{"id":"975d62e1-3794-4ce7-b1ec-3807fb4b51ff","arxiv_id":"2607.03290","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"Directional half-space conditioning defines v-variograms with closed forms in standard multivariate Pareto models and a half-space-mass-driven bias–variance tradeoff that ensemble estimators can exploit.","lead":"The paper generalizes extremal variograms by conditioning multivariate extremes on arbitrary half-spaces instead of single components. Choosing and combining directions yields a bias–variance tradeoff and can cut estimation error for rare-event dependence.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The reader correctly isolates the domain-of-attraction approximation and the HR-centric simulation design as the softest points, yet these are standard, openly stated limitations of the genre and do not threaten the mathematical core or the reported simulation patterns. The proofs for Propositions 3.2–3.8, Lemmas 4.1–4.2, Proposition 4.3 and Corollaries 4.6–4.7 are self-contained and use only classical tools (tilting, Hausdorff measure on hyperplanes, properties of conditionally negative definite matrices). Ensemble variance reduction is a direct consequence of positively correlated but overlapping half-space samples and is visible in Figures 9–10. Because no load-bearing flaw was located that would alter the ACCEPT verdict, the recommendation is UNCHANGED.","tokens_in":25553,"tokens_out":516,"duration_ms":4635,"concrete_test":"Independently re-derive the HR density identity (4.3) and the centering claim of Corollary 4.6 for a concrete low-dimensional Γ (e.g., d=3 with known Γ) by direct integration of the generator representation (2.2) against the half-space indicator, without invoking the pseudo-determinant lemmas; if the normalizing constant and mean of W^{v0} match the paper, the load-bearing HR theory is confirmed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central mathematical claims (decomposition Y^v =^d W^v + E1, closed-form Γ^v for logistic/Dirichlet/HR, HR density representations via non-canonical half-spaces, and the resistance-curvature characterization of v0) rest on standard EVT constructions (exponent measures, generators, exponential tilting, oblique projections) and are supported by complete proofs in Appendix C. The statistical claims about bias–variance trade-offs and ensemble gains are simulation-based and explicitly scoped to HR limits and domain-of-attraction data; the paper does not over-claim transfer. Finite second moments of differences under half-space conditioning (Definition 3.9) are required for Γ^v to exist, but this is the same regularity already used for classical extremal variograms and holds for the three parametric families treated. No internal inconsistency or hidden assumption that would undermine the strongest claim was found.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper generalizes the extremal variogram of Engelke and Volgushev (2022) by conditioning a multivariate generalized Pareto vector Y on arbitrary half-spaces H_v = {y : v^T Y > 0} for v in the probability simplex. It defines Y^v = (Y | v^T Y > 0) and the v-variogram Gamma^v_ij = Var(Y_i^v - Y_j^v), proves the decomposition Y^v =^d W^v + E 1 with W^v the v-extremal function on v^perp and E independent Exp(1), and relates the laws of Y^v and W^v across directions via exponential tilting. Closed forms for Gamma^v are obtained for logistic, Dirichlet, and Husler-Reiss models; for HR, Gamma^v equals the parameter matrix for every v, new density representations are derived, and the resistance-curvature vector v0 is shown to center W^{v0} and minimize half-space mass Lambda(H_v). Empirical v-variograms and ensemble averages over collections V are introduced, and simulations (mainly HR limits and domain-of-attraction data) document a bias-variance trade-off driven by Lambda(H_v) and variance reduction from ensembles.","tokens_in":25764,"tokens_out":1051,"duration_ms":8018,"significance":"The work cleanly extends a standard dependence summary for multivariate extremes from axis-aligned conditioning to general half-spaces, with complete proofs of the stochastic representations, tilting identities, and closed forms (Appendix C). The HR density formulae and the characterization of v0 (least-mass half-space and centered Gaussian extremal function) are new and link extremal statistics to Euclidean-distance / resistance geometry. The simulation study is carefully designed (M=2000, fixed and random Gamma, multiple thresholds and ensemble constructions) and shows that ensembles can improve MSE relative to the classical average of unit-vector variograms. Strengths include full proofs, recovery of known special cases, and transparent scoping of the statistical claims to HR-type settings. The contribution is solid and useful for inference and graphical modeling of extremes.","major_comments":[{"comment":"Section 5 and Figures 6-10: the bias-variance and ensemble conclusions are demonstrated only for Husler-Reiss limits and max-stable HR data in the domain of attraction. Because Gamma^v is independent of v only for HR (Example 3.13), the practical recommendation to average over multiple v is less clear for logistic or Dirichlet models, where Gamma^v itself depends on v. A short simulation or discussion for at least one non-HR family would make the statistical claims more transferable; without it the ensemble gains remain HR-specific.","section":null},{"comment":"Definition 3.9 and Section 5.1: existence of Gamma^v requires finite second moments of component differences under every half-space conditioning. This is the same regularity used for classical extremal variograms and holds for the three parametric families treated, but the paper does not state sufficient conditions on the exponent measure (or generator) that guarantee these moments for general Y. A brief remark on when the empirical v-variogram is well-defined would strengthen the general theory.","section":null}],"minor_comments":[{"comment":"Notation: the paper uses both Lambda_d and Delta^{d-1} for the simplex in places (e.g., Example 3.13 vs. Section 2.1); unify to one symbol.","section":null},{"comment":"Figure 1 (right) and Table 1: the ensemble constructions are clear in the text but the figure legend could name the two ensembles more explicitly (e.g., {e1,e2} vs. 11-point grid).","section":null},{"comment":"Remark 4.8 / Eq. (4.4): the convex program for v0* when v0 has negative entries is well-motivated; a one-line note that standard QP solvers suffice would help practitioners.","section":null},{"comment":"Appendix B: the Hausdorff-measure convention for densities on hyperplanes is carefully defined; a forward reference from Lemma 3.6 / Proposition 3.8 would help readers who skip the appendix.","section":null},{"comment":"Typos: 'H¨ usler' spacing and occasional missing spaces after commas in the arXiv text; also 'Lambda_d' vs. 'Delta' inconsistency noted above.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is technically careful and the proofs check out against known special cases. The main limitation is the HR-centric simulation study; that is a scope issue rather than a correctness issue and is fixable with a short additional experiment or clearer scoping language. Fit for a methods journal in extremes is good."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This is a genuine extension of Engelke & Volgushev (2022), not a re-labeling. They replace axis conditioning with an arbitrary simplex direction v, define Y^v and the v-variogram, prove the decomposition Y^v =^d W^v + E 1, and get the exponential-tilting relations that connect different directions. Closed forms for logistic, Dirichlet, and Hüsler–Reiss are new and useful; for HR they also give density representations via non-canonical half-spaces and identify the resistance-curvature vector v0 as the unique centering / least-mass direction. The math is standard EVT toolkit (generators, tilting, oblique projections) done carefully, with full proofs and checks against the known unit-vector and equal-weight cases.\n\nWhat the paper does well is make the statistical consequence concrete. Simulations (M=2000, fixed and random Γ, several thresholds) show a clean half-space-mass bias–variance trade-off: unit vectors give low variance / high bias, v0 the opposite, and ensembles sit in the sweet spot and beat the classical average of the d axis variograms. That is practical and immediately usable with the existing graphicalExtremes package.\n\nSoft spots are the usual ones for this genre and do not sink the claims. Everything statistical is scoped to HR limits and max-stable HR data in the domain of attraction; transfer to other dependence structures is assumed. No real-data example, and the free parameters (threshold, choice of V, sparsity of candidate v) still need user judgment. Finite second moments under every half-space are required, but that is the same regularity the classical extremal variogram already needs and holds for the three families treated. Circularity is negligible: Γ^v = Γ for all v in HR is derived, not imposed.\n\nThis is for people who already work with extremal graphical models or threshold-based multivariate extremes and want a better summary statistic without full likelihood. It deserves a serious referee. I would cite it when I next need a flexible extremal dependence summary, and I would bring it to reading group.","headline":"Clean, usable generalization of extremal variograms to arbitrary half-spaces, with solid theory and a clear bias–variance picture for HR.","tokens_in":26423,"tokens_out":545,"would_cite":true,"duration_ms":5504,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["62G32","62H12","60G70"],"pacs":[],"model":"grok-4.5","headline":"Conditioning multivariate extremes on arbitrary half-spaces yields directional variograms whose ensembles cut estimation variance while trading bias against half-space mass.","keywords":["multivariate extremes","generalized Pareto distribution","extremal variogram","v-variogram","Hüsler–Reiss","extremal function","half-space conditioning","ensemble estimation"],"falsifier":"In a high-dimensional simulation from a non-Hüsler–Reiss domain-of-attraction model, check whether ensembles of empirical v-variograms still reduce MSE relative to the classical single-component average once the threshold is varied; systematic failure would refute the claimed generality of the bias–variance mechanism.","tokens_in":26426,"feed_emoji":"📐","tokens_out":956,"duration_ms":8056,"temperature":0.7,"pith_summary":"Standard extremal variograms condition a multivariate generalized Pareto vector on a single component exceeding zero. That choice is arbitrary and can waste data or inflate bias. This paper replaces the axis with any direction v from the probability simplex, defining the conditioned vector Y^v and its v-variogram of pairwise differences. The vector always factors as an independent exponential shift plus a v-extremal function living on the hyperplane orthogonal to v. Closed forms are obtained for logistic, Dirichlet and Hüsler–Reiss models; for the last family the v-variogram equals the model parameter for every v, and a special resistance-curvature direction centers the Gaussian law while minimizing expected sample size. Empirical ensembles that average several directions substantially reduce variance relative to any single conditioning, with the bias–variance balance governed by the mass of the half-space. The practical payoff is a tunable, multi-direction estimator that can be matched to the extremity of the threshold at hand.","feed_headline":"Directional half-spaces cut variance in extremal variograms","feed_subtitle":"Ensembles over many conditioning directions beat single-axis methods via a bias–variance trade-off","key_machinery":"The v-extremal function W^v obtained by oblique projection of the half-space-conditioned vector Y^v (or equivalently by exponential tilting of a generator), which carries the entire dependence structure of the v-variogram and links densities across directions by exponential tilting.","core_discovery":"For any direction v in the probability simplex the conditioned multivariate generalized Pareto vector admits the decomposition Y^v =^d W^v + E 1, where W^v is the v-extremal function on v^perp and E is independent standard exponential; the associated v-variogram equals the Hüsler–Reiss parameter matrix for every v, admits closed forms for logistic and Dirichlet models, and its empirical ensembles improve mean-squared error by exploiting the bias–variance trade-off controlled by half-space mass.","pith_inferences":["The same half-space geometry should extend immediately to other summary statistics (extremal coefficients, madograms, tail dependence functions) that are currently computed only on axis-aligned exceedances.","Because half-space mass is a continuous function of v, one can treat direction choice itself as a continuous tuning parameter and optimize it by cross-validation on held-out extremes.","The resistance-curvature vector links the statistical construction to discrete curvature of Euclidean distance matrices, suggesting that geometric invariants of the variogram may systematically control estimation efficiency."],"forward_implications":["Practitioners can replace the classical average of d axis-aligned extremal variograms by a weighted or unweighted ensemble over any collection of directions, with expected variance reduction from overlapping half-spaces.","The resistance-curvature direction v0 (or its simplex projection) supplies a canonical low-bias estimator whose half-space contains the fewest expected exceedances.","New closed-form Hüsler–Reiss densities parameterized by arbitrary v simplify likelihood construction and graphical-model inference.","Bias-dominated low-threshold regimes favor directions of small half-space mass; variance-dominated high-threshold regimes favor unit vectors.","Logistic and Dirichlet models obtain explicit v-dependent variogram formulae usable for moment matching or goodness-of-fit."],"fun_headline_variants":["Half-space directions unlock v-variograms for multivariate GP extremes","Multi-v ensembles cut MSE via half-space mass bias-variance trade-off","Any simplex v yields Y^v = W^v + E1 and closed-form extremal variograms","Resistance-curvature v0 centers Hüsler-Reiss law on least-mass half-space","Directional conditioning generalizes single-axis extremal variogram inference"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"Thresholded, empirically margin-transformed samples behave like exact draws from the limiting multivariate generalized Pareto law, and the reported bias–variance patterns transfer beyond Hüsler–Reiss dependence.","fun_headline_variants_meta":{"raw":{"variants":["Half-space directions unlock v-variograms for multivariate GP extremes","Multi-v ensembles cut MSE via half-space mass bias-variance trade-off","Any simplex v yields Y^v = W^v + E1 and closed-form extremal variograms","Resistance-curvature v0 centers Hüsler-Reiss law on least-mass half-space","Directional conditioning generalizes single-axis extremal variogram inference"]},"model":"grok-4.5","effort":"low","cost_usd":0.00485,"raw_usage":{"total_tokens":1471,"prompt_tokens":898,"num_sources_used":0,"completion_tokens":93,"cost_in_usd_ticks":48500000,"prompt_tokens_details":{"text_tokens":898,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":480,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":898,"tokens_out":93,"duration_ms":4028,"temperature":1.0,"reasoning_tokens":480,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-12T03:28:53.783893+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"In a high-dimensional simulation from a non-Hüsler–Reiss domain-of-attraction model, check whether ensembles of empirical v-variograms still reduce MSE relative to the classical single-component average once the threshold is varied; systematic failure would refute the claimed generality of the bias–variance mechanism.","supporting_citations":[],"review_version":1}