How CAT 2025 Normalisation Works

CAT 2025 is conducted in three slots (morning, afternoon, evening). Each slot gets a different test form, so raw marks from one slot cannot be directly compared with raw marks from another. To ensure fairness, IIMs use a two-stage process:

• Stage 1: Convert raw scores to scaled scores using slot-wise normalisation.
• Stage 2: Convert scaled scores to percentiles using the official ranking formula.

The same method is applied separately for each section (VARC, DILR, Quant) and then for the overall score.

1. Slot-wise scaling: Raw score → Scaled score

For each slot, CAT computes the mean and standard deviation of raw scores in a section:

• Morning slot: mean = M₁, standard deviation = S₁, reference G₁ = M₁ + S₁
• Afternoon slot: mean = M₂, standard deviation = S₂, reference G₂ = M₂ + S₂
• Evening slot: mean = M₃, standard deviation = S₃, reference G₃ = M₃ + S₃

Across all three slots combined, CAT also calculates:

• Overall mean raw score = M
• Overall standard deviation = S
• Global reference point G = M + S

2. Top 0.1% anchor candidates

To align the difficulty of different slots, CAT uses the top 0.1% candidates as anchor points in each section:

• Morning slot: mean raw score of top 0.1% = M₁^0.1
• Afternoon slot: mean raw score of top 0.1% = M₂^0.1
• Evening slot: mean raw score of top 0.1% = M₃^0.1

For all slots together, CAT computes the overall top 0.1% mean: M^0.1 (combined).

3. Scaled score formula

Suppose a candidate in a given slot scores raw marks R in a section. Their scaled score R̂ is obtained by:

R̂ = (R − G_s) × ( M^0.1 − G ) / ( M_s^0.1 − G_s ) + G

where:

• G_s = slot-specific reference (G₁, G₂ or G₃)
• M_s^0.1 = top 0.1% mean for that slot
• M^0.1 = overall top 0.1% mean across all slots
• G = overall reference point (M + S)

Intuitively, this formula:

• Shifts your raw score relative to your slot’s difficulty (via G_s).
• Stretches or compresses it so that the top 0.1% of your slot align with the overall top 0.1%.
• Places everyone on a common CAT-wide scale (via G).

This gives a scaled score that is comparable across all three slots. The same process is repeated for VARC, DILR and Quant, and the overall scaled score is the sum of the three section scaled scores.

Note: If any scaled score exceeds 100, it is capped at 100 as per the official document.

4. Scaled score → Percentile

Once scaled scores are computed, CAT converts them into percentiles:

• Let N = total number of candidates who appeared for CAT 2025.
• Candidates are ranked in descending order of scaled scores (rank 1 = highest score).
• If a candidate has rank r in a section, their percentile P is:

P = ((N − r) / N) × 100

Percentiles are then rounded to two decimal places with special rules at the extreme top end (for example, values ≥ 99.995 are reported as 100.00). The same procedure is followed for each section and for the overall score.

About AzuCATion

AzuCATion is India’s most data-driven and student-centric CAT preparation ecosystem. Built on precision, transparency, and discipline, AzuCATion has guided 90,000+ aspirants toward top B-schools with a philosophy rooted in analytics and real student outcomes.

We believe that CAT preparation should not rely on guesswork or hearsay. Every mock analysis, percentile predictor, difficulty assessment, and strategy comes from data-backed modelling, large-scale student surveys, and RTI-sourced insights from IIMs. This commitment to evidence-based mentoring sets AzuCATion apart in the test-prep landscape.

Whether it is daily Quant practice, LRDI sets, VARC strategy, WAT–PI mentorship, or post-CAT decision guidance, the AzuCATion system ensures that every learner receives clarity, direction, and structure.

Our motto is simple:
“Right Data → Right Direction → Right Results.”

CAT 2025 Scaled Score Calculator

Note: We do NOT save or store your data.

Predictor Confidence & Model Accuracy

This predictor uses large-scale CAT 2025 response dataset and advanced AzuCATion statistical modelling for extremely high accuracy.

Last Update on 12 December, 08:30 AM - 1000+ Iterations with 1,93,243 Data Set (99%+ Accuracy in Extrapolation)

Section	Confidence Level	Model Type	Model Accuracy (R²)	Reliability
VARC	98%	Gen. Logistic	99.95%	Very High
DILR	98%	Bi-Logistic	99.93%	Ultra High
Quant	97.5%	Gen. Logistic	99.87%	High
Overall Percentile	98% confidence (for 90+ percentile range)	Bi-Logistic	99.98%	Perfect Fit

Slot Slot 1 Slot 2 Slot 3 VARC Raw Score DILR Raw Score Quant Raw Score Get Scaled Score → Percentile

Based on Data as on 12 December, 08:30 AM 1000+ Iterations with 1,93,243 Data Set (99%+ Accuracy in Extrapolation) | - Prev Updates 09 December, 12:30 PM | 7 December, 03:31AM & 01:00 AM |7 December, 12:05 AM | 6 December, 9:00 AM (AzuCATion Dataset)

Your Results

Slot -

Section	Raw	Scaled Score	%ile
Total

AzuCATion Courses for CAT 2025 WAT–PI

Choose the right level of support for your B-School Journey

To Get More Details or To Join → WhatsApp 990 5050 159