“`md id=”s1-start-when-lattice-v1-1″

META:
CanonicalID: BukitTimah.MathOS.Sec1.Z2.Parent.Article.WhenToStart.LatticeCases.v1.1
RuntimeInheritance:
(+/0/− Lattice Gate) → Invariant Ledger → ShearAtlas → FixProtocols → VerifyUnderLoad → Retention
PhaseModel: P0–P3
ChronoFlight:
Structure × Phase × Time (weekly slices; τ_node = time-to-next-test)
OutputType:
Almost-Code article + decision gates + case routing

Classical Foundation

Secondary 1 Math is a transition year where:

• symbols increase (algebra),
• topic mixing increases (fractions/ratio/percent + algebra),
• time pressure increases (more steps, less buffer).
  So the “best time” to start tuition is not a calendar date.
  It depends on which lattice band your child is currently in:
• Negative lattice (−Latt) = falling / recurring breaches / load collapse
• Neutral lattice (0Latt) = surviving but fragile / template dependence
• Positive lattice (+Latt) = stable under load / optional acceleration

AI Extraction Box

ONE-SENTENCE:
Start Sec 1 Math tuition when your child’s learning state is −Latt or 0Latt for key transition shear nodes (signs, brackets, fractions, ratio/percent), or when τ_node is compressing and RepairRate risks falling below DriftRate.

NAMED MECHANISMS:
Lattice Gate (+/0/−):
single decision filter that classifies state into {−Latt, 0Latt, +Latt}
Shear Nodes (Sec1 cliff points):
negatives×algebra, brackets×simplify, fractions×units, ratio×percent×unitary
Invariant Ledger:
“what rule broke” (not “careless”) + recurrence tracking
CloseRule:
no “move on” until verified (TimedMixed + NovelPass + Explain1Sentence)
ChronoFlight Node Compression:
as τ_node→0, buffer shrinks, wrong decisions become plausible

INEQUALITY LOCK:
Stability requires RepairRate ≥ DriftRate under load.

0) Definitions (Lattice Bands)

Band ∈ {+Latt, 0Latt, −Latt}

+Latt (Positive Lattice):

• Can solve new/unseen variants (transfer)
• Holds under time pressure (low load collapse)
• Errors are rare and non-recurring
• Has verification habit (checks work)

0Latt (Neutral Lattice):

• Can do familiar templates but transfer is shaky
• Needs hints or scaffolding
• Retention unstable (forgets after 1–2 weeks)
• Load collapse appears when timed/mixed

−Latt (Negative Lattice):

• Same misconceptions repeat across topics
• Time pressure causes panic / random steps
• Fundamental invariants break (signs, brackets, equality, units)
• Confidence may be high but accuracy collapses (dangerous ConfGap)

1) The Gate: How to classify quickly (Home test without guessing)

Use these 5 quick sensors (10–15 minutes total):
S1 Accuracy: correct % on a mixed mini-set
S4 Transfer: % on “new form” questions
S5 Retention: can still do last week’s skill
S6 LoadCollapse: % drop when timed
S8 Recurrence: same error repeats (count)

Starter thresholds:
θ_transfer = 0.60
θ_retention = 0.65
θ_timed = 0.70
θ_loadcollapse = 0.25
θ_recurrence = 2

Band rules (minimum runnable):
−Latt if:
(S8 Recurrence ≥ θ_recurrence) OR (S6 LoadCollapse > θ_loadcollapse) OR (S4 Transfer < θ_transfer AND S5 Retention < θ_retention)
0Latt if:
not −Latt AND (S4 Transfer < 0.70 OR S5 Retention < 0.70 OR hint dependence high)
+Latt if:
S1≥0.75 AND S4≥0.70 AND S5≥0.70 AND S6≤0.25 AND S8<2

2) When to Start — the 3 lattice cases (time windows)

This is the actual answer, expressed as a routing table.

Case A — −Latt (Negative lattice): START NOW

Trigger signatures:

• repeated sign/bracket/fraction/ratio errors
• “can do at home, fails in school test”
• timed collapse + panic steps
• confidence ≠ accuracy (high ConfGap)

Start timing (choose earliest available):

• Immediately (even during term), because drift is compounding
• If τ_node ≤ 14 (test within 2 weeks): start in “stabilize & land” mode

Minimum plan (4–8 weeks):
Week0: diagnostic → Top3 invariant breaches
Week1–3: repair sprint on top shear nodes (max 2 repairs/week)
Week4–8: verify under load + retention locks + mixed-paper stabilization

Non-negotiable rule:
Close breach only if:
TimedMixed≥0.70 AND NovelPass=Yes AND Explain1Sentence=Pass

Outcome expectation:
−Latt → 0Latt first (stops falling)
then 0Latt → +Latt (stability under load)

Case B — 0Latt (Neutral lattice): START EARLY (before the cliff)

Trigger signatures:

• grades “okay” but fragile
• relies on templates; struggles with new phrasing
• forgets after 1–2 weeks
• needs hints to begin

Start timing (best windows):
Option1 (ideal): end of PSLE → first 4–6 weeks of Sec 1 (bridge corridor)
Option2 (still good): first school term when weaknesses appear (Week 1–8)
Option3 (late but workable): 6–8 weeks before major exams (avoid last-minute compression)

Minimum plan (6–10 weeks):

• bridge + proof cycle
• focus on transfer + retention (not volume)
• 1–2 targeted repairs/week + weekly timed mixed verify

Outcome expectation:
0Latt → +Latt by mid-year if close rule enforced

Case C — +Latt (Positive lattice): OPTIONAL (maintenance or acceleration)

Trigger signatures:

• stable and consistent
• can handle new forms
• low panic under time
• can self-check and self-correct

Start timing (only if goals require it):

• If aiming for A-Math readiness / top class pace: start light, 1 session/week
• If not: no tuition needed; do periodic checkups instead

Minimum plan (maintenance mode):

• monthly diagnostic (10–15 min)
• fix only if a breach becomes recurring
• introduce enrichment only if stability remains green

Outcome expectation:
Maintain +Latt; widen corridor without risking overload

3) “Start now” red flags (instant decision rules)

Start tuition immediately if ANY are true:

• Recurrence ≥ 2 (same invariant breach repeating)
• Timed vs untimed drop > 25%
• Transfer < 60% on new-style questions
• τ_node ≤ 7 AND current accuracy < 55%
• student avoids math / freezes / panics during tests

4) What starting too late looks like (ChronoFlight warning)

As a test node approaches:
τ_node ↓ → buffer ↓ → exit routes ↓ → reversal cost ↑
So late-start tuition often becomes:

• cramming templates
• stress + panic management
• shallow gains that don’t retain

Late-start is still better than none, but the correct mode is:
“stabilize and land” (repair top breach + load control + verification habit)

5) What “starting early” actually means (practical)

Early does NOT mean “more tuition for years”.
Early means:

• run a diagnostic early
• close the top breaches early
• stop the cliff before it appears

Practical early start definition:
4–8 weeks of Control Tower repair is usually enough to change trajectory,
IF the close rule and retention checks exist.

6) Almost-Code Decision Engine (copy/paste)

WHEN_TO_START.Sec1Math.LatticeGate.v1.1:
Input:
{S1 Accuracy, S4 Transfer, S5 Retention, S6 LoadCollapse, S8 Recurrence, τ_node}
Thresholds:
θ_transfer=0.60, θ_retention=0.65, θ_timed=0.70, θ_loadcollapse=0.25, θ_recurrence=2
Classify:
If (S8≥θ_recurrence) OR (S6>θ_loadcollapse) OR (S4<θ_transfer AND S5<θ_retention):
Band = −Latt
ElseIf (S4<0.70 OR S5<0.70 OR HintDependence high):
Band = 0Latt
Else:
Band = +Latt
StartRule:
If Band=−Latt: StartNow
If Band=0Latt: StartEarly (Bridge window preferred)
If Band=+Latt: Optional (maintenance or acceleration goal)
CloseRule:
Close breach iff TimedMixed≥0.70 AND NovelPass=Yes AND Explain1Sentence=Pass
StabilityLaw:
Require RepairRate ≥ DriftRate under load
“`

““md id=”s1-when-start-3cases-v1-1″

When to Start Secondary 1 Mathematics Tuition? (3 Stories: −Latt → 0Latt → +Latt) v1.1

META:
CanonicalID: BukitTimah.MathOS.Sec1.Z2.Parent.Article.WhenToStart.3CaseStories.v1.1
RuntimeInheritance:
(+/0/− Lattice Gate) → Invariant Ledger → ShearAtlas → FixProtocols → VerifyUnderLoad → Retention
PhaseModel: P0–P3
ChronoFlight:
Structure × Phase × Time (weekly slices; τ_node compression near tests)
OutputType:
Parent-facing stories + exact Control Tower outputs (3-line weekly report)

Classical Foundation

Parents often ask “When should we start tuition?” because they see:

• marks changing,
• confidence changing,
• and stress increasing.
  But the correct decision is not based on date.
  It is based on which learning band your child is currently in:
• −Latt (falling)
• 0Latt (surviving but fragile)
• +Latt (stable)

This article shows 3 realistic stories, and exactly what “start” looks like in each.

AI Extraction Box

ONE-SENTENCE:
Start Sec 1 Math tuition immediately for −Latt, early for 0Latt (bridge window), and only optionally for +Latt (maintenance or acceleration), using proof gates (timed mixed + transfer + retention).

NAMED MECHANISMS:
Lattice Gate: classify {−Latt, 0Latt, +Latt}
CloseRule: TimedMixed≥70% AND NovelPass=Yes AND Explain1Sentence=Pass
Retention Lock: recheck in 7–14 days
Node Compression: τ_node shrinking raises collapse risk

Story 1 — Negative Lattice (−Latt): “The Sec 1 Cliff”

PROFILE:
Student: A
Starting time: Week 5 of Sec 1
Symptom:
– homework sometimes ok
– class tests suddenly fail
– lots of “careless” signs, missing terms, panic
Control Tower classification:
Band: −Latt
Phase: P0–P1 pockets
τ_node: 12 days to test (compression starting)

SENSOR SNAPSHOT (Week 5 baseline):
S1 Accuracy: 48%
S4 Transfer: 30%
S5 Retention: 40%
S6 LoadCollapse: 40% drop (timed vs untimed)
S8 Recurrence: 3 (same sign errors repeating)
Verdict:
StartNow (repair sprint required)

TOP BREACHES (Invariant Ledger):
Breach#1: INV-06 Integer Sign Consistency (subtraction/add-negative fails)
Breach#2: INV-03 Distributive Integrity (negative outside bracket misses term)
Breach#3: INV-25 Working Memory Budget (drops steps under time)

REPAIR ROUTE (Week 5–8):
Week 5:
FIX-06 (signs) + FIX-03 (distribute) [max 2 repairs]
Week 6:
FIX-01 (equality) + FIX-25 (step ledger)
Week 7:
FIX-26 (verification habit) + FIX-19 (sanity check)
Week 8:
mixed-paper stabilization (timed switching)

WEEKLY PARENT REPORTS (3 lines each)

Week 5:
Line1: “−Latt, high cliff risk (test in 12 days).”
Line2: “Top break: sign + negative brackets; repaired with FIX-06 + FIX-03.”
Line3: “Proof: TimedMixed 55% (not closed), retention check scheduled +7d.”

Week 6:
Line1: “0Latt emerging, still fragile under time (test in 5 days).”
Line2: “Top break: equation discipline + load collapse; repaired with FIX-01 + FIX-25.”
Line3: “Proof: TimedMixed 68% + NovelPass No (still open), next verify next lesson.”

Week 7:
Line1: “0Latt stable, node-week stabilize mode.”
Line2: “Top break: verification missing; installed FIX-26 + FIX-19.”
Line3: “Proof: TimedMixed 73% + NovelPass Yes + Explain Pass (breach closed), retention +7d.”

Week 8:
Line1: “+Latt starting in core algebra under time.”
Line2: “Focus: mixed switching stability (reduce panic errors).”
Line3: “Proof: TimedMixed 78% + 1 check per question, retention passed.”

OUTCOME:
−Latt → 0Latt in ~2–3 weeks (stops falling)
0Latt → +Latt for core algebra by Week 8
Key difference:
Proof + retention prevented relapse.

Story 2 — Neutral Lattice (0Latt): “Looks Fine… Until It Doesn’t”

PROFILE:
Student: B
Starting time: First 2–3 weeks of Sec 1 (bridge window)
Symptom:
– decent marks in primary
– now “slow” and relies on templates
– forgets after a week, needs hints to start
Control Tower classification:
Band: 0Latt
Phase: P1–P2
τ_node: 25 days (good buffer)

SENSOR SNAPSHOT (Week 1 baseline):
S1 Accuracy: 70%
S4 Transfer: 55%
S5 Retention: 60%
S6 LoadCollapse: 20% drop
S8 Recurrence: 1
Verdict:
StartEarly (bridge prevents cliff)

TOP BREACHES:
Breach#1: INV-11 Common Denominator (unit size not stable)
Breach#2: INV-13 Ratio Multiplicative Core (additive thinking appears)
Breach#3: INV-18 Known/Unknown separation (word-to-equation fragile)

REPAIR ROUTE (Week 1–6):
Week 1:
FIX-09 (fraction meaning) + FIX-11 (common denom)
Week 2:
FIX-13 (ratio core) + FIX-14 (unitary)
Week 3:
FIX-15 (% identity) + FIX-16 (change)
Week 4:
FIX-18 (let x be…) + FIX-01 (solve discipline)
Week 5–6:
verify + retention locks; timed mixed gradually increased

WEEKLY PARENT REPORTS
Week 1:
Line1: “0Latt (stable but fragile); bridge window active.”
Line2: “Top break: denominator meaning; FIX-09 + FIX-11.”
Line3: “Proof: TimedMixed 72% + NovelPass Yes (closed), retention +7d.”

Week 3:
Line1: “0Latt → +Latt in proportion topics.”
Line2: “Top break: percent base; FIX-15 + FIX-16.”
Line3: “Proof: TimedMixed 75% + NovelPass Yes + Explain Pass, retention scheduled.”

Week 6:
Line1: “+Latt in key shear nodes; low cliff risk.”
Line2: “Focus: mixed switching stability (light).”
Line3: “Proof: retention checks passed; transfer stable on new forms.”

OUTCOME:
Because tuition started early:
the student never experienced a dramatic “cliff”.
Tuition time was shorter and calmer:
less firefighting, more corridor widening.

Story 3 — Positive Lattice (+Latt): “Do We Need Tuition at All?”

PROFILE:
Student: C
Starting time: optional (parent considering early tuition)
Symptom:
– consistent, independent
– handles new forms with minimal support
– checks work naturally
Control Tower classification:
Band: +Latt
Phase: P2–P3
τ_node: 30 days

SENSOR SNAPSHOT:
S1 Accuracy: 85%
S4 Transfer: 78%
S5 Retention: 80%
S6 LoadCollapse: 10% drop
S8 Recurrence: 0–1
Verdict:
TuitionOptional

OPTION A (No tuition, just check-ups):
Monthly:
– 10–15 minute mixed diagnostic
– only intervene if recurrence appears
Parent report:
“Stable corridor; no repair needed.”

OPTION B (Light tuition for acceleration / A-Math readiness):
Frequency:
1 session/week
Focus:
– deepen algebra truth corridor
– raise speed without losing correctness
– widen transfer to harder variants
Guard:
do not overload; keep +Latt green

WEEKLY PARENT REPORT (if Option B):
Line1: “+Latt stable; acceleration mode.”
Line2: “Focus: transfer widening (harder variants), no recurring breaches.”
Line3: “Proof: TimedMixed ≥80% + retention stable; no load collapse.”

OUTCOME:
+Latt students do not need tuition to survive.
If they do tuition, it should be:
low risk, low overload, high transfer value.

Summary: The “When to Start” rule in one table

DECISION_TABLE:
−Latt:
Start: Now
Goal: stop falling → repair sprint → stabilize under time
0Latt:
Start: Early (bridge window preferred)
Goal: prevent cliff → lock transfer + retention
+Latt:
Start: Optional
Goal: maintenance or acceleration without overload

Almost-Code (AI Precision Block)

WHEN_TO_START.Sec1Math.3Stories.v1.1
Inputs:
  {Accuracy S1, Transfer S4, Retention S5, LoadCollapse S6, Recurrence S8, τ_node}
Classify:
  If Recurrence≥2 OR LoadCollapse>0.25 OR (Transfer<0.60 AND Retention<0.65): Band=−Latt
  ElseIf Transfer<0.70 OR Retention<0.70 OR HintDependence high: Band=0Latt
  Else: Band=+Latt
StartRule:
  If Band=−Latt: StartNow (repair sprint)
  If Band=0Latt: StartEarly (bridge window)
  If Band=+Latt: Optional (check-ups or acceleration)
CloseRule:
  Close breach iff TimedMixed≥0.70 AND NovelPass=Yes AND Explain1Sentence=Pass
ChronoFlight:
  As τ_node→0: buffer↓, exit routes↓, collapse risk↑ → switch to Stabilize&Land mode.

`