{"id":"b2ce83ca-9f27-4d8d-8c3c-707b71d98321","arxiv_id":"2605.26507","paper_version":3,"verdict":"CONDITIONAL","confidence":"LOW","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Conditional tie weighting replaces unobserved higher-priority genuine ties by their conditional probability so partially censored pairs still inform restricted win statistics.","lead":"This paper proposes conditional tie weighting so censored patient pairs can still contribute fractionally to hierarchical win statistics in clinical trials. It aims for the same restricted-time estimand as existing methods while recovering efficiency lost to censoring-induced ties.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.5","headline":"Full text is the wrong paper (SIKA-GP); identification of conditional genuine-tie probability for the win-statistic estimand remains unchecked.","rationale":"The reader correctly flagged that only the abstract of 2605.26507 is available and that the cached body is a different paper, yielding CONDITIONAL with LOW confidence. Stress-testing confirms the same load-bearing gap: the method’s value rests on identification of the conditional higher-priority tie probability so fractional contributions leave the restricted-time estimand unchanged. That step is not present in the supplied full text. No stronger endorsement or rejection is warranted until the correct manuscript is examined; the reader’s weakest_assumption is exactly the right one. Verdict therefore stays CONDITIONAL (abstract is methods-shaped and plausible; body and proofs still required).","tokens_in":22312,"tokens_out":489,"duration_ms":14330,"concrete_test":"Retrieve the actual arXiv 2605.26507 PDF/source. Locate the identification theorem for the conditional genuine-tie probability; check whether it requires only observable pairwise data and standard independent censoring (or weaker MAR-type) assumptions, or whether it imposes untestable restrictions on the joint event/censoring process. If the latter, or if the proof is missing/incomplete, the claim that the estimator targets the same restricted-time win probabilities fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that replacing the higher-priority genuine-tie indicator by its conditional probability given the observed pairwise data preserves the restricted-time win probabilities (same estimand as all-or-nothing restricted win statistics). That identification is asserted in the abstract but is the step that must hold under the censoring and nuisance models. The CACHEABLE full manuscript is SIKA-GP (arXiv 2605.26509 on sparse inducing kernels for GPs), not the win-statistics paper (2605.26507). Consequently the identification argument, the form of the two-sample U-statistic with estimated nuisances, the sandwich variance derivation, and the simulation/HF-ACTION design cannot be inspected. Without those, one cannot confirm that fractional weights do not shift the estimand when lower-priority comparisons are informative only under censoring-induced higher-priority ties.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"From the abstract alone, the paper proposes conditional tie weighting for hierarchical composite survival endpoints under right censoring. Existing restricted win-statistic estimators require higher-priority genuine ties to be fully observed before lower-priority comparisons can contribute; the proposed method replaces the unobserved higher-priority genuine-tie indicator by its conditional probability given the observed pairwise data, so partially observed pairs can contribute fractionally. The abstract claims this targets the same restricted-time win probabilities as the all-or-nothing restricted estimators, develops identification and large-sample theory for two-sample U-statistics with estimated nuisances, supplies sandwich variances for win ratio, net benefit, and win odds, and reports efficiency gains plus an HF-ACTION reanalysis. The body text supplied with the submission is not this manuscript: it is an unrelated paper (SIKA-GP on sparse inducing-kernel approximations for Gaussian processes).","tokens_in":22492,"tokens_out":809,"duration_ms":12664,"significance":"If the identification claim holds under standard censoring and nuisance models, conditional tie weighting would be a useful methodological contribution: it would preserve a clinically interpretable restricted-time estimand while recovering information that current restricted win statistics discard under heavy censoring or long restriction horizons. That would matter for cardiovascular and other trials that use death-first hierarchical composites. The claimed sandwich variances and U-statistic theory with estimated nuisances would also be practically valuable. These strengths cannot be credited on the present file, because the proofs, assumptions, simulations, and HF-ACTION analysis are not present in the supplied full text.","major_comments":[{"comment":"Manuscript mismatch: the title, abstract, and arXiv id (2605.26507, stat.ME) describe conditional tie weighting for win statistics, but the full manuscript text is SIKA-GP (arXiv 2605.26509, cs.LG) on sparse inducing kernels for GPs. None of the claimed identification argument, U-statistic theory, sandwich variances, simulations, or HF-ACTION reanalysis appears in the body. The central claim cannot be refereed from the abstract alone.","section":null},{"comment":"Load-bearing identification (asserted only in the abstract): the claim that replacing the higher-priority genuine-tie indicator by its conditional probability given the observed pairwise data preserves the restricted-time win probabilities (rather than shifting the estimand) is the step on which the whole contribution rests. Without the derivation, censoring assumptions, and nuisance-model conditions, one cannot verify that fractional weights for censoring-induced ties leave the restricted-time estimand unchanged when lower-priority comparisons are informative.","section":null},{"comment":"Uninspectable large-sample theory and inference: the abstract asserts two-sample U-statistics with estimated nuisance functions and sandwich variances for win ratio, net benefit, and win odds. The form of the kernel, the influence-function expansion, regularity conditions on the nuisance estimators, and the sandwich construction are not available in the supplied text, so asymptotic validity and variance correctness cannot be assessed.","section":null},{"comment":"Uninspectable empirical support: efficiency gains under heavier censoring and longer restriction horizons, and the HF-ACTION death-first hospitalization reanalysis, are asserted but not present. Without design, data-generating mechanisms, competitor estimators, and numerical results, the practical claims cannot be checked.","section":null}],"minor_comments":[],"recommendation":"uncertain","confidential_remarks":"The submission appears to have the wrong full manuscript attached (SIKA-GP instead of the win-statistics paper). This is almost certainly a production/upload error rather than a scientific defect of the intended work. I recommend the editor request the correct PDF for 2605.26507 and reassign for a full review; I would be willing to re-review the correct manuscript. Until then a scientific accept/reject decision is not possible."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The only usable material for 2605.26507 is the abstract. The cached full text is SIKA-GP (2605.26509), a completely different paper on sparse inducing kernels for GPs. So this is an abstract-only read with low confidence.\n\nWhat is new, on the abstract’s own terms: conditional tie weighting. Existing restricted win-statistic estimators keep the restricted-time estimand by requiring higher-priority genuine ties to be observed; that all-or-nothing rule throws away pairs whose lower-priority comparison is already informative but whose higher-priority tie is only censoring-induced. The proposal replaces the missing genuine-tie indicator by its conditional probability given the observed pairwise data, so those pairs contribute fractionally. They claim identification of the same restricted-time win probabilities, large-sample theory for two-sample U-statistics with estimated nuisances, sandwich variances for win ratio / net benefit / win odds, efficiency gains under heavy censoring and longer horizons, and an HF-ACTION reanalysis (death-first hospitalization).\n\nThat is a clean, well-motivated estimator idea for an active niche in trial methodology. The circularity burden looks low: they say they keep the existing restricted-time estimand and only change the weighting. Free parameters (restriction horizon, nuisance models for the conditional genuine-tie probability) are the usual ones for this literature.\n\nThe soft spot is load-bearing and currently uncheckable. Everything turns on whether that conditional probability is correctly identified under the censoring mechanism and the nuisance models so that fractional weights do not shift the estimand. The abstract asserts identification and theory; without proofs, assumptions, simulation design, or the trial reanalysis, we cannot verify it. The stress-test is right: wrong full text means the central claim stays open.\n\nWho it is for: biostatisticians working on hierarchical composites and win statistics under censoring. If the identification holds and the efficiency gains are real, it is useful secondary-analysis and design material. It is serious enough in framing to deserve a full manuscript and a serious referee once the correct PDF is available. I would not cite or bring it to reading group on the abstract alone. Send to peer review only after the right paper is in hand; desk-rejecting the idea itself would be premature.","headline":"Abstract-only methods idea for fractional weighting of censoring-induced ties in hierarchical win statistics; full text is the wrong paper, so identification and efficiency claims stay unchecked.","tokens_in":23101,"tokens_out":556,"would_cite":false,"duration_ms":5907,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["62N01","62G05","62P10"],"pacs":[],"model":"grok-4.5","headline":"Fractional weights let censored patient pairs still inform hierarchical win statistics without changing the target.","keywords":["win statistics","hierarchical composite endpoints","right censoring","conditional tie weighting","restricted-time estimand","U-statistics","win ratio","net benefit"],"falsifier":"In a simulation with known restricted-time win probabilities, heavy censoring, and deliberately misspecified tie-probability nuisances, check whether the conditional-tie-weight estimator remains unbiased for the restricted-time target (or whether bias appears while variance still falls).","tokens_in":23146,"feed_emoji":"📉","tokens_out":814,"duration_ms":12428,"temperature":0.7,"pith_summary":"In trials that rank outcomes (for example death first, then hospitalization), a pair of patients is compared only on the next outcome when the higher-priority comparison is a genuine tie. Right censoring often leaves that higher-priority tie unconfirmed even when the lower-priority outcome is fully observed, so standard restricted win-statistic estimators discard the pair entirely. This paper replaces the missing genuine-tie indicator with its conditional probability given what was observed for that pair. The resulting estimator still targets the same restricted-time win probabilities, net benefit, and win odds, but lets partially observed pairs contribute a fractional weight when their lower-priority comparison is informative. Theory for two-sample U-statistics with estimated nuisance functions, sandwich variances, simulations, and a reanalysis of a heart-failure trial support large efficiency gains under heavier censoring and longer restriction times.","feed_headline":"Censored pairs still count in hierarchical win statistics","feed_subtitle":"Conditional probabilities replace hard ties so lower-priority outcomes can contribute without changing the target.","key_machinery":"Conditional tie weight: the unavailable higher-priority genuine-tie indicator is replaced by its conditional probability given the observed pairwise data, turning a hard exclusion into a fractional contribution inside two-sample U-statistics with estimated nuisance functions.","core_discovery":"Conditional tie weighting recovers the same restricted-time hierarchical win probabilities as existing all-or-nothing restricted win-statistic estimators while allowing pairs with censoring-induced higher-priority ties to contribute fractionally whenever the lower-priority comparison is observed and informative.","pith_inferences":["The same conditional-probability idea may extend to hierarchies with more than two priority levels if intermediate ties can be modeled analogously.","If the nuisance models for tie probabilities can be made robust or doubly robust, the method could tolerate more realistic censoring and dependence structures common in multi-event survival data.","Efficiency gains large enough under heavy censoring may change power calculations and sample-size planning for hierarchical composite primary endpoints."],"forward_implications":["Under heavy censoring or long restriction horizons, hierarchical win ratio, net benefit, and win odds can be estimated with substantially smaller variance without changing the scientific target.","Pairs that currently contribute nothing in death-first hospitalization hierarchies can still inform treatment comparisons when hospitalization is observed.","Sandwich variance formulas for win ratio, net benefit, and win odds become available for the fractionally weighted U-statistics.","Completed trials with hierarchical composite endpoints can be reanalyzed to recover information previously discarded by all-or-nothing restricted estimators."],"fun_headline_variants":["Conditional ties let censored pairs count in hierarchical win stats","Fractional weighting recovers info from censoring-induced ties","Censoring-induced ties contribute via conditional probabilities","Same restricted win probabilities with partial credit for censored pairs","Hierarchical win stats that fractionally use censored pair data"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The method works only if the conditional probability of a true higher-priority tie, given the observed pair data, is correctly identified and estimated; if that nuisance model is wrong, the fractional weights can shift the estimand rather than merely improve precision.","fun_headline_variants_meta":{"raw":{"variants":["Conditional ties let censored pairs count in hierarchical win stats","Fractional weighting recovers info from censoring-induced ties","Censoring-induced ties contribute via conditional probabilities","Same restricted win probabilities with partial credit for censored pairs","Hierarchical win stats that fractionally use censored pair data"]},"model":"grok-4.5","effort":"low","cost_usd":0.00364,"raw_usage":{"total_tokens":1184,"prompt_tokens":770,"num_sources_used":0,"completion_tokens":81,"cost_in_usd_ticks":36400000,"prompt_tokens_details":{"text_tokens":770,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":333,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":770,"tokens_out":81,"duration_ms":3059,"temperature":1.0,"reasoning_tokens":333,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-12T15:56:52.836304+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"In a simulation with known restricted-time win probabilities, heavy censoring, and deliberately misspecified tie-probability nuisances, check whether the conditional-tie-weight estimator remains unbiased for the restricted-time target (or whether bias appears while variance still falls).","supporting_citations":[],"review_version":3}