Corridor selection — how Gauge Intelligence delineates freight corridors

A corridor is the unit of analysis that distinguishes Gauge Intelligence from operator-level or destination-level reporting. It groups performance by the infrastructure route a train runs over, rather than by who runs it or where it ends. Because the corridor is the lens, the way corridors are drawn determines every figure published against them. This page sets out the rule we apply, and why that rule is the one an independent record can defend.

This page is part of the standing methodology for the public archive (see also Methodology) and complements the data-window page, which sets out what the TRUST feed can and cannot see. Where the data-window page governs what is measured, this page governs where it is measured.

The rule

A corridor is one of the named freight corridors delineated by Network Rail’s Freight Network Study, published in April 2017, which identified eleven key freight corridors across the network. Gauge Intelligence represents each named corridor as an ordered sequence of TIPLOC timing points, running from origin to destination. The corridor is the route; the ordered point list is its machine-readable form.

Gauge Intelligence does not invent corridor boundaries. We adopt Network Rail’s strategic delineation as published, and our work is confined to expressing that delineation as an ordered point sequence that TRUST movement data can be matched against. The corridor set is external-authoritative: it comes from a published Network Rail study, not from anything in our own data or our own editorial judgement.

The Great Eastern Main Line corridor — Felixstowe to Stratford, extended to London Gateway — is the worked example throughout this page. It runs from the port’s container terminals through Ipswich, Colchester, Chelmsford and Shenfield to Stratford, and onward via Barking and Tilbury Town to the London Gateway terminal. Its full point sequence is published with the corridor’s register entry.

Why external-authoritative provenance matters

The most serious objection a methodology reviewer can raise against any corridor-level statistic is that the corridor was drawn to produce the result. Move a boundary a few miles, include or exclude a difficult junction, start the measurement after the worst bottleneck, and a corridor’s punctuality can be made to look materially better or worse without a single train changing its behaviour. This is the gerrymandering objection, and for a publisher that positions itself as an independent record it is the objection that matters most.

Adopting Network Rail’s published delineation pre-empts it. Because the corridor set comes from the Freight Network Study rather than from Gauge Intelligence’s own analysis, we cannot have defined a corridor to manufacture a performance result. We did not choose the eleven corridors, we did not choose where each one begins and ends, and we did not choose which routes count as strategic freight corridors and which do not. A reviewer who wishes to contest a corridor boundary is contesting Network Rail’s strategic planning document, not a Gauge Intelligence editorial decision. The provenance of the boundary is the defence of the figure measured inside it.

This is the load-bearing argument of the page. Every other choice we make about corridors — how a journey is matched to one, how multiple corridors are handled, how gauge and route availability are recorded — is downstream of it. The point of measurement is not ours to move.

From named corridor to measurable segment

A named corridor in a strategic study is a line on a map. To measure performance against it, the corridor has to become a concrete sequence of timing points that real-world movement messages can be tested against.

Each corridor is stored as an ordered list of points. Every point carries its STANOX code (from CORPUS), its TIPLOC code (from the CIF timetable feed), and a human-readable name. The ordering is the origin-to-destination direction of the corridor, so that a journey’s progress along the corridor can be read off the sequence in which its movement reports arrive.

Journey segments are assigned to corridors by matching TRUST movement locations against these point definitions. As a freight service reports at successive TIPLOCs and STANOXs through its journey, those reported locations are tested against each corridor’s point list. A run of consecutive matches identifies the segment of that journey which traversed the corridor. Performance figures — arrival punctuality at the corridor exit, scheduled-versus-actual delay at any intermediate point, cancellation counts — are then computed over the population of segments assigned to the corridor.

A single journey may traverse more than one corridor. A container service from Felixstowe to the West Midlands shares its first several timing points with the Great Eastern Main Line corridor before branching at Haughley Junction onto the cross-country route to the West Midlands. Each corridor’s figures count only the segment of the journey that ran over it; the journey itself is one continuous freight service, and its overall punctuality is reported separately from the per-corridor segment figures. This is why a corridor figure is always a statement about a stretch of railway, not about a set of trains.

Two infrastructure attributes are carried per point because they bound what can run over the corridor at all. Gauge clearance is recorded on the W-series scale, which describes the loading-gauge envelope a container wagon needs; a corridor cleared to a higher W-number can carry taller intermodal boxes on standard wagons. Route availability is recorded on the RA 1-to-10 scale, which describes the axle-load the route can bear. On the Great Eastern Main Line the points are cleared to W10 — the standard for nine-foot-six high-cube containers on standard platform wagons — and carry route availability of RA 8 at the port terminals rising to RA 9 along the main line. These attributes are not performance figures; they are the physical context that explains which traffic a corridor can host and therefore which services its performance population contains.

The inaugural corridor

The Great Eastern Main Line was the first corridor Gauge Intelligence published. That is a decision about publication order, not about how corridors are selected. The rule above applies identically to all eleven; this corridor simply went first.

The reasons it went first are straightforward. Felixstowe is Britain’s busiest container port, so the corridor carries the largest single intermodal flow on the network and produces the deepest population of freight movements to measure. It is the primary intermodal route from that port, which makes it the corridor where corridor-level analysis has the most to say. And it is cleared to W10 throughout, so the physical constraints that complicate gauge-restricted routes do not muddy the inaugural worked example. Publishing it first put the methodology in front of the largest and cleanest freight flow available, which is the right place to prove it.

None of this changes the selection rule. The corridor is not on the register because it is busy; it is on the register because the Freight Network Study named it one. Its prominence governed when it was published, not whether it qualified.

Publication order and what is next

The eleven corridors of the Freight Network Study are the universe. Gauge Intelligence publishes them progressively rather than all at once, so that each corridor’s point sequence can be validated against live TRUST data and each corridor report can be brought up to the standing methodology before it is added to the public register.

Not all of the named corridors are published yet. The register will grow toward the full set as each corridor is validated and its first period report is prepared. The selection rule does not change as the register grows: every corridor added is a corridor the Freight Network Study named, expressed as an ordered point sequence, measured the same way as the Great Eastern Main Line. The provenance argument that defends the inaugural corridor defends every corridor that follows it.

Where this breaks

Adopting an external delineation is the right trade for defensibility, but it is a trade, and the seams are worth stating plainly.

• Flows that do not map onto any single corridor. Not every freight working runs origin-to-destination along one named corridor. A service may run over a stretch of a corridor, leave it for a route the Freight Network Study did not name, and rejoin a different corridor later. The middle of such a journey is unattributed at the corridor level. We report the corridor segments we can identify and do not force the rest into a corridor it did not run over.

• Terminals served by branches off the named corridor. A corridor’s point sequence follows the main line. Terminals reached by a short branch off that main line — a quarry, a private siding, a distribution park a mile from the through route — are not on the corridor point list. A journey to such a terminal will match the corridor as far as the branch divergence and then leave the point sequence. The branch mileage is real but is not part of the corridor as delineated, so its performance is not counted in the corridor figures.

• TIPLOC matching gaps. Corridor assignment depends on TRUST reporting a location that appears in the corridor’s point sequence. When TRUST reports a location that is not in the sequence — a passing point, a recess loop, a STANOX that maps to a TIPLOC we have not included — that report contributes nothing to corridor matching. If reporting along a stretch of the corridor is sparse, the matched segment can be shorter than the journey actually ran. We treat persistent matching gaps as a register-maintenance task, not as a reason to widen the point list arbitrarily.

• The 2017 delineation ageing against current flows. The Freight Network Study describes the network as Network Rail saw it in April 2017. Freight flows move: a terminal opens, a flow re-routes, a port’s catchment shifts. Where current freight no longer matches the 2017 delineation, the corridor as published can lag the railway as run. We hold to the published delineation rather than quietly re-drawing it, because re-drawing is precisely the discretion the provenance argument forbids; material divergence is recorded and surfaced rather than absorbed silently into a moved boundary.

• Journeys traversing multiple corridors. A journey that runs over two named corridors is counted once in each corridor’s segment population, for the stretch it ran over that corridor. This is the intended behaviour, but it means the same train can appear in two corridor reports, and a reader who sums corridor populations will double-count any multi-corridor service. Corridor figures are per-segment, not per-train; they are not additive across corridors, and the journey-level total is the canonical count of the service itself.

Where a corridor boundary materially affects a published figure, the boundary’s provenance — the Freight Network Study delineation — is cited alongside the figure, so that any reader can trace the line back to the document that drew it.