The evidence

What it shows. The first public record that our branch was dropped from the modernisation — and it dates the decision to a cost-and-schedule review in September 2020, prompted by pandemic finances. TfL’s own words leave the door open to converting the signalling later. (The same meeting’s Signalling & Controls paper, Item 12 §5.10, separately refers to the “4LM signalling descope areas”.)

What it shows. TfL restates the decision and names exactly where the new signalling stops: East Putney (our branch) and Stamford Brook (the Richmond and Ealing Broadway branches). The “continue to assess” wording is repeated word for word. The paper also weighs — and rejects, on cost grounds — giving the Metropolitan line the same treatment.

What it shows. TfL admits, in writing, that the branch’s train-tracking equipment is life-expired and prone to faults — the root of the daily board errors and delays. (This describes the train-tracking/information layer specifically, not the whole signalling system.)

What it shows. TfL confirms verbatim that the modern signalling reaches only as far as East Putney, and that the one improvement coming to the branch is better information boards — now due in the second half of 2026, having slipped from an earlier “early 2026”.

What it shows. TfL keeps a record of every service-status message it puts on the platform boards. Put its three releases together and you have an unbroken eleven-year account of the District line — and it shows our branch going one way while the rest of the line goes the other.

Counting the days on which TfL’s own records place a signal, points or track fault at a station on our branch: 14 days in 2015, 27 in 2023, 26 in 2024 — and 49 in 2025. Over the same period the equivalent figure for the whole District line fell, from around 150 days a year to around 120. Our branch’s share of the line’s fault days went from 9% to 37%. In 2025, four stations accounted for more than a third of the fault days on a line that runs from Upminster to Ealing Broadway. Southfields and East Putney account for most of that jump.

2026 so far is better. To 26 July there were 15 such days, which works out at about 26 across a full year — back to the 2023 and 2024 level. We are not claiming the branch is still getting worse. We are saying it has been consistently worse than the rest of the line for a decade, and that 2025 was the worst year on record.

TfL originally refused the data from March 2025 onwards under section 22 of the Freedom of Information Act, on the basis that it was intended for future publication. It published it on 28 August 2026, and we withdrew our challenge.

How to check this. These figures are our analysis of TfL’s raw records, not TfL’s own published statistics. TfL cautions that each row is a status message, not an incident, and that messages overlap — so we count distinct days on which a fault is named at a branch station, never message totals, and never hours. We ignore 2020 and 2021 when comparing periods, because pandemic timetables make them unlike other years. All three datasets are public: the 2015–2019 file is on the request thread below, 2020–2025 is on TfL’s disclosure log as FOI-3588-2425, and 2022–2026 is on TfL’s performance page.

Why this one matters. Every other set of numbers here is TfL reporting on itself. This is not. Network Rail attends signalling faults on part of our branch under contract to TfL, and it keeps its own log. That gives us a way to check TfL’s public account against somebody else’s record of the same equipment.

Network Rail logged 27 signalling faults on the branch between 2022 and April 2026 — 2 in 2022, rising to 10 in 2025. Its worst year is 2025, the same as TfL’s.

We checked all 27 dates against what TfL told passengers that day. On 9 of them TfL named a fault on the branch. On 7 the branch was disrupted but TfL gave a different reason. On 10 more, TfL’s record does not mention the branch at all — and on 10 November 2025, when Network Rail logged a points failure at Wimbledon Park, TfL’s record for the whole day contains nothing but the routine overnight notice.

What this does and does not prove. Network Rail gave us dates but no times. A fault logged on a given day might have happened after the branch closed for the night, or been fixed before any train was affected. So these are questions worth asking, not proof that anything was hidden — and we have asked Network Rail for the timings.