No opinions. Only formulas.
Every insight on Exam Neeti is produced by a five-stage analytical pipeline governed by predefined business rules — auditable, repeatable, and explainable.
Five stages from raw test to strategy.
Data Ingestion
Test responses, timings and answer keys are ingested per question — every choice, every second, every mark.
Metric Extraction
Raw responses are classified into correct, incorrect, skipped and rule-based guess attempts.
Normalization
Scores are normalized across difficulty layers and test formats so every comparison is apples-to-apples.
Pattern Detection
Business rules isolate recurring behaviours: late-test fatigue, guess clusters, time sinks, easy-question skips.
Insight Generation
Each pattern is converted into a quantified, prioritised finding with its exact mark value attached.
A look inside the formulas.
Simplified forms of the actual business rules the engine executes on every ingested test.
Attempt Classification Rule
guess = (time < 25s) ∧ (topic_accuracy < 40%)An attempt is flagged as a guess when answer time falls below the reading threshold and the student's historical accuracy in that topic is weak — no self-reporting required.
Time Utilization Index
TUI = Σ(t_optimal) / Σ(t_total) × 100The share of total test time spent inside the optimal 60–120 second band. Students above 75 TUI consistently outperform their accuracy-matched peers.
Negative Mark Attribution
loss_avoidable = Σ(neg | cause ∈ {silly, misread})Every negative mark is attributed to one of four root causes. Silly errors and misreads are classified avoidable — typically 60%+ of all losses.
Recoverable Marks Formula
RM = loss_avoidable + skips_easy×M + Δt_misalloc×accAvoidable losses, plus skipped questions from high-accuracy chapters at full marks, plus misallocated time re-priced at the student's real accuracy.
Deterministic, not probabilistic
The same test data always produces the same analysis. Every number can be traced to its rule.
Behaviour-first metrics
We measure decisions (guessing, pacing, skipping) — because decisions change faster than knowledge.
Marks as the universal unit
Every insight is priced in marks, so students always know what a fix is worth.
See the methodology run on your own test.
Upload one mock and watch the pipeline price every decision you made.