Day 1 versus Day 2 grades

Day 2 of a two-day EBT module is deliberately more demanding than Day 1: Evaluation diagnoses, scenario-based training develops under stretch. A raw grade dip between days can follow from that demand change and from grading process rather than outcome. It is not a programme goal, a proof of seriousness, or a reason to mark Day 2 harder by default.

#Position

#Day 1 diagnoses; Day 2 develops

An EBT module is two simulator sessions and four components:

Needs found in Evaluation are worked mainly in SBT on Day 2. That is the operational difference between traditional checking and EBT: Day 1 surfaces gaps; Day 2 develops them.

Difficulty: Day 2 is designed to be more demanding. SBT stretches capability in a low-jeopardy environment; assistance is legitimate when learning would otherwise stall. ICAO's module outline matches the asymmetry: EVAL captures realistic baseline; SBT is the largest phase and develops competencies against generation-critical threats and errors.

#Day 1 and Day 2 are not the same measurement

Comparing raw grade averages across days without context is a measurement error.

Day 1 (end-of-day overall)

• Rate all nine competencies at the end of Evaluation (first look, natural performance, simple line-oriented scenario).
• MV is rated per manoeuvre for grades 1, 2, or 5; end-of-Day-1 overall is based primarily on Evaluation when MV was at 2 or above.
• MV failure conditions EVAL grades for Flight Path Management (Manual) and/or Application of Procedures when those gaps appeared in MV.

Day 2 (end-of-session overall)

• Overall rating after SBT only. Do not include ISI.
• The crew already completed the check elements the previous day under lower difficulty.
• Weigh scenario demand, whether poor performance improved with assistance, and receptivity to learning.
• Repositioning and device artifacts induce errors that would not appear on the line; do not over-weight that noise.

SBT is graded on process, not outcome. How the crew applied competencies to solve the problem matters more than whether the scenario ended clean. What was learned and what they take away is the assessment surface.

So a Day-2 grade is not "the same score under a harder exam." It is higher demand, formative assistance allowed, process-weighted judgment, ISI excluded. A numerical dip can be legitimate without meaning the pilot got worse overnight.

#"Day 2 grades should be lower" conflicts with criterion-referenced grading

Grading is criterion-referenced testing (CRT) against word pictures and VENN grading (HOW WELL / HOW OFTEN / HOW MANY / OUTCOME), not against yesterday's numbers or the cohort. COMPETENT means all competencies ≥2. ADDITIONAL TRAINING REQUIRED means any competency at level 1, with the demand filter below. Consecutive 2s in the same competency across cycles trigger follow-up even when each session alone was COMPETENT.

If the cadre's model is "Day 2 grades should be lower," three distortions follow:

1. Normative pressure. Instructors mark harder on Day 2 to match the culture, independent of observed indicators.
2. Soft Day-1 calibration. Day 1 stays at default 3s and generous first-look ratings so Day 2 can look "serious" by contrast. That contaminates the diagnostic data the rest of the module depends on.
3. Outcome-weighted SBT. Scenarios that "go poorly" under high demand get punished even when process, recovery, and learning were strong.

All three damage Inter-rater reliability and the fleet-level data EBT uses to set training emphasis. Grades then describe instructor culture more than crew competency.

A mild average dip can be a side effect of higher demand and a different construct. It must not be a success criterion for the programme.

#The ATR filter already accounts for high demand

ADDITIONAL TRAINING REQUIRED is not "any level 1 anywhere." It may be considered only when level 1 is observed:

• under conditions that do not place a high demand on the crew;
• repeated at the same level on several occasions; and
• where learning is not being achieved.

An isolated level 1 under high demand may be a reactive failure, not a competency gap. Day-2 stretch is exactly where a single rough cell under high demand should not automatically become a module failure.

#Healthy progressive demand vs unhealthy pain curve

#Bottom line

Progressive difficulty is normal. Manufacturing or expecting a grade drop is a symptom that assessment culture has drifted from the grading rules.

#How to tell which state the company is in

Anecdote is not enough. Use evidence the training system already produces or can produce cheaply.

#Score distributions

• Plot per-competency grade distributions for Day-1 end-of-day versus Day-2 overall across many modules (same fleet, same season if possible).
• Healthy: modest shift toward 2s under competencies SBT was designed to stress; COMPETENT outcomes still dominate; level-1 rates rare and clustered where Day 1 already flagged need.
• Symptom: near-ceiling Day 1 (almost all 3–4–5) with sharp Day-2 collapse; or Day-2 level-1 rate that tracks scenario "messiness" rather than competency indicators.
• Intermediate "fleet performance" marks during SBT must not bleed into end-of-session grades.

#Construct alignment

• For each SBT scenario, list intended competency pressure from the lesson plan.
• Check whether low Day-2 grades land on those axes or on untaught surprise, repositioning artifacts, or pure outcome.
• If failures cluster on skills Day 1 never diagnosed and Day 2 never trained toward, the jump is curriculum failure, not rigor.

#Demand-aware failure logic

• Sample modules that ended ADDITIONAL TRAINING REQUIRED after Day 2.
• Was level 1 under low demand, repeated, with stalled learning? Or a single high-demand rough cell?
• High fraction of the latter means the cadre is not applying the demand filter.

#Inter-rater calibration

• Same recorded SBT video graded by multiple instructors.
• Compare Day-1 and Day-2 harshness by instructor.
• Systematic "I mark Day 2 one grade lower" without indicator evidence is cultural bias, not standard.

#Process versus outcome language in debriefs

• Process-aligned: decision sequence, tasksharing under concurrent threats, recovery after first error, receptivity when prompted.
• Outcome-aligned: "didn't land," "scenario fell apart," "messier than Day 1."
• Outcome-dominant language with thin process language breaches SBT grading rules.

#Learning transfer

• Day-1 needs (two or three focus areas from the examiner debrief) should reappear as SBT design and Day-2 debrief themes.
• If Day 2 is generically "harder" but uncoupled from Day-1 diagnostics, the module is two independent checks, not an EBT module.

#Environmental confounds

Scores are partly a function of environment: fatigue, time of day, examiner attitude, non-standard instructions. Day 2 often carries more cumulative fatigue. A grade drop that tracks roster load and night-sim slots more than competency indicators is measurement contamination, not proof of higher standards.

#Instructor use

1. Brief Day 2 as stretch training. Check elements are behind them; today develops process under higher demand; assistance is legitimate when learning would stall.
2. Grade SBT on process. Weigh improvement with assistance and receptivity. Do not let a messy scripted outcome alone drive a level 1. Exclude ISI.
3. Do not use Day 1 as the Day 2 bar. Same word pictures, different phase conditions (CRT).
4. Before ATR, apply the three-part filter. Level 1, low demand, repeated, learning not achieved. High-demand one-offs are stretch evidence first.
5. Keep Day 1 honest. Soft first-look grades to make Day 2 look serious destroy the diagnostic half of the module.
6. Debrief two or three competency areas. Not trivial SOP noise; not outcome shame.
7. In standardisation, pair Day-1 / Day-2 video grading so the cadre sees its own Day-2 harshness bias if any.
8. Stratify fleet metrics by phase/day. Unstratified averages misread SBT stretch as fleet decay.
9. Audit SBT against Day-1 needs. Generically harder scenarios that ignore the diagnostic themes are not progressive training.

#Connections

• Evidence-based training. Module phase structure that creates the Day 1 / Day 2 asymmetry.
• VENN grading. Four-dimension grades; Day 2 process-over-outcome rule; high-demand level-1 caution for ATR.
• Word pictures. Same grade language on Day 1 and Day 2; phase rules change, not the cells.
• Evaluation cycle. Criterion-referenced testing versus normative ranking; environmental confounds; cohort-wide failure as design signal.
• Inter-rater reliability. Phase mix-up and cultural Day-2 harshness as concordance failures.
• Facilitation. Primary Day-2 instructional technique; learning over check theater.
• Just culture. Low-jeopardy stretch depends on crews not managing image for a second check.
• Skill development model. Stretch under Monitor and facilitation develops proficiency; demand rise is not automatic remediation failure.
• Competency-based training. Integrated competence across contexts, not a single-day league table.
• Fault analysis. Root-cause reading of Day 2 breakdowns under high demand.
• From competency to observable behaviour. Observe–record–classify–evaluate chain that must respect phase rules.

#Sources

• A4.2.4 Grading Methodology for Recurrent Training and Checking. When to grade by phase; end Day 1 vs end Day 2; process not outcome on SBT; ATR demand filter; default level 3.
• A4.2.2 Guidance for Instructors. Day 2 converts Day-1 diagnostics into development; SBT stretch; overall rating excluding ISI; prior day lower difficulty; improvement with assistance and receptivity; focus two or three competency areas.
• A4.2.1 Guidance for Examiners. EVAL first look; needs mainly in Day 2 SBT; repositioning induces non-line errors.
• Doc 9995, Part I Ch 3. EVAL / MT / SBT module outline; SBT as largest development phase.
• Doc 9995, Part I Ch 7. Continuous assessment; criterion-referenced competence; facilitation debrief.
• 11.1 Purpose of Evaluation. Criterion-referenced testing: grade against standards, not peers or days.
• 11.3 The Evaluation Cycle. Cohort always fails same item; environment affects scores; close the loop on trainee, training, and instructor.
• 2.2 The Learning Process. Workload–performance relationship under rising demand.