Careless-Error Elimination

At the top, you don't have a knowledge problem — you have an error-rate problem, and careless errors are predictable signatures, not bad luck. Above ~1450, the content is mostly known; what stands between you and 1600 is a handful of avoidable slips per test. One wrong scored question can drop a section ~10–20 points, so driving the careless rate toward zero is the strategy for the stretch goal.

Sources: UWorld (800 guides), MySatCoach (plateau), Catalyst. Point convention: each item = 1 question; a Module-1 slip is a ceiling-risk loss (System Map).

---

#THE CORE INSIGHT

Careless errors repeat. The same student misses the same types — solving for x instead of x+1, missing "EXCEPT/LEAST," dropping a negative, a transition whose logic actually reverses. Because they repeat, they are trainable: catalogue your signatures, then run a targeted check that catches exactly those.

---

#THE FOUR HABITS THAT KILL CARELESS ERRORS

1. Annotate the ask. Before solving, underline/box what the question actually wants — the variable, "greatest/least," units, rounding, "EXCEPT." Why: the #1 careless loss is solving correctly and answering the wrong thing. [saves ~1–3 questions/test]

2. Slow down on the easy/early ones. The questions you "already know" are where slips happen, and in Module 1 they're ceiling-risk. Deliberately add 3–5 seconds to confirm the ask on routine items. Why: rushing easy items to "save time for hard ones" trades cheap points for expensive ones.

3. Re-verify, don't re-feel. Math: plug your answer back in or re-graph in Desmos. R&W: point to the text that proves it. Why: confidence is not correctness; proof is.

4. Run your personal checklist on flagged items. Your own recurring signatures (from the Error Log), checked before submitting anything you marked. Why: a generic "check your work" does nothing; checking your specific failure modes catches them.

---

#BUILD YOUR PERSONAL CARELESS CHECKLIST

From your Error Log, list every miss you tagged careless or misread. Cluster into signatures. A typical starter list (replace with yours):

MATH                          R&W
☐ solved for the asked value?  ☐ answered the exact question (not a close-but-wrong idea)?
☐ negative sign kept?          ☐ "EXCEPT / LEAST / NOT" caught?
☐ units / rounding as asked?   ☐ transition logic direction right?
☐ both roots / right denom?    ☐ answer provable from the text?
☐ re-entered into Desmos?      ☐ didn't change a right answer on review?

Memorise it; run it on every flagged question. Update it weekly from new misses.

---

#SAME ITEM, TWO LEVELS (MERIDIAN-original)

Question asks: "Which of the following is NOT supported by the passage?"

• ~700 (trap): finds a supported statement and picks it — read "which is supported." Trap: missed "NOT." [−1; pure giveaway]
• 800 (precise): boxed "NOT" while reading the stem; eliminated the three supported options; picked the unsupported one. Move: annotate-the-ask caught the negation.

---

#DRILLS (self-marked)

Drill 1 — Ask-annotation (every practice question, 1 week): physically mark the ask before solving. PASS: zero "answered-wrong-quantity" misses in your next full section. Drill 2 — Checklist pre-submit (10 flagged items): run your personal checklist on each before confirming. PASS: the checklist catches ≥1 real slip you'd otherwise have submitted. Drill 3 — Easy-item slowdown: on a timed set, deliberately confirm the ask on every "easy" item. PASS: accuracy on easy items = 100%. Repetition: any careless miss → add its signature to the checklist; re-run drills.

---

#TRANSFER

Abstract rule: errors are signatures; catalogue them, then check for exactly those. Works across both sections and any topic. Constant: annotate-ask, slow-easy, re-verify, personal-checklist. Changes: which signatures are on your list (and the list shrinks as you train).

#REFERENCE CARD — CARELESS-ERROR

1 ANNOTATE THE ASK (box: variable · EXCEPT/LEAST · units · rounding)
2 SLOW DOWN ON EASY/EARLY (Module 1 = ceiling risk)
3 RE-VERIFY (Math: plug back / Desmos · R&W: prove from text)
4 RUN YOUR personal checklist on flagged items · update it weekly from the Error Log
One slip ≈ −10–20 scaled. The stretch to 1600 = error rate → 0.

---

MERIDIAN™ · © 2026 · For Neo's personal study use. Not affiliated with or endorsed by the College Board. SAT® is a trademark of the College Board.

MERIDIAN · FORGE · CARELESS-ERROR ELIMINATION | v1.0