2024-Q1 — Backfill: opening the index
The index opens with a retrospective quarter. What was accepted, what was refused, and the four problems that were obvious from the first hundred records.
11 reports, one per quarter since the index opened. Each covers what was submitted, what the numbers did, what changed in method or rule set, and what the index got wrong. They are the record of how this dataset came to look the way it does, and they are the only place the index writes about itself at length.
| Quarter | Report | Records | Batches | Median purity | Flagged | What changed |
|---|---|---|---|---|---|---|
| 2024-Q1 | Backfill: opening the index | 163 | 27 | 99.09 | 107 | Backfill quarter. Reports assayed earlier were accepted retrospectively where custody could be documented. |
| 2024-Q2 | First live quarter: the purity ceiling | 349 | 50 | 99.08 | 280 | First quarter of live submission. |
| 2024-Q3 | A capable laboratory arrives, and the short-peptide problem becomes visible | 522 | 86 | 98.97 | 356 | Nordanalyt Laboratories began contributing. |
| 2024-Q4 | Impurity fingerprinting: nineteen bulks, twenty-nine labels | 434 | 76 | 98.97 | 321 | Impurity-fingerprint matching introduced; first shared-bulk links published. |
| 2025-Q1 | The provenance effect: 0.31 pp | 702 | 106 | 99.12 | 531 | Anomaly rule set expanded from seven rules to eleven. |
| 2025-Q2 | Closing the mass balance | 556 | 99 | 99.07 | 410 | Gravimetric mass-balance reconciliation became mandatory on new records. |
| 2025-Q3 | A fast laboratory arrives, and gradient length becomes a number | 819 | 157 | 98.79 | 639 | Kestrel Bioanalytical began contributing; gradient-length normalisation introduced. |
| 2025-Q4 | Digit preference: what a round number tells you | 883 | 144 | 99.00 | 703 | Provenance classification became a required field. |
| 2026-Q1 | Integration is where purity is decided | 832 | 149 | 99.08 | 625 | Rule AX-R12 added. Dataset v2 schema published. |
| 2026-Q2 | Buyer-sourced sensitivity analysis on every scorecard | 1,009 | 187 | 98.90 | 771 | Buyer-sourced sensitivity analysis added to every supplier scorecard. |
| 2026-Q3 | Quarter in progress | 453 | 73 | 98.92 | 365 | Quarter in progress. Aggregates for this quarter are provisional. |
The index opens with a retrospective quarter. What was accepted, what was refused, and the four problems that were obvious from the first hundred records.
Live submission opens. The purity distribution turns out to be compressed at the top, which makes purity a poor discriminator and content a good one.
Nordanalyt begins contributing ion-pair, SEC and ICP-MS work. Its first records show that a large part of the index had been measuring the wrong thing.
Named-impurity data reaches critical mass and the index can identify lots that are the same material sold twice.
With provenance recorded on every record, the index can measure whether supplier-supplied samples read better than material bought independently. They do.
Gravimetric reconciliation becomes mandatory on new records. Doing the arithmetic reveals how many certificates cannot be reconciled at all.
Kestrel Bioanalytical begins contributing on a 15-minute gradient. Quantifying its offset turns a methodological argument into a measurement.
Terminal-digit analysis on declared purity figures, and what happens when you put the round numbers next to independent measurements of the same lots.
Rule AX-R12 is added. The uncertainty budget shows that the choice between two integration conventions is larger than every other error source combined.
Supplier pages now report their buyer-sourced subset separately. For some suppliers the two views differ enough to change the reading.
A provisional report for the current quarter. Figures will change as records arrive; the standing observations will not.