Three readings set the frame. Throughput: in the week to 12 September Japanese refiners processed 2,638,109 cubic metres of crude, equal to 2,370,600 barrels a day, down 11,100 b/d on the week, while average crude distillation unit utilisation rose from 82.9% to 87.2%. Runs: for 27 September to 3 October, refinery utilisation was 75.9% against a prior 77.9%. Stocks: in the week to 17 October commercial crude stocks stood at 11,294,800 kilolitres, a 652,200 kL rebuild from 10,642,600 kL a week earlier, roughly 71.04 million barrels at 6.2898 barrels per kilolitre. Throughput falling while stocks first bottomed and then rebuilt inside a month tells its own story: this is not missing demand, it is refiners cutting runs on purpose while cargoes arrived in a bunch.

The dual-capacity gap deserves more attention. On 17 October utilisation was 83% on operable capacity (BPSD) but only 76.4% on design capacity (BPCD), a 6.6 percentage point spread; a week earlier the figures were 86.4% and 78.8%, a 7.6 point gap. BPSD excludes units idled long-term that cannot restart quickly; BPCD keeps them in the denominator. A persistent six to eight point wedge means Japan's genuinely callable refining capacity is structurally below nominal nameplate, and that gap does not close when a turnaround round ends. Anyone sizing Japanese supply against design capacity will systematically overestimate what refiners can commit.

A second, easily missed signal is the loss of visibility. The Petroleum Association of Japan has suspended its weekly breakdown of gasoline, jet fuel, kerosene and gasoil stocks, citing changes in the structure of Japan's oil product supply. For outside analysts the high-frequency read on the product-side stock buffer is gone, pushing observation back to monthly prints. At the same time utilisation was still falling week on week into 17 October, BPSD down from 86.4% to 83%, so maintenance season has not yet passed its trough. When refiners steer constrained feedstock towards gasoil and jet, base oil yield and export availability, being a co-produced stream, typically tighten with it.

Three implications for SNSUC. First, spot availability of naphtha and base oil out of Japan will swing wider through the fourth quarter, so inquiries into Japanese refinery-origin barrels should be prepared for volume not being filled rather than price failing to agree. Second, quotation validity has to be set in weeks, not months, since both cracking and base oil are repricing weekly right now. Third, SNSUC's addressable categories are LNG and LPG, SN150 and SN500 base oil, paraffin and lubricants; hazardous chemicals are out of scope without exception, and that same boundary applies to any Japanese-origin discussions. One caution: this is market observation built on public data and is not investment advice, and because PAJ has already changed its reporting basis once, always reconcile units and coverage when splicing historical series.