What separates 75% RC from 95% RC
The bottom-tier mistake on hard RC sections is reading slowly and not predicting. The middle-tier mistake is reading fast and not noticing trap options. The top tier — the 99%-ile cohort — does two specific things:
- They tag every paragraph with a one-word stance ("agree", "qualify", "rebut", "extend"). Not summary. Stance.
- They predict the answer before looking at options. The options are a confidence check, not a search.
Trap option taxonomy
Every CAT and UPSC RC question has 4 options. One is right. The other three are constructed from a small, finite set of traps:
- The shifted-scope trap — true of part of the passage, presented as if true of all of it.
- The reversed-causality trap — author says A causes B, option says B causes A.
- The plausible-but-unsupported trap — sounds reasonable to a generalist; the passage never actually argues it.
- The half-right trap — first half is exact, second half is wrong. Most missed-by-1 errors live here.
Once you can name the trap, you can mark options with S/R/P/H in 5 seconds and almost always have one option left.
The 4-pass method (works on dense passages)
- Pass 1 (30s): Read the first sentence of each paragraph. Tag the stance. Skip the rest.
- Pass 2 (90s): Read the full passage at normal speed. Don''t re-read. Trust your tags.
- Pass 3 (per question): Read the question. Predict the answer in your own words. Don''t look at options.
- Pass 4 (per question): Match your prediction to options. Eliminate trap options.
This isn''t faster than reading start-to-finish. It''s about the same. But it produces a 15-20 percentage-point lift in accuracy on hard inference questions in our internal A/B with CAT mocks.
How FocusRead drills this
Inside the app, the Edge track runs you through real CAT/UPSC RC passages with auto-graded inference questions and trap-option breakdowns after every quiz. The pass system above is built into the reader UI — paragraph-level stance tagging is one tap.
You don''t need an app for any of this. You can do it on PDFs. But you''ll need someone to grade your inferences, and that''s the unglamorous part our reviewers do for you.