Mathematics Training (eduKateSG) — Fence Learning Systems for Math (Primary → Secondary)

TITLE: Mathematics Training (eduKateSG) — Fence Learning Systems for Math (Primary → Secondary)
SUBTITLE: Reach P3 reliability under load: method selection + accuracy + speed + transfer
VERSION: CivOS Unified Spec v1.x
MODE: Almost-Code / WordPress paste-ready
SCOPE: Z0 learner + Z1 parent + Z2 tutor/school; PSLE Math, Sec E-Math, Sec A-Math spine
GOAL: Math becomes reliable execution under time pressure (not “can do it when guided”)
========================================================
0) DEFINITION LOCK BOX
========================================================
Math mastery = reliable correct execution under load.
Math.P3 :=
A learner can:
- choose the right method quickly,
- execute steps accurately with low careless rate,
- finish within time,
- transfer to unfamiliar question forms,
- and remain stable under stress.
Fence principle:
Math collapses when:
- prerequisites are missing,
- method selection is weak,
- false competence hides gaps,
- careless drift leaks marks,
- time pressure triggers panic,
and TTC shrinks.
So math training must be instrumented, fenced, repaired, and verified.
========================================================
1) THE MATH STACK (WHAT MATH IS MADE OF)
========================================================
Math = 4 sub-systems:
M1 Prerequisite Spine (foundations)
- number sense, fractions/ratios, algebra basics, indices/surds/logs (as level rises)
M2 Method Selection (recognition)
- identify chapter/tool from problem cues
M3 Execution Accuracy (procedure)
- step discipline, arithmetic reliability, algebra manipulation
M4 Load & Transfer Reliability
- timed performance + mixed-form transfer (inversion)
If any subsystem fails, the whole paper collapses.
========================================================
2) UNIVERSAL DASHBOARD (MATH EDITION)
========================================================
Metrics:
- TTC: days to next test/exam
- Buffer: sleep/routine/stress slack (math is sensitive to cognitive bandwidth)
- R: are weak nodes shrinking weekly?
- Θ: do we fence within 24–72h after a drop?
- Λ: can routines/boundaries be enforced same day?
- Coupling Index (secondary recommended): does one gap break many topics?
Math-specific sensors:
- first-step latency (seconds to start)
- method selection accuracy (right tool? Y/N)
- careless error rate (% marks lost)
- stuck-time episodes (per paper)
- marks-per-minute (timed throughput)
========================================================
3) Z0 ROUTER: FAST DIAGNOSIS (10 MINUTES)
========================================================
Step 1: Identify the failing lane:
- prerequisite spine (M1)?
- method selection (M2)?
- execution accuracy (M3)?
- load/transfer (M4)?
Step 2: Classify dominant failure mode (pick ONE):
F1 Missing prerequisite (cannot start / repeated same wrong step)
F2 Method confusion (wrong chapter/tool)
F3 False competence (works with guidance, fails on new)
F4 Careless drift (slips dominate losses)
F5 Time collapse (too slow, panic, blanks)
F6 Word problem translation failure (text → math model)
F7 Multi-step coupling collapse (secondary/A-Math especially)
Step 3: Fence (truncate):
- shrink to smallest failing node
- apply stop-loss timers
- rebuild buffer if low
Step 4: Stitch (repair loop):
- 4–7 day targeted loop
- verify with timed mini-set + inversion set
========================================================
4) FAILURE MODES → SENSORS → REPAIR LOOPS
========================================================
--------------------------------------------------------
F1 Missing Prerequisite (Spine hole)
--------------------------------------------------------
Sensors:
- cannot start without help
- same error repeats across questions
Repair Loop MR1 (Micro-Spine Repair):
- isolate the micro-skill
- 10 micro-drills + 2 mixed questions
- spaced retest next day
Fence:
- stop new topics until this node stabilizes (R<1)
--------------------------------------------------------
F2 Method Confusion (Recognition failure)
--------------------------------------------------------
Sensors:
- uses wrong formula/chapter
- “I don’t know which one to use”
Repair Loop MR2 (Method Cue Cards + Mixed Sorting):
- sort 20 questions by method (no solving first)
- then solve 8 mixed questions timed
Fence:
- stop-loss: if unsure > 90s → mark, skip, return after solving others
--------------------------------------------------------
F3 False Competence (Guided illusion)
--------------------------------------------------------
Sensors:
- can do workbook examples; fails on fresh exam questions
Repair Loop MR3 (Inversion Practice):
- remove scaffolds
- randomize forms
- explain 1-sentence “why this method”
Fence:
- no new sets until transfer score improves (R<1)
--------------------------------------------------------
F4 Careless Drift (Execution leakage)
--------------------------------------------------------
Sensors:
- high marks lost from slips
- errors are not conceptual
Repair Loop MR4 (Error Taxonomy + Checklist):
- tag every mistake:
  sign / copy / arithmetic / last step / units
- “forced check” routine:
  last line check + re-substitute + estimate reasonableness
Fence:
- reduce speed demands temporarily; restore accuracy first
--------------------------------------------------------
F5 Time Collapse (Speed + panic)
--------------------------------------------------------
Sensors:
- cannot finish paper
- accuracy drops sharply when timed
Repair Loop MR5 (Timed Micro-Sets / Marks-per-Minute):
- 3 questions / 6 min
- 5 / 10
- then mini-section
- train decision speed: “start within 30 seconds”
Fence:
- “move-on fence”: if stuck 2–3 minutes, skip and return
- protect buffer (sleep) before pushing timed work
--------------------------------------------------------
F6 Word Problem Translation Failure (Modeling)
--------------------------------------------------------
Sensors:
- cannot convert text to equations/diagrams
Repair Loop MR6 (Model Template):
- Given / Find / Let / Relationship / Equation / Check
- draw bar model (primary) or define variables (secondary)
Fence:
- do modeling-only drills first (no full solving) until stable
--------------------------------------------------------
F7 Multi-step Coupling Collapse (Secondary / A-Math)
--------------------------------------------------------
Sensors:
- one early slip destroys entire question
- loses confidence mid-problem
Repair Loop MR7 (Checkpointing + Step Discipline):
- mandatory checkpoints every 2–3 lines
- verify intermediate results (factorization, substitution, units)
- redo hardest questions until “survival rate” improves
Fence:
- truncate: master 3 high-weight patterns first; then expand
========================================================
5) PRIMARY MATH (PSLE) — FAST CLUSTERS
========================================================
High-yield clusters:
P1 Fractions / Ratio / Percent
P2 Speed / Time / Distance (incl average speed basics)
P3 Whole numbers + heuristics (estimation, patterns)
P4 Geometry + measurement (area/perimeter/volume basics)
P5 Word problem modeling (bar model)
Rule (PSLE):
Most collapse is F6 (modeling) + F5 (time) + F4 (careless).
So always train:
- modeling template
- timed micro-sets
- error checklist
========================================================
6) SECONDARY E-MATH — SPINE EMPHASIS
========================================================
E-Math spine nodes:
- algebra manipulation
- linear equations/inequalities
- graphs and interpretation
- geometry reasoning
- probability/statistics basics
Rule:
If algebra spine is weak, everything breaks.
So F1/F4 repairs dominate early.
========================================================
7) SECONDARY A-MATH — THE “ALGEBRA SPINE FIRST” LAW
========================================================
A-Math collapses mainly from:
- algebra spine holes (factorization, indices, surds, logs)
- function understanding
- calculus method selection + step discipline
A-Math minimal order:
- Algebra spine → Functions → Differentiation → Applications → Integration
Fence rule:
Never expand calculus volume while algebra spine is unstable (R>1).
========================================================
8) Z1 PARENT TOOLKIT (MATH)
========================================================
Parent role = enforce fences (Λ<1) and protect buffers; not teach solutions.
Home fences:
- fixed daily slot (20–40 min)
- device boundary
- insist on:
  - “start within 30 seconds”
  - show workings (step discipline)
  - check using checklist
- sleep fence before tests
Low-effort prompts:
- “What is given?”
- “What is asked?”
- “Which chapter is this?”
- “What’s your check?”
========================================================
9) Z2 TUTOR/SCHOOL ENGINE (TRANSFER RELIABILITY)
========================================================
Weekly cycle:
- diagnose dominant failure mode
- 4 days targeted repair loop (MR1–MR7)
- 2 timed micro-sets/week
- 1 mixed inversion set/week
- post-mortem redo of weakest questions
- simulation every 2–4 weeks (weekly as TTC shrinks)
Non-negotiables:
- inversion sets (kills false competence)
- timed training (kills load collapse)
- error taxonomy (kills repeat leaks)
- fast retest cycles (stitching)
========================================================
10) MINIMUM VIABLE 14-DAY PLAN (TTC-RED)
========================================================
Day 1–3:
- MR2 method sorting + MR4 error checklist (stabilize quickly)
Day 4–7:
- MR1 micro-spine repair of weakest node + spaced retest
Day 8–10:
- MR5 timed micro-sets + move-on fence
Day 11–14:
- 2 mixed inversion sets (MR3) + mini-sim + redo protocol
Rule:
Do not expand scope; shrink weakness until R<1.
========================================================
11) FAILURE MODE TRACE (REQUIRED)
========================================================
Trace:
prerequisite gap + method confusion + TTC short
→ time collapse → careless leaks rise
→ R>1 (weak nodes grow)
→ late reaction (Θ>1) + unenforceable routine (Λ>1)
→ P2→P0 near exam
Fence Repair:
detect TTC-RED → truncate to MR1+MR2+MR5+MR4
→ enforce routine same day (Λ<1)
→ fence within 24h (Θ<1)
→ stitch with inversion + retest → R<1 → restore P3 reliability
========================================================
12) WORDPRESS PASTE-READY STRUCTURE
========================================================
H1: Mathematics Training: How to Reach P3 Reliability Under Time Pressure (Fence Learning Systems)
H2: What Math Mastery Really Is (Method + Accuracy + Speed + Transfer)
H2: The Math Stack (Spine, Method, Execution, Load)
H2: Dashboard (TTC, Buffer, R, Θ, Λ) + Math sensors
H2: Fast Diagnosis (pick the failing lane)
H2: Failure Modes → Repair Loops (MR1–MR7)
H2: PSLE Math Clusters (what to fix first)
H2: Secondary E-Math Spine
H2: Secondary A-Math “Algebra Spine First” Law
H2: Parent Toolkit (Z1)
H2: Tutor/School Engine (Z2)
H2: 14-Day TTC-Red Plan
H2: Failure Mode Trace (Short)
H2: Copy/Paste Weekly Plan
========================================================
13) COPY/PASTE WEEKLY PLAN (MATH)
========================================================
Daily (25–45 min):
- 10–15 min micro-spine drills (MR1) OR method sorting (MR2)
- 10 min timed micro-set (MR5) (2–3 times/week minimum)
- 5 min error log + checklist habit (MR4)
- 1 word problem modeling drill (MR6) (primary emphasis)
Weekly:
- 1 mixed inversion set (MR3)
- redo 5 hardest questions until survival rate improves
- dashboard review (TTC/Buffer/R/Θ/Λ)
END SPEC
========================================================