Starting Mathematics Support | State, Evidence & Next Action

Mathematics support should not begin by teaching. It should begin by finding out what kind of situation it has been asked to enter.

A learner may arrive with a low mark, a specific question, a vague feeling of confusion, an examination paper, a request for more practice, a parent concern, a transition into a new stage, or simply the question: “What should I do next?”

If the system responds to all of these with the same action, it is not yet operating as a learning system. It is only serving content.

The Starting Mathematics Support is the public startup sequence for Bukit Timah Tutor’s Mathematics Learning System. The starting sequence is deliberately simple: establish the learner’s state, gather enough evidence, choose the right route and provide no more help than the situation justifies.

The connected routes are shown in the BTT Mathematics System Map.

Starting support does not mean “start explaining”. It means “establish the state, choose the route, and preserve the learner’s control.”

The starting sequence

ENTER → STATE → EVIDENCE → ROUTE → HELP → PRACTISE → TRANSFER → VALIDATE → PERFORM → FADE → RETURN.

This is not a compulsory path through eleven pages. It is a decision sequence. Some learners may skip several stages. A learner who is stable and ready for stretch does not need diagnosis. A learner with a clear examination-execution problem may move directly toward examination craft. A learner with a single known arithmetic break may need only a small HELP intervention and a changed-question retest.

The starting sequence therefore asks one question at each decision point: what does the current evidence justify next?

STEP 01 — ENTER

The support process begins with the learner’s actual request, not with BTT’s preferred article.

A learner or parent can enter through:

a current Mathematics question;
a marked school paper;
a recent drop in marks;
a repeated error;
a stage transition;
a request for tuition;
a request for revision;
a topic or syllabus route;
or a general uncertainty about what to do next.

The broad public entry route is Start Here at Bukit Timah Tutor. The wider Mathematics domain route is the Singapore Mathematics Hub.

Starting rule: do not force a learner to understand the whole estate before receiving a useful route.

STEP 02 — STATE

The next task is to decide what kind of mathematical state is most plausible.

This is not a personality assessment. It is a working description of the current learning condition.

Common starting states include:

missing prerequisite;
fragile representation;
procedure known only in familiar form;
transfer failure;
retention failure;
task-language interference;
examination-execution instability;
support dependence;
or stable capability ready for stretch.

The public route is Find My Mathematics State | What Should I Fix First?.

Starting rule: a score is evidence, not the state itself.

STEP 03 — EVIDENCE

Before a large intervention, collect enough evidence to make the next decision better.

Useful evidence can be surprisingly small:

one marked question with working;
one recent paper;
one prerequisite probe;
one changed-form question;
one comparison between supported and independent work;
one timed versus untimed comparison;
or one explanation from the learner of where they became uncertain.

The purpose is not to gather every possible datum. It is to gather enough information to discriminate between plausible explanations.

If the learner’s Mathematics has started to deteriorate in a way that is not yet understood, When Mathematics Slips | Find the First Useful Break helps separate visible symptoms from the likely useful repair point.

Starting rule: do not respond to uncertainty by maximising volume. Increase evidence resolution first.

STEP 04 — ROUTE

Once the state is sufficiently clear, route to the page that owns the job.

Current problem	Best route
Broad learner state unclear	Find My Mathematics State
Earliest weak link unclear	How Mathematics Diagnosis Works
Intervention depth unclear	How Mathematics Help Works
Feedback needs to change the next action	How Mathematical Feedback Works
Practice needs design	How Mathematical Practice Works
Familiar success needs transfer testing	How Mathematical Transfer Works
Repair needs validation	BTT Mathematical Lab
Known Mathematics is not becoming marks	Mathematics Examination Craft
Support may be ready to reduce	How Independent Mathematics Works

The public routing map is maintained in the BTT Mathematics System Map.

Starting rule: do not make neighbouring pages compete for the same job.

STEP 05 — DIAGNOSE BEFORE REPAIRING LARGE

If several explanations remain plausible, diagnosis becomes the route before teaching or large-scale practice.

The diagnostic question is:

Where does the Mathematics first become unreliable enough to explain the downstream errors we can actually observe?

This may require distinguishing between a weak object and a weak connection. A learner can know fractions and know algebra but fail when fractions enter algebra. A learner can know equations and know graphs but fail to translate one into the other.

The public route is How Mathematics Diagnosis Works | Finding the Earliest Weak Link.

Starting rule: repair the smallest justified cause, not the largest visible chapter.

STEP 06 — HELP

Once the likely break is sufficiently clear, BTT decides how much intervention is justified.

The public HELP ladder is conceptually simple:

WAIT → ASK → REFRAME → CUE → HINT → MODEL → TEACH → GUIDED PRACTICE → INDEPENDENT ATTEMPT → VERIFY → FADE.

Not every learner begins at WAIT and passes through every stage. The ladder exists to prevent two opposite errors:

too little support, where productive struggle becomes unproductive looping;
too much support, where the learner’s mathematical decisions are replaced by the helper.

The route is How Mathematics Help Works | Minimum Justified Help.

Starting rule: HELP should rise only as far as necessary and should begin to fall again as soon as capability returns.

STEP 07 — FEEDBACK

After an attempt, the support process decides what information should return to the learner.

Good feedback should:

preserve what is already correct;
identify the last trustworthy point;
reduce the right uncertainty;
give the learner a next action;
avoid turning every error into a full worked solution;
and create a repair that can be tested.

The public route is How Mathematical Feedback Works | Turn Errors Into the Next Useful Action.

Starting rule: feedback is not complete because the learner has been told the answer. It is complete when the learner can make a better next move.

STEP 08 — PRACTISE

After useful movement is restored, practice must be chosen according to the capability that needs to change.

Practice can have different jobs:

acquire;
stabilise;
retrieve;
space;
interleave;
transfer;
perform;
and fade support.

Repeating the same surface form can be useful during stabilisation. It becomes less useful once the learner is accurate and the real question becomes whether the relationship survives delay, mixed work or changed representation.

The public route is How Mathematical Practice Works | From Repetition to Reliable Performance.

Starting rule: practise for a reason, not because a worksheet exists.

STEP 09 — TRANSFER

Once familiar performance is stable, practice changes the question.

Transfer asks whether the relationship survives changes in:

numbers;
unknown position;
wording;
representation;
context;
topic mixing;
method selection;
or delay.

The public route is How Mathematical Transfer Works | When Learning Survives a Changed Question.

A transfer failure can reopen diagnosis. It may reveal that the learner memorised a procedure, relied on a representation cue, cannot choose among methods, or needs a prerequisite repaired.

Starting rule: changed questions are not punishment. They are evidence of whether the learner owns the relationship rather than the template.

STEP 10 — VALIDATE

A plausible repair should make a prediction.

If signed-number weakness was the real cause of several algebra errors, then a successful signed-number repair should improve embedded sign handling when algebra is retested. If the predicted improvement does not appear, the diagnosis should be reopened rather than defended.

The validation route is the BTT Mathematical Lab | Observe, Probe, Repair, Validate, Release.

Observe → Probe → Repair → Validate → Release.

Starting rule: the evidence owns the conclusion.

STEP 11 — PERFORM

When the Mathematics is independently available and transfer is sufficiently stable, the learner can test whether it survives examination conditions.

Examination performance introduces additional demands:

time allocation;
question sequencing;
command words;
answer forms;
calculator discipline;
working and communication;
verification;
and the ability to recover after a difficult question.

The route is Mathematics Examination Craft | Converting Knowledge Into Marks.

Starting rule: do not turn a knowledge problem into an examination-craft problem merely because it first appeared in an examination.

STEP 12 — FADE

After the learner becomes more capable, support must test whether support can fall.

Support can include obvious help and hidden help:

tutor prompts;
worked examples;
chapter labels;
parent confirmation;
solution keys;
step-by-step digital guidance;
AI hints;
or external pacing reminders.

The public route is How Independent Mathematics Works | When the Learner Can Carry the Route.

The key test is:

What can the learner now do without the support that used to be necessary?

Starting rule: do not leave support at yesterday’s level when today’s evidence says the learner can carry more.

STEP 13 — RETURN

The final step is a return, not another intervention.

The learner returns to the current curriculum, the next mathematical stage, the wider Mathematics estate, or independent work with more control than before.

This matters because diagnosis, HELP and practice can become self-perpetuating if the system has no exit condition.

The support process should therefore ask:

Is the original break repaired?
Did the predicted downstream improvement occur?
Can the learner retrieve the Mathematics after delay?
Can the learner handle a changed form?
Can support be reduced?
Can the learner resume the current stage without being trapped in remedial work?

The learner may then route back into the Singapore Mathematics Hub, the appropriate stage route, or the Mathematics Journey.

Starting rule: support is successful when the learner can leave the layer that was helping them.

The 90-second start

For a parent, learner or tutor who needs the shortest version, use these five questions:

What happened? Look at the actual work or result.
Where did it first become unreliable? Find the earliest useful break.
What is the smallest action worth trying? Do not rebuild everything automatically.
What should improve if this action is correct? Make the repair testable.
Can the learner now do more without help? Return control when the evidence allows.

If any answer is unclear, choose the route that can resolve that uncertainty.

The parent start

Parents do not need to operate every layer. The parent-facing sequence is deliberately simpler:

Bring evidence → identify the state → ask what the first useful repair is → ask how it will be tested → ask what support can later be reduced.

The public route is the Parents’ Guide to Mathematics | What Should I Do Next?.

The learner start

A learner can use an even more compact version:

What is the question asking?
What do I know?
What representation makes this clearer?
Where exactly did I stop trusting my work?
What can I try before asking for the full solution?
How will I check whether the answer makes sense?

This is already a form of mathematical independence. The learner is beginning to carry parts of the decision process internally.

The tutor start

For a tutor, the startup discipline is:

Observe before explaining. Probe before assuming. Intervene at the smallest justified level. Return the task. Watch what changes.

In BTT’s three-student setting, small-group value comes from this visibility. The tutor can see enough working to distinguish three different states even when all three students are studying the same nominal topic.

Three students can therefore receive different intervention depths without requiring three separate curricula.

Starting support should fail safely

A useful support system must know what to do when its first interpretation is wrong.

If a repair fails to produce the predicted improvement:

do not simply increase the same intervention;
return to the evidence;
ask whether the diagnosis was too broad or too narrow;
look for interacting weak links;
test whether representation, retention, task language or examination conditions are involved;
then reroute.

The system should be allowed to revise itself. A confident wrong diagnosis is worse than a provisional one that remains responsive to evidence.

Starting support should stop when the learner is already stable

Not every entry is a repair problem.

If the evidence shows that the learner can retrieve, transfer, verify and perform the current Mathematics independently, the support process should not manufacture a weakness merely because diagnostic tools exist.

The correct route may be stretch, wider Mathematics or the next stage.

Use the Mathematics Journey for stage progression and the Singapore Mathematics Hub for the wider learning estate.

Starting route map

Starting function	Public route
Enter	Start Here at BTT
Explore Mathematics	Singapore Mathematics Hub
State	Find My Mathematics State
Read slippage	When Mathematics Slips
Diagnose	How Mathematics Diagnosis Works
HELP	How Mathematics Help Works
Feedback	How Mathematical Feedback Works
Practise	How Mathematical Practice Works
Transfer	How Mathematical Transfer Works
Validate	BTT Mathematical Lab
Perform	Mathematics Examination Craft
Fade / Return	How Independent Mathematics Works
Check routing boundaries	BTT Mathematics System Map

The starting principle

BTT’s public Mathematics Learning System does not need to know everything before it begins. It needs to know enough to make the next decision safer and more useful.

That produces a disciplined startup:

Enter through the learner’s real problem. Establish the state. Collect enough evidence. Route to the correct route. Intervene minimally. Test what changed. Reduce support. Return control.

The starting sequence is not successful because it activates more machinery. It is successful when it finds the smallest useful route through the system and gives the Mathematics back to the learner.

System routes: Mathematics Next-Step Decision Guide · System Map · complete Mathematics directory.

Starting Mathematics Support | State, Evidence and Next Action