Railway AFA LDCE Pass Marks: 90 in the CBT, 30 in Viva and Record

To reach the AFA/ADFM panel through the 30% LDCE you clear four bars in order: 90 of 150 in the CBT (three fifths of that for SC/ST, 54 by our arithmetic); a medical fitness test; 30 of 50 in viva and record of service together (18 for SC/ST); and, inside that, at least 15 of 25 from record of service. Then merit decides who takes the seats. East Coast Railway published its full 2026 AFA marks table, so this guide also shows what those bars looked like in a real cohort — as counts only, with no names or roll numbers.

Board letter of 28.06.2022; ECoR notification of 23.12.2025; RBE 69/2026.

The record-of-service floor is the rule people forget. The viva and record carry 25 marks each and the pass is on the total, but 15 of the 30 must come from record of service. A perfect viva cannot carry a weak record. Record of service is assessed from your APARs — ECoR asked controlling officers to keep the last five (2018-19 to 2022-23) ready — and your service record. IREM grades record-of-service scores of 80% and above as "Outstanding" and 60% to 79% as "Good", and says there is no moderation.

East Coast Railway published RRB/Ajmer's "Declaration of Results" for the AFA post: each candidate's final mark and normalised score out of 150. We counted it. 69 candidates are listed (60 UR, 9 SC, no PwBD); 50 were present and 19 absent. 7 reached a normalised 90 (5 UR, 2 SC). All seven were empanelled on 21.08.2026 against 8 seats; the ST seat was not filled.

Two readings of this table, both ours. First, the viva list matches the SC/ST rule exactly: the five UR candidates called are the five UR candidates at 90 or more, and the six SC candidates called are every SC candidate at 54 or more. The letter itself does not state the basis, and six SC calls is more than five times the one SC vacancy. Second, the panel order is not the CBT order: an SC candidate with a higher CBT score than some UR empanelees sits on the SC seat, and panel positions 4 to 6 do not follow the normalised scores — the viva and record of service marks, which ECoR did not publish for AFA, account for the difference.

The 2025-26 CBT ran on two dates, 24.05.2026 and 31.05.2026, and the Board's letter of 02.04.2026 adopted normalisation "to ensure parity". The AFA table carries two roll series, "A" (22 present) and "C" (28 present); it does not say which sat on which date. In series C the normalised score is the final mark scaled up by a factor between 150/149 and 150/147 — consistent with deleted questions being removed and the mark rescaled to 150, as the AEN table shows in detail. In series A the published score moved from the final mark by between −1.39 and +3.62 marks. The one candidate whose final mark of 91.33 became 89.94 was in series A. All figures are our arithmetic over the official table.

No normalisation method was published. A widely read exam-analysis site analysed normalisation across 35,188 candidates for AOM, ACM, AME, AEE, APO, AMM, AEN and ASTE — AFA was not in its analysis (reported). The detailed mechanics, from a zone that published every column, are in our AEN guide on deletions, rescaling and normalisation. The lesson for AFA is the same: within about four marks of 90, do not count on either side of the line until your zone publishes the normalised score.

RBE 69/2026 (Board letter No. 2022-E(SCT)I/25/14 of 31.08.2026) says that in the 30% LDCE SC/ST candidates' written qualifying marks "will continue to be" three fifths of the general standard, and that in viva and record of service they need 18 in place of 30, still with 15 in record of service. The phrase "will continue" means the rule was not new in 2026, which fits the ECoR viva call of 06.08.2026 made before the letter (our reading). The letter concerns non-safety categories and does not list posts; treating AFA, a finance post, as non-safety is our reading. A candidate who qualifies on general standards without any relaxation is placed against an unreserved seat, exactly as happened to one SC candidate on the ECoR panel.