors observed for a set of ten minimally empirical dispersion-corrected density-functional approximations (DFAs), ensuring fair treatment across all benchmarks. The performance of 115 DFAs is then reassessed using WTMAD-4 and we highlight a literature example where a DFA parametrised by minimising WTMAD-2 underperforms for benchmarks marginalised by that metric.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper spots real under-weighting in standard GMTKN55 WTMAD schemes and supplies a new WTMAD-4 based on errors from ten specific DFAs, plus a re-ranking of 115 methods, but the weights rest on an untested reference set.

read the letter

The main point is that existing WTMAD definitions for GMTKN55 have been assigning some benchmarks weights that are orders of magnitude smaller than others, and this paper builds a new version called WTMAD-4 by taking typical errors from ten minimally empirical dispersion-corrected DFAs and setting weights inversely to those errors. They then re-score 115 functionals and flag one case where a method tuned to the old WTMAD-2 ends up weaker on the benchmarks that had been down-weighted before. That is a concrete, usable output for anyone who relies on GMTKN55 rankings. The identification of the weighting imbalance is straightforward and the re-assessment gives numbers that show how rankings can shift, which is the kind of evidence that makes the proposal worth considering. The work stays grounded in the actual error data rather than abstract arguments about fairness. The soft spot is the dependence on exactly those ten DFAs. They share a common design philosophy, so their relative error sizes may not match what you would see from hybrids, range-separated functionals, or more empirical variants. The paper reports the new rankings but does not test how much the weights or the final order change if you pick a different reference group or add some variation. That leaves the claim of fair treatment resting on an assumption that is not checked. No code or raw error tables are provided either, so anyone who wants to use or adjust the scheme has to rebuild it. This is aimed at people who benchmark density functionals or other electronic-structure methods against GMTKN55. A reader who compares many DFAs or who has published rankings based on the older metrics will get direct value from the re-ranking data and the example of the tuned functional. It is coherent enough and addresses a practical issue in a widely used benchmark, so it deserves a serious referee. I would recommend sending it to peer review, with the main request being a sensitivity check on the reference-set choice.

Referee Report

2 major / 2 minor

Summary. The manuscript identifies a flaw in existing WTMAD definitions for the GMTKN55 benchmark set, which under-weight certain component benchmarks by orders of magnitude. It proposes a new WTMAD-4 metric whose benchmark weights are set inversely proportional to the typical errors observed across a fixed set of ten minimally empirical dispersion-corrected DFAs. The performance of 115 DFAs is then reassessed with WTMAD-4, and a literature example is highlighted in which a DFA parametrized by minimizing WTMAD-2 underperforms on benchmarks that were marginalized by the earlier weighting.

Significance. If the central construction holds, WTMAD-4 would provide a more balanced ranking of density functionals on GMTKN55 by ensuring that benchmarks with intrinsically larger errors (e.g., certain reaction barriers or non-covalent interactions) receive appropriate weight. The reassessment of 115 methods supplies a substantial body of comparative data, and the explicit demonstration of a practical consequence for a previously published parametrization strengthens the case for adoption. The approach is grounded in observed error statistics rather than arbitrary scaling factors.

major comments (2)

[§3] §3 (Definition of WTMAD-4): the weights are constructed from the mean absolute errors of a hand-selected reference set of exactly ten minimally empirical dispersion-corrected DFAs. No sensitivity test is reported that quantifies how the resulting weights, or the final ordering of the 115 evaluated DFAs, change when the reference set is replaced by another plausible collection (e.g., including range-separated hybrids). Because the claim of “fair treatment across all benchmarks” rests on the representativeness of these ten error profiles, the absence of such a test is load-bearing.
[Results] Results section (reassessment of 115 DFAs): while aggregate WTMAD-4 values are tabulated, the manuscript provides no uncertainty estimates or bootstrap-style error bars on the weights themselves. Small differences in reported WTMAD-4 scores between closely ranked functionals could therefore be statistically insignificant, weakening the interpretation of the ranking shifts relative to WTMAD-2.

minor comments (2)

[Abstract] The abstract refers to “a literature example” without naming the functional or the original reference; adding the specific DFA and citation would improve immediate clarity.
[§2] Notation for the ten reference DFAs is introduced without a compact table listing their exact names, dispersion corrections, and basis sets; a short summary table would aid reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed report. The comments identify valid opportunities to strengthen the robustness of WTMAD-4. We address each major comment below and will incorporate the suggested analyses in the revised manuscript.

read point-by-point responses

Referee: [§3] §3 (Definition of WTMAD-4): the weights are constructed from the mean absolute errors of a hand-selected reference set of exactly ten minimally empirical dispersion-corrected DFAs. No sensitivity test is reported that quantifies how the resulting weights, or the final ordering of the 115 evaluated DFAs, change when the reference set is replaced by another plausible collection (e.g., including range-separated hybrids). Because the claim of “fair treatment across all benchmarks” rests on the representativeness of these ten error profiles, the absence of such a test is load-bearing.

Authors: We selected the ten minimally empirical dispersion-corrected DFAs because they span a range of functional forms while avoiding heavy empirical parametrization that could bias the error profiles toward particular benchmarks. This choice was intended to provide a balanced, representative sampling of typical errors. Nevertheless, we agree that an explicit sensitivity test would reinforce the claim of fair treatment. In the revised manuscript we will add a new subsection (or appendix) that repeats the weight derivation using an alternative reference set that includes range-separated hybrids and report the resulting changes (or lack thereof) in benchmark weights and the ordering of the 115 DFAs. revision: yes
Referee: [Results] Results section (reassessment of 115 DFAs): while aggregate WTMAD-4 values are tabulated, the manuscript provides no uncertainty estimates or bootstrap-style error bars on the weights themselves. Small differences in reported WTMAD-4 scores between closely ranked functionals could therefore be statistically insignificant, weakening the interpretation of the ranking shifts relative to WTMAD-2.

Authors: We concur that uncertainty quantification would improve the interpretation of small ranking differences. We will perform a bootstrap resampling analysis on the underlying MAE values used to derive the weights, propagate the resulting variability to the WTMAD-4 scores, and include error bars (or confidence intervals) in the revised tables and figures. This will allow readers to assess whether observed differences between closely ranked functionals are statistically meaningful. revision: yes

Circularity Check

0 steps flagged

No significant circularity in WTMAD-4 derivation

full rationale

The paper defines WTMAD-4 weights from typical errors observed on a fixed external reference set of ten minimally empirical dispersion-corrected DFAs. This reference set is chosen independently of the 115 DFAs later reassessed, and the provided abstract and description contain no equations or steps in which the new metric or benchmark weights reduce by construction to a fit or prediction performed on the target evaluation set itself. The construction is a definitional reweighting scheme grounded in external benchmarks rather than a self-referential loop, satisfying the criteria for a self-contained derivation against external data.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that typical errors from a chosen reference set of DFAs can define fair weights, plus the free parameter of which ten DFAs are selected; no new physical entities are postulated.

free parameters (1)

Choice of ten minimally empirical DFAs
The specific selection of which ten DFAs supply the typical-error values directly determines the numerical weights in WTMAD-4.

axioms (1)

domain assumption GMTKN55 component benchmarks should receive weights inversely related to typical DFA errors to achieve fair treatment.
This premise is invoked to justify replacing prior WTMAD definitions and is not derived from first principles within the paper.

pith-pipeline@v0.9.0 · 5676 in / 1415 out tokens · 39021 ms · 2026-05-22T13:12:37.417666+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

WTMAD-4 weights are chosen such that each benchmark contributes roughly evenly (between 1% and 3%) to the overall WTMAD-4... based on the magnitudes of expected errors rather than on the absolute energy scales.

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

44 extracted references · 44 canonical work pages · 1 internal anchor

[1]

Goerigk, A

L. Goerigk, A. Hansen, C. Bauer, S. Ehrlich, A. Najibi and S. Grimme, Phys. Chem. Chem. Phys., 2017, 19, 32184--32215

work page 2017
[2]

Santra, M

G. Santra, M. Cho and J. M. Martin, J. Phys. Chem. A, 2021, 125, 4614--4627

work page 2021
[3]

Wittmann, H

L. Wittmann, H. Neugebauer, S. Grimme and M. Bursch, J. Chem. Phys., 2023, 159, 224103

work page 2023
[4]

A. D. Becke, J. Chem. Phys., 2024, 160, 204118

work page 2024
[5]

o rling, T. Gould, S. Grimme, O. Gritsenko, H. J. A. Jensen, E. R. Johnson, R. O. Jones, M. Kaupp, A. M. K\

A. M. Teale, T. Helgaker, A. Savin, C. Adamo, B. Aradi, A. V. Arbuznikov, P. W. Ayers, E. J. Baerends, V. Barone, P. Calaminici, E. Canc\` e s, E. A. Carter, P. K. Chattaraj, H. Chermette, I. Ciofini, T. D. Crawford, F. D. Proft, J. F. Dobson, C. Draxl, T. Frauenheim, E. Fromager, P. Fuentealba, L. Gagliardi, G. Galli, J. Gao, P. Geerlings, N. Gidopoulos,...

work page 2022
[6]

A. J. A. Price, A. Otero-de-la-Roza and E. R. Johnson, Chem. Sci., 2023, 14, 1252--1262

work page 2023
[7]

V. Blum, R. Gehrke, F. Hanke, P. Havu, V. Havu, X. Ren, K. Reuter and M. Scheffler, Comput. Phys. Commun., 2009, 180, 2175--2196

work page 2009
[8]

X. Ren, P. Rinke, V. Blum, J. Wieferink, A. Tkatchenko, A. Sanfilippo, K. Reuter and M. Scheffler, New J. Phys., 2012, 14, 053020

work page 2012
[9]

S. V. Levchenko, X. Ren, J. Wieferink, R. Johanni, P. Rinke, V. Blum and M. Scheffler, Comput. Phys. Commun., 2015, 192, 60--69

work page 2015
[10]

Kokott, F

S. Kokott, F. Merz, Y. Yao, C. Carbogno, M. Rossi, V. Havu, M. Rampp, M. Scheffler and V. Blum, J. Chem. Phys., 2024, 161, 024112

work page 2024
[11]

V. W.-z. Yu, F. Corsetti, A. Garc \'i a, W. P. Huhn, M. Jacquelin, W. Jia, B. Lange, L. Lin, J. Lu, W. Mi, A. Seifitokaldani, A. V\' a zquez-Mayagoitia, C. Yang, H. Yang and V. Blum, Comput. Phys. Commun., 2018, 222, 267--285

work page 2018
[12]

V. Havu, V. Blum, P. Havu and M. Scheffler, J. Chem. Phys., 2009, 228, 8367--8379

work page 2009
[13]

A. C. Ihrig, J. Wieferink, I. Y. Zhang, M. Ropo, X. Ren, P. Rinke, M. Scheffler and V. Blum, New J. Phys., 2015, 17, 093020

work page 2015
[14]

E. R. Johnson, gmtkn55-fhiaims, 2024, https://github.com/erin-r-johnson/gmtkn55-fhiaims

work page 2024
[15]

Roadmap on Advancements of the FHI-aims Software Package

J. W. Abbott, C. M. Acosta, A. Akkoush, A. Ambrosetti, V. Atalla, A. Bagrets, J. Behler, D. Berger, B. Bieniek, J. Bj\" o rk, V. Blum, S. Bohloul, C. L. Box, N. Boyer, D. S. Brambila, G. A. Bramley, K. R. Bryenton, M. Camarasa-G\' o mez, C. Carbogno, F. Caruso, S. Chutia, M. Ceriotti, G. Cs\' a nyi, W. Dawson, F. A. Delesma, F. D. Sala, B. Delley, R. A. D...

work page internal anchor Pith review Pith/arXiv arXiv 2025
[16]

F. O. Kannemann and A. D. Becke, J. Chem. Theory Comput., 2010, 6, 1081--1088

work page 2010
[17]

A. D. Becke, J. Chem. Phys., 1986, 84, 4524--4529

work page 1986
[18]

J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865

work page 1996
[19]

J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1997, 78, 1396--1396

work page 1997
[20]

Zhang and W

Y. Zhang and W. Yang, Phys. Rev. Lett., 1998, 80, 890--890

work page 1998
[21]

A. D. Becke, J. Chem. Phys., 1986, 85, 7184--7187

work page 1986
[22]

A. D. Becke, Phys. Rev. A, 1988, 38, 3098

work page 1988
[23]

A. D. Beck, J. Chem. Phys., 1993, 98, 5648--5652

work page 1993
[24]

C. Lee, W. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785

work page 1988
[25]

P. J. Stephens, F. J. Devlin, C. F. Chabalowski and M. J. Frisch, J. Phys. Chem., 1994, 98, 11623--11627

work page 1994
[26]

S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200--1211

work page 1980
[27]

Adamo and V

C. Adamo and V. Barone, J. Chem. Phys., 1999, 110, 6158--6170

work page 1999
[28]

A. D. Becke, J. Chem. Phys., 1993, 98, 1372--1377

work page 1993
[29]

K. R. Bryenton, A. A. Adeleke, S. G. Dale and E. R. Johnson, WIRES: Comput. Mol. Sci., 2023, 13, e1631

work page 2023
[30]

O. A. Vydrov and G. E. Scuseria, J. Chem. Phys., 2006, 125, 234109

work page 2006
[31]

O. A. Vydrov, G. E. Scuseria and J. P. Perdew, J. Chem. Phys., 2007, 126, 234109

work page 2007
[32]

J. A. Pople, M. Head-Gordon, D. J. Fox, K. Raghavachari and L. A. Curtiss, J. Chem. Phys., 1989, 90, 5622--5629

work page 1989
[33]

L. A. Curtiss, C. Jones, G. W. Trucks, K. Raghavachari and J. A. Pople, J. Chem. Phys., 1990, 93, 2537--2545

work page 1990
[34]

Tkatchenko and M

A. Tkatchenko and M. Scheffler, Phys. Rev. Lett., 2009, 102, 073005

work page 2009
[35]

Tkatchenko, R

A. Tkatchenko, R. A. DiStasio Jr, R. Car and M. Scheffler, Phys. Rev. Lett., 2012, 108, 236402

work page 2012
[36]

Ambrosetti, A

A. Ambrosetti, A. M. Reilly, R. A. DiStasio Jr and A. Tkatchenko, J. Chem. Phys., 2014, 140, 18A508

work page 2014
[37]

Hermann and A

J. Hermann and A. Tkatchenko, Phys. Rev. Lett., 2020, 124, 146401

work page 2020
[38]

M\" u ller and S

M. M\" u ller and S. Ehlert, gmtkn55, 2021, https://github.com/grimme-lab/GMTKN55, https://github.com/grimme-lab/GMTKN55

work page 2021
[39]

Santra, N

G. Santra, N. Sylvetsky and J. M. Martin, J. Phys. Chem. A, 2019, 123, 5129--5143

work page 2019
[40]

A. J. A. Price, K. R. Bryenton and E. R. Johnson, J. Chem. Phys., 2021, 154, 230902

work page 2021
[41]

Korth and S

M. Korth and S. Grimme, J. Chem. Theory Comput., 2009, 5, 993–1003

work page 2009
[42]

Gasevic, M

T. Gasevic, M. Müller, J. Schöps, S. Lanius, J. Hermann, S. Grimme and A. Hansen, ChemRxiv preprint chemrxiv-2025-rdsd0, 2025

work page 2025
[43]

C. A. Goodhart, in Monetary theory and practice: The UK experience, Springer, 1984, pp. 91--121

work page 1984
[44]

Strathern, Eur

M. Strathern, Eur. Rev., 1997, 5, 305--321 mcitethebibliography main.tex0000664000000000000000000021315215066047405011237 0ustar rootroot [twoside,twocolumn,9pt] article extsizes [super,sort&compress,comma] natbib [version=3] mhchem [left=1.5cm, right=1.5cm, top=1.785cm, bottom=2.0cm] geometry balance mathptmx sectsty graphicx lastpage [format=plain,justi...

work page doi:10.1039/cxcp00000x/ 1997

[1] [1]

Goerigk, A

L. Goerigk, A. Hansen, C. Bauer, S. Ehrlich, A. Najibi and S. Grimme, Phys. Chem. Chem. Phys., 2017, 19, 32184--32215

work page 2017

[2] [2]

Santra, M

G. Santra, M. Cho and J. M. Martin, J. Phys. Chem. A, 2021, 125, 4614--4627

work page 2021

[3] [3]

Wittmann, H

L. Wittmann, H. Neugebauer, S. Grimme and M. Bursch, J. Chem. Phys., 2023, 159, 224103

work page 2023

[4] [4]

A. D. Becke, J. Chem. Phys., 2024, 160, 204118

work page 2024

[5] [5]

o rling, T. Gould, S. Grimme, O. Gritsenko, H. J. A. Jensen, E. R. Johnson, R. O. Jones, M. Kaupp, A. M. K\

A. M. Teale, T. Helgaker, A. Savin, C. Adamo, B. Aradi, A. V. Arbuznikov, P. W. Ayers, E. J. Baerends, V. Barone, P. Calaminici, E. Canc\` e s, E. A. Carter, P. K. Chattaraj, H. Chermette, I. Ciofini, T. D. Crawford, F. D. Proft, J. F. Dobson, C. Draxl, T. Frauenheim, E. Fromager, P. Fuentealba, L. Gagliardi, G. Galli, J. Gao, P. Geerlings, N. Gidopoulos,...

work page 2022

[6] [6]

A. J. A. Price, A. Otero-de-la-Roza and E. R. Johnson, Chem. Sci., 2023, 14, 1252--1262

work page 2023

[7] [7]

V. Blum, R. Gehrke, F. Hanke, P. Havu, V. Havu, X. Ren, K. Reuter and M. Scheffler, Comput. Phys. Commun., 2009, 180, 2175--2196

work page 2009

[8] [8]

X. Ren, P. Rinke, V. Blum, J. Wieferink, A. Tkatchenko, A. Sanfilippo, K. Reuter and M. Scheffler, New J. Phys., 2012, 14, 053020

work page 2012

[9] [9]

S. V. Levchenko, X. Ren, J. Wieferink, R. Johanni, P. Rinke, V. Blum and M. Scheffler, Comput. Phys. Commun., 2015, 192, 60--69

work page 2015

[10] [10]

Kokott, F

S. Kokott, F. Merz, Y. Yao, C. Carbogno, M. Rossi, V. Havu, M. Rampp, M. Scheffler and V. Blum, J. Chem. Phys., 2024, 161, 024112

work page 2024

[11] [11]

V. W.-z. Yu, F. Corsetti, A. Garc \'i a, W. P. Huhn, M. Jacquelin, W. Jia, B. Lange, L. Lin, J. Lu, W. Mi, A. Seifitokaldani, A. V\' a zquez-Mayagoitia, C. Yang, H. Yang and V. Blum, Comput. Phys. Commun., 2018, 222, 267--285

work page 2018

[12] [12]

V. Havu, V. Blum, P. Havu and M. Scheffler, J. Chem. Phys., 2009, 228, 8367--8379

work page 2009

[13] [13]

A. C. Ihrig, J. Wieferink, I. Y. Zhang, M. Ropo, X. Ren, P. Rinke, M. Scheffler and V. Blum, New J. Phys., 2015, 17, 093020

work page 2015

[14] [14]

E. R. Johnson, gmtkn55-fhiaims, 2024, https://github.com/erin-r-johnson/gmtkn55-fhiaims

work page 2024

[15] [15]

Roadmap on Advancements of the FHI-aims Software Package

J. W. Abbott, C. M. Acosta, A. Akkoush, A. Ambrosetti, V. Atalla, A. Bagrets, J. Behler, D. Berger, B. Bieniek, J. Bj\" o rk, V. Blum, S. Bohloul, C. L. Box, N. Boyer, D. S. Brambila, G. A. Bramley, K. R. Bryenton, M. Camarasa-G\' o mez, C. Carbogno, F. Caruso, S. Chutia, M. Ceriotti, G. Cs\' a nyi, W. Dawson, F. A. Delesma, F. D. Sala, B. Delley, R. A. D...

work page internal anchor Pith review Pith/arXiv arXiv 2025

[16] [16]

F. O. Kannemann and A. D. Becke, J. Chem. Theory Comput., 2010, 6, 1081--1088

work page 2010

[17] [17]

A. D. Becke, J. Chem. Phys., 1986, 84, 4524--4529

work page 1986

[18] [18]

J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865

work page 1996

[19] [19]

J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1997, 78, 1396--1396

work page 1997

[20] [20]

Zhang and W

Y. Zhang and W. Yang, Phys. Rev. Lett., 1998, 80, 890--890

work page 1998

[21] [21]

A. D. Becke, J. Chem. Phys., 1986, 85, 7184--7187

work page 1986

[22] [22]

A. D. Becke, Phys. Rev. A, 1988, 38, 3098

work page 1988

[23] [23]

A. D. Beck, J. Chem. Phys., 1993, 98, 5648--5652

work page 1993

[24] [24]

C. Lee, W. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785

work page 1988

[25] [25]

P. J. Stephens, F. J. Devlin, C. F. Chabalowski and M. J. Frisch, J. Phys. Chem., 1994, 98, 11623--11627

work page 1994

[26] [26]

S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200--1211

work page 1980

[27] [27]

Adamo and V

C. Adamo and V. Barone, J. Chem. Phys., 1999, 110, 6158--6170

work page 1999

[28] [28]

A. D. Becke, J. Chem. Phys., 1993, 98, 1372--1377

work page 1993

[29] [29]

K. R. Bryenton, A. A. Adeleke, S. G. Dale and E. R. Johnson, WIRES: Comput. Mol. Sci., 2023, 13, e1631

work page 2023

[30] [30]

O. A. Vydrov and G. E. Scuseria, J. Chem. Phys., 2006, 125, 234109

work page 2006

[31] [31]

O. A. Vydrov, G. E. Scuseria and J. P. Perdew, J. Chem. Phys., 2007, 126, 234109

work page 2007

[32] [32]

J. A. Pople, M. Head-Gordon, D. J. Fox, K. Raghavachari and L. A. Curtiss, J. Chem. Phys., 1989, 90, 5622--5629

work page 1989

[33] [33]

L. A. Curtiss, C. Jones, G. W. Trucks, K. Raghavachari and J. A. Pople, J. Chem. Phys., 1990, 93, 2537--2545

work page 1990

[34] [34]

Tkatchenko and M

A. Tkatchenko and M. Scheffler, Phys. Rev. Lett., 2009, 102, 073005

work page 2009

[35] [35]

Tkatchenko, R

A. Tkatchenko, R. A. DiStasio Jr, R. Car and M. Scheffler, Phys. Rev. Lett., 2012, 108, 236402

work page 2012

[36] [36]

Ambrosetti, A

A. Ambrosetti, A. M. Reilly, R. A. DiStasio Jr and A. Tkatchenko, J. Chem. Phys., 2014, 140, 18A508

work page 2014

[37] [37]

Hermann and A

J. Hermann and A. Tkatchenko, Phys. Rev. Lett., 2020, 124, 146401

work page 2020

[38] [38]

M\" u ller and S

M. M\" u ller and S. Ehlert, gmtkn55, 2021, https://github.com/grimme-lab/GMTKN55, https://github.com/grimme-lab/GMTKN55

work page 2021

[39] [39]

Santra, N

G. Santra, N. Sylvetsky and J. M. Martin, J. Phys. Chem. A, 2019, 123, 5129--5143

work page 2019

[40] [40]

A. J. A. Price, K. R. Bryenton and E. R. Johnson, J. Chem. Phys., 2021, 154, 230902

work page 2021

[41] [41]

Korth and S

M. Korth and S. Grimme, J. Chem. Theory Comput., 2009, 5, 993–1003

work page 2009

[42] [42]

Gasevic, M

T. Gasevic, M. Müller, J. Schöps, S. Lanius, J. Hermann, S. Grimme and A. Hansen, ChemRxiv preprint chemrxiv-2025-rdsd0, 2025

work page 2025

[43] [43]

C. A. Goodhart, in Monetary theory and practice: The UK experience, Springer, 1984, pp. 91--121

work page 1984

[44] [44]

Strathern, Eur

M. Strathern, Eur. Rev., 1997, 5, 305--321 mcitethebibliography main.tex0000664000000000000000000021315215066047405011237 0ustar rootroot [twoside,twocolumn,9pt] article extsizes [super,sort&compress,comma] natbib [version=3] mhchem [left=1.5cm, right=1.5cm, top=1.785cm, bottom=2.0cm] geometry balance mathptmx sectsty graphicx lastpage [format=plain,justi...

work page doi:10.1039/cxcp00000x/ 1997