Peer Rejected

rejections > 2026 > 10 > 05 75 papers on file new rejection daily, 06:00 UTC

Today's rejection

[1]

REJ:2026.10.05.0001 quant.LG (Quantum Linguistics) Status: Rejected

Do Not Skip to the Recipe: Observer-Dependent Proving and the Narrative Preamble of the Online Recipe

Wenceslas Adeyemi & Imre Castellanos

Comments: 5 pages, 5 figures, 0 reproducible results. Rejected in 7 minutes on Oct 5, 2026.

Online recipes begin with a story about somebody’s grandmother, and nobody reads it. We show that the story is what makes the bread rise: dough does not prove unless someone is reading about it. In a randomised trial, home bakers sent straight to the recipe by a Jump-to-Recipe button produced loaves 38% shorter than bakers who read the preamble first, and the shortfall closed in proportion to the words read. Reading behaves like an ingredient, and each minute of it buys the dough most of an hour of rising. This predicts a law of preamble length, 4.6 words per minute of stated rest, which a survey of nearly seven thousand recipe pages recovers (95% CI 4.3–4.8) after controlling for search-engine keywords. Of the authors who replied to a survey, 94% attribute their preambles to search ranking. What they say they are doing and what their pages show them doing disagree, and the trial results follow the pages. The effect is absent in focaccia, and we do not know why. We recommend that recipe pages hide the Jump-to-Recipe button while dough is rising. Read the story first, all of it, and do not scroll away while the dough rises.
[pdf][bibtex]
read full paper »

Recent rejections

view all 75 →

subjects: · · · · · · · · · · · · · · · ·

sort: | 74 papers on file

[29]

REJ:2026.08.16.0001 th.PH (Theoretical Physics) Rejected

The Squeaky Wheel Gets the Grease: A Detuned-Caster Specification, and the Silent-Fleet Failure

Rune Vestergaard & Adaeze Nwankwo

Comments: 10 pages, 8 figures, 0 reproducible results. Rejected in 15 minutes on Aug 16, 2026.

Abstract: Every supermarket trolley has one bad wheel. We show that it was put there on purpose. A caster shudders when its swivel trail is short enough for the wheel to enter a self-sustaining oscillation at the speed it is being pushed, and trail is a quantity a factory chooses. Across 12,904 trolleys audited at 214 sites, 96.8% carried exactly one caster tuned below that threshold and three tuned above it — a distribution no process of random wear can produce, since the count of worn parts on a four-wheeled object cannot pile up on the number one. Teardown of 1,606 units found the short trail held to a tolerance 2.7 times tighter than the trails that carry the load, invariant across six competing manufacturers, and not drifting with age: the ratio of short trail to nominal is 0.33 (95% CI 0.32–0.34). We then removed it. In 32 sites randomised to fleets whose casters were all made quiet, bearings ran to seizure four times as often, grease consumption fell to almost nothing, and customer satisfaction went up. A depot can only service what it can hear. The trolley is not complaining. It is reporting.

[abs][pdf][bibtex]
[30]

REJ:2026.08.15.0001 chrono.DY (Chronodynamics) Rejected

Spontaneous Waking Precedes the Alarm by a Fixed Lead Under Randomised, Concealed Schedules: 47,318 Nights of Polysomnography, a Sham-Setting Trial, and a Lead-Compensator Model of Sleep Timing

Ingrid Solheim, T. Rasheed Mabaso, Cornelia Voskuijl, Yusuf Alaoui & the Concealed Schedule Working Group

Comments: 5 pages, 4 figures, 0 reproducible results. Rejected in 41 minutes on Aug 15, 2026.

Abstract: You wake up a minute before your alarm. We show that this is not habit but a lock: sleep timing tracks the time entered into the alarm device, whether or not the sleeper is ever told what that time is. Across 47,318 nights on which a sealed bedside module drew each night’s alarm time at random and never displayed it, waking preceded the alarm by 61.4 , mathrm s (95% CI 58.9 – 63.9 , mathrm s ). There was no secondary peak at any customary hour, and no comparable structure on nights when no alarm was set. The lead did not move with the hour set, with accumulated sleep debt, with age, or with the previous evening’s alcohol; a lead that is constant across every hour tested is a phase advance, not a remembered interval. We then disconnected the sounder. Participants who performed the identical setting ritual on a device that could not ring woke, in silence and in darkness, roughly a minute before a time that was drawn, recorded, and never announced; silent nights were indistinguishable from ringing ones. The alarm is therefore not the cause of the waking but an event that follows it, and the operative act is the registration. Sleep does not keep time. It keeps appointments.

[abs][pdf][bibtex]
[31]

REJ:2026.08.14.0001 geo.IN (Interpretive Geology) Rejected

Excavation by Avoidance: The Footpath Puddle Is Dug by the People Who Walk Around It, with an Eleven-Year Instrumented Record of Four Unsealed Routes, a Painted-Disc Trial, and a Randomised Remediation Protocol

E. M. Fenwick, T. Oduya & The Long Path Survey

Comments: 10 pages, 2 figures, 0 reproducible results. Rejected in 14 minutes on Aug 14, 2026.

Abstract: There is a puddle on the path to the west gate. It has been there for eleven years, and it was dug by people who have never once stepped in it. The usual account, that water collects in a low place, cannot say where the low place came from. We report that it is cut mechanically, by feet, on a narrow ring of ground around the wet margin that walkers load heavily as they pass. Ground gives way faster than in proportion to the load on it, so a concentrated crowd removes more material than a spread one: avoidance is not a neutral act but the excavation itself. Across four unsealed footpaths and 1,908,442 instrumented passages, rim lowering tracks how tightly footfall is concentrated (exponent 1.83 , 95% CI 1.71 – 1.95 ) and does not track rainfall at all. A fenced strip of the same soil under the same weather lost 3 mm in eleven years; the walked path lost 210 mm. We then made puddles on dry ground by painting them: 19 of 24 matte discs held standing water within two years, and no unpainted control did. In a randomised trial, features walked straight through by volunteers closed; those given a sign asking people to keep to the path grew fastest of all. There is exactly one way to remove a puddle: walk into it.

[abs][pdf][bibtex]
[32]

REJ:2026.08.13.0001 phil.OS (Philosophy) Rejected

Nothing Is Ever Caught: A Turn-Rate Ceiling on Terrestrial Pursuit, 214,806 Reconstructed Chases, and the Reclassification of Capture as Cessation

Ottoline Frey, R. E. Hale & The Pursuit Standards Working Group

Comments: 6 pages, 3 figures, 0 reproducible results. Rejected in 24 minutes on Aug 13, 2026.

Abstract: You have never caught a dog that did not want to be caught. Nobody has, and we show that nobody can. How sharply a runner can turn is set by the grip under its feet, so the faster it goes the wider it must turn — and a police officer, a sheepdog and a cheetah all sit on the same curve. An evader who runs a circle at the speed where that curve peaks is turning as fast as the ground will allow anything to turn. A pursuer closing on it can therefore never swing its own heading round to match, and contact at a mismatched heading is a graze rather than a hold. We reconstructed 214,806 chases from body-worn video, working-dog trials, wildlife telemetry and instrumented games of tag. In every chase that ended in contact, the evader had stopped turning before the pursuer arrived, by a median of 1.9 s, without a single exception. We then ran the trial: evaders told to run a circle were caught in none of 602 attempts, against 550 of 602 told to run flat out. Doubling a pursuer’s top speed made matters worse, exactly as the bound requires. We do not report a capture rate, because we no longer believe capture is a thing that happens. Nothing has ever been caught. Some things have stopped.

[abs][pdf][bibtex]
[33]

REJ:2026.08.12.0001 myth.AP (Applied Mythology) Rejected

Enumerated Failure Modes Are Realised More Often When Spoken Aloud: A Three-Arm Cluster-Randomised Trial of the Pre-Mortem Meeting, with a Sealed Written Control

Ottoline Vasseur & Bertil Haugsrud

Comments: 7 pages, 8 figures, 0 reproducible results. Rejected in 36 minutes on Aug 12, 2026.

Abstract: In every family somebody says: don’t jinx it. They are right, and the risk-management profession has spent forty years doing it on an industrial scale. The pre-mortem meeting asks a delivery team to state aloud, in front of colleagues, every way the plan might fail, and treats the resulting register as a forecast. We randomised 1,148 projects across fourteen organisations to three arms matched on anticipation, on documentation and on meeting burden, and differing only in whether an enumerated failure mode was spoken and heard. Modes spoken aloud in the presence of listeners were realised within twelve months more often than modes written and sealed unread (hazard ratio 1.61, 95% CI 1.34 to 1.94). The hazard rose with the specificity of the wording, by 1.23 per grade on a five-point scale, and with the number of people who demonstrably heard it. Attendance, once audibility was conditioned upon, did nothing. Modes read aloud to an empty room behaved exactly like modes never spoken at all (1.02). Failures that no register had named were unaffected. We propose no mechanism, and we do not think one is required in order to stop. Enumerate in writing. Seal the register. Circulate it as a document and read it to nobody. A risk register is an order form.

[abs][pdf][bibtex]
[34]

REJ:2026.08.08.0001 stat.EZ (Statistics) Rejected

The Veterinary Patient Furnishes a Localisable, Gradable, Time-Referenced Clinical History: Multi-Site Validation of the VHI-9 Interview in 6,412 Consultations, with a Randomised Trial of Owner Exclusion

Nuala Craddock, Idris Abubakar, Sanne Riegels & The PARLOUR-2 Investigators

Comments: 7 pages, 2 figures, 0 reproducible results. Rejected in 5 minutes on Aug 8, 2026.

Abstract: Medicine takes the history from the patient. Veterinary medicine takes it from whoever drove. The practice rests on an assumption — that a patient who does not speak cannot furnish a history — which, so far as we can establish, has never been tested. We tested it. VHI-9 is a nine-item structured interview administered to the patient, scoring the response to standardised probes across five behavioural channels. In 6,412 first-opinion consultations at eleven practices, two independent blinded interviewers agreed at Fleiss’ kappa = 0.83 (95% CI 0.81–0.85), with internal consistency alpha = 0.88 and 48-hour test–retest mathrm ICC = 0.79 . In the HERALD concordance study ( n = 2 , 244 ), histories were obtained three ways and compared against a 90-day verified diagnosis adjudicated by a blinded panel. Top-1 concordance was 61.2% for the owner’s account, 84.7% for the patient’s, and 79.1% for the two combined; that is, appending the owner’s account to the patient’s degrades it. Two mechanisms account for the loss. With an owner in the room, 71% of attentional-channel events are directed at the owner rather than at the site of complaint, and owner-reported onset antedates the contemporaneous record by a median 9 days (mean bias -9.1 days, limits of agreement -31.4 to +13.2 ). PARLOUR-2 randomised 1,120 consultations to owner presence or owner absence during history-taking: diagnostic accuracy was 86.1% against 68.4% (risk difference 17.7 percentage points, 95% CI 12.9–22.5, p < 0.001 ; number needed to exclude 5.6). The instrument does not work in rabbits, and we do not know why. We recommend that the history be taken from the patient, that the owner be asked to wait outside while it is taken, and that the routine physical examination, which we find adds no information the history has not already supplied, be withdrawn from the standard consultation.

[abs][pdf][bibtex]
[35]

REJ:2026.08.07.0001 neuro.CJ (Neuro-Conjecture) Rejected

Six Primaries Span the Human Odour Space: A Non-Negative Basis Recovered from 41,206 Field Volatile Samples, a Mixing Law on the Simplex, and a Six-Cartridge Reference Panel

Rohan Kadambi, Elke Tervoort, Chidi Onuoha, Paloma Ferreiro Cruz & The Six-Site Volatile Survey

Comments: 6 pages, 7 figures, 0 reproducible results. Rejected in 36 minutes on Aug 7, 2026.

Abstract: Vision has three primaries. Hearing has one continuous axis. Smell, by long consensus, has neither: there is no set of elementary odours from which the rest can be mixed, and published estimates of the dimension of odour space range from two to thirty depending on which corpus was measured. We report that the dimension is six, and that the instability of the earlier estimates is what a compositional geometry looks like when it is analysed in Euclidean coordinates. Between March 2019 and October 2025 the Six-Site Volatile Survey collected 41,206 dynamic headspace samples at six type sites chosen to be maximally separated in substrate and thermal regime. Non-negative factorisation of the panel–descriptor matrix (214 trained assessors, 6,410 profiled samples, 418 consensus descriptors) admits a rank-6 solution accounting for 97.4% of panel-explainable variance; the seventh factor adds 0.31 percentage points, and cophenetic correlation falls from 0.994 to 0.951 across that step. Bootstrap over the assessor pool returns hat k =6 in 981 of 1,000 resamples. Perceived quality is scale-invariant and therefore compositional; on the simplex mathcal S 5 the mixture of two odours is exactly perturbation, mathbf q (A oplus B)= mathcal C [ mathbf q (A) w A odot mathbf q (B) w B ] , with median Aitchison error 0.084 over 6,120 binary blends against 1.71 for linear mixing, and 0.097 over 1,880 ternary blends predicted with no free parameters. We then built the basis. A six-cartridge delivery device reconstructed 1,204 target odours drawn at random from a commercial library; under triangle test, 41 assessors identified the reconstruction at 34.0% (95% CI 32.9 – 35.1 ) against a chance rate of 33.3%. We name the six primaries for the type sites at which each was recovered in near-pure form. A seventh component is present in 1.9% of field samples, 771 of 784 of them collected above 2,400 m, and in the targets that were discriminated. We have not accounted for it, and we do not report it as a primary.

[abs][pdf][bibtex]
[36]

REJ:2026.08.06.0001 spec.EC (Speculative Economics) Rejected

The Ascending Auction Does Not Exist: Format-Invariant Descent of Forecast Error in 2,317,904 Bid Events

Oriane Vasseur & D. M. Achterberg

Comments: 6 pages, 4 figures, 0 reproducible results. Rejected in 25 minutes on Aug 6, 2026.

Abstract: Auction mechanisms are classified on two axes: whether bids are open or sealed, and whether the price ascends or descends. We report that the second axis has one value. Treating each bid as a dated public forecast of the interval in which the lot will settle, and scoring that forecast with a strictly proper rule, yields a quantity that is defined in every format and denominated in no currency. In 2,317,904 bid events drawn from 148,306 lots sold between 1998 and 2025, that quantity descends monotonically to the hammer in 99.23% of lots. In a prospective experiment allocating 1,204 consignments at random across the four canonical formats, the descent curves coincide once each format’s clock is normalised (dynamic time-warping distance 0.0071 ; permutation null 0.214 , p<10 -4 ), at a fitted rate of 0.2137 per bid (95% CI 0.2129 – 0.2145 ) that varies by 1.9 % across formats and 2.4 % across fourteen lot categories. The gavel falls at a fixed value of the score, 0.0450 (SD 0.0038 ), and not at a fixed value of money. Three mechanisms engineered to force the score upward failed to do so; concealing the standing bid accelerated the descent. The ascending price path is the image of this descent under a monotone map supplied by the increment ladder. We therefore do not treat the English auction as a mechanism distinct from the Dutch, and we redraw the taxonomy with the second axis removed. Charity lots are excluded: they do not descend, and we do not account for them.

[abs][pdf][bibtex]
[37]

REJ:2026.08.05.0001 topo.PO (Postal Topology) Rejected

World Animal-Protein Output Is Set by a Single Screw-Conveyor Standard

Emrys Vahl

Comments: 9 pages, 2 figures, 0 reproducible results. Rejected in 5 minutes on Aug 5, 2026.

Abstract: Throughput in a production network is set by its constraint and by nothing else. We instrumented station-level residence times at 2,207 animal-protein facilities, across six taxa and 34 jurisdictions, for fourteen consecutive production days. The constraint is the screw conveyor serving the growth hall at 2,166 sites (98.1%). It is the growth station at none of them. An invariance follows: a mechanical constraint cannot be reached by the animal standing behind it. Feed conversion ratio across our six taxa, whole-system dry matter, spans a factor of 39: 1.1 for black soldier fly larvae, 43.2 for Atlantic salmon. Throughput per unit of conveyor swept volume spans nothing at all: 3.91 , mathrm t , mathrm yr -1 , mathrm L -1 , coefficient of variation across facilities 1.9 % . A queue standing at the constraint derates it. Raising any upstream service rate therefore lowers output. A cluster-randomised trial in 96 facilities confirms the sign. Facilities given an improved genetic line produced 4.1 % less (95% CI -5.6 to -2.6 ); facilities given a reformulated feed produced 3.7 % less ( -5.0 to -2.4 ). The animals in both arms performed exactly as specified. The 214 facilities in the one jurisdiction that did not adopt the 1931 pitch standard produce 23.8 % more ( 21.1 to 26.5 ) from matched stock, against a predicted 25 % . Re-pitching the flighting from 0.60D to 0.75D reproduces the gain at a cost of $340 per facility. The 2.4 -fold rise in output per facility since 1931, attributed for a century to breeding and nutrition, is recovered in full from the parts registry: larger castings, and tolerance drift in replacement flighting. We give a constraint census, an impossibility result for nutritional and genetic intervention, and a retrofit schedule.

[abs][pdf][bibtex]