eduKateSG · Why Science?
Meet an unfamiliar challenge, share the thinking and turn scientific knowledge into a calm team decision under time pressure
Explore the official 2026 National STEM Championship: Game On through its new format, four-person school teams, phased challenge route, scientific inquiry, collaboration, communication and reflective preparation.
Science Centre Singapore says 2026 marks an evolution for the National STEM Championship. The refreshed National STEM Championship: Game On retains rigorous academic standards and collaborative spirit while introducing a more dynamic format.
The official 2026 page states that each school may register one team of four students who are currently in Secondary 3, IP Year 3 or Grade 9. Its timeline runs from March registration through phases in May, July and August, with broadcast activity listed for October to November.
Did you know? A challenge can test far more than recall. Teams may need to identify the real question, choose an approach, estimate, test, revise and explain while the clock keeps moving. The science matters, and so does the team’s system for using it.
This article teaches durable preparation habits without inventing unpublished challenge details. Current organiser instructions control tasks, materials, safety, judging and schedules for each phase.
Chapter index
Chapters 1–9: programme identity, Game On format, eligibility, phases, preparation and role design
- 1. Game On is the 2026 championship identity
- 2. The published team route is specific
- 3. A phased timeline changes preparation
- 4. Breadth needs connected foundations
- 5. The first task is to frame the problem
- 6. Roles should match the task, not ego
- 7. Estimation creates an early reality check
- 8. Evidence should outrank the loudest voice
- 9. Testing must preserve the comparison
Chapters 10–17: problem solving, verification, communication, pressure, reflection and future STEM pathways
- 10. A failed attempt contains directional information
- 11. Communication is part of the solution
- 12. Time pressure needs checkpoints
- 13. Practice should mix known and unfamiliar tasks
- 14. Past episodes are examples, not forecasts
- 15. Wellbeing supports reasoning
- 16. A result is one data point
- 17. A complete NSTEMC receipt supports the next challenge
The official page describes NSTEMC:GO as a dynamic revamp of the former National Science Challenge while preserving rigour and collaboration. Students should prepare for applied reasoning rather than trying to predict secret questions from older episodes.
The energy in game on is the 2026 championship identity comes from reasoning together while the problem is still unfamiliar. The official page describes NSTEMC:GO as a dynamic revamp of the former National Science Challenge while preserving rigour and collaboration. Students should prepare for applied reasoning rather than trying to predict secret questions from older episodes. Keep this boundary: Current official instructions control unpublished tasks and judging.
Try this
Run a timed challenge for game on is the 2026 championship identity. Write the goal, constraints and evidence needed for game on is the 2026 championship identity. Save the team’s first assumption and final correction. Keep this limit: Current official instructions control unpublished tasks and judging.
A strong science note separates observation from interpretation. Write the goal, constraints and evidence needed for game on is the 2026 championship identity. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Write the goal, constraints and evidence needed for game on is the 2026 championship identity. Keep this boundary visible: Current official instructions control unpublished tasks and judging.
For 2026, one four-student team per school is listed, with students in Secondary 3, IP Year 3 or Grade 9. Families should not generalise those conditions to future years; school coordinators and the live organiser page control participation.
For the published team route is specific, speed improves when the team makes assumptions visible. For 2026, one four-student team per school is listed, with students in Secondary 3, IP Year 3 or Grade 9. Families should not generalise those conditions to future years; school coordinators and the live organiser page control participation. Preserve this guardrail: A fast answer is useful only when its assumptions and units survive checking.
Try this
Create a role card for the published team route is specific. Assign a solver, checker, recorder and explainer for the published team route is specific. Rotate roles in the next practice. Preserve this boundary: A fast answer is useful only when its assumptions and units survive checking.
A strong science note separates observation from interpretation. Assign a solver, checker, recorder and explainer for the published team route is specific. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Assign a solver, checker, recorder and explainer for the published team route is specific. Keep this boundary visible: A fast answer is useful only when its assumptions and units survive checking.
The 2026 page lists registration, briefing and four competition phases before broadcast. Each transition can require a new kind of readiness. Teams benefit from a reusable method for reading constraints, assigning roles, checking claims and saving decisions.
A strong solution to a phased timeline changes preparation is easy for a teammate to inspect. The 2026 page lists registration, briefing and four competition phases before broadcast. Each transition can require a new kind of readiness. Teams benefit from a reusable method for reading constraints, assigning roles, checking claims and saving decisions. Remember: Past episodes do not predict the next challenge.
Try this
Explain a phased timeline changes preparation in ninety seconds. Estimate a phased timeline changes preparation before using detailed calculation or construction. Include units, uncertainty and one check. Respect this limit: Past episodes do not predict the next challenge.
A strong science note separates observation from interpretation. Estimate a phased timeline changes preparation before using detailed calculation or construction. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Estimate a phased timeline changes preparation before using detailed calculation or construction. Keep this boundary visible: Past episodes do not predict the next challenge.
Championship questions may cross science, technology, engineering and mathematics. Breadth is useful when a learner can connect a physical mechanism, a quantitative relationship, a design constraint and a clear explanation instead of memorising disconnected trivia.
The energy in breadth needs connected foundations comes from reasoning together while the problem is still unfamiliar. Championship questions may cross science, technology, engineering and mathematics. Breadth is useful when a learner can connect a physical mechanism, a quantitative relationship, a design constraint and a clear explanation instead of memorising disconnected trivia. Keep this boundary: Competition placement is not a complete measure of scientific ability.
Try this
Run a timed challenge for breadth needs connected foundations. Review breadth needs connected foundations and identify the clue that should transfer to a new problem. Save the team’s first assumption and final correction. Keep this limit: Competition placement is not a complete measure of scientific ability.
A strong science note separates observation from interpretation. Review breadth needs connected foundations and identify the clue that should transfer to a new problem. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Review breadth needs connected foundations and identify the clue that should transfer to a new problem. Keep this boundary visible: Competition placement is not a complete measure of scientific ability.
Before calculating or building, restate the goal, inputs, constraints and required output. Teams lose time when members solve different versions of the question. A thirty-second shared problem statement can prevent several minutes of elegant work on the wrong target.
For the first task is to frame the problem, speed improves when the team makes assumptions visible. Before calculating or building, restate the goal, inputs, constraints and required output. Teams lose time when members solve different versions of the question. A thirty-second shared problem statement can prevent several minutes of elegant work on the wrong target. Preserve this guardrail: Current official instructions control unpublished tasks and judging.
Try this
Create a role card for the first task is to frame the problem. Write the goal, constraints and evidence needed for the first task is to frame the problem. Rotate roles in the next practice. Preserve this boundary: Current official instructions control unpublished tasks and judging.
A strong science note separates observation from interpretation. Write the goal, constraints and evidence needed for the first task is to frame the problem. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Write the goal, constraints and evidence needed for the first task is to frame the problem. Keep this boundary visible: Current official instructions control unpublished tasks and judging.
One member may track constraints, another calculate, another test and another prepare the explanation. Roles can rotate as the problem changes. Every member should understand the final reasoning so the team never depends on one silent expert.
A strong solution to roles should match the task, not ego is easy for a teammate to inspect. One member may track constraints, another calculate, another test and another prepare the explanation. Roles can rotate as the problem changes. Every member should understand the final reasoning so the team never depends on one silent expert. Remember: A fast answer is useful only when its assumptions and units survive checking.
Try this
Explain roles should match the task, not ego in ninety seconds. Assign a solver, checker, recorder and explainer for roles should match the task, not ego. Include units, uncertainty and one check. Respect this limit: A fast answer is useful only when its assumptions and units survive checking.
A strong science note separates observation from interpretation. Assign a solver, checker, recorder and explainer for roles should match the task, not ego. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Assign a solver, checker, recorder and explainer for roles should match the task, not ego. Keep this boundary visible: A fast answer is useful only when its assumptions and units survive checking.
Order-of-magnitude reasoning can expose impossible answers before detailed work. Write units, plausible bounds and assumptions. An estimate is not a guess when another person can see how it was built and which observation would improve it.
The energy in estimation creates an early reality check comes from reasoning together while the problem is still unfamiliar. Order-of-magnitude reasoning can expose impossible answers before detailed work. Write units, plausible bounds and assumptions. An estimate is not a guess when another person can see how it was built and which observation would improve it. Keep this boundary: Past episodes do not predict the next challenge.
Try this
Run a timed challenge for estimation creates an early reality check. Estimate estimation creates an early reality check before using detailed calculation or construction. Save the team’s first assumption and final correction. Keep this limit: Past episodes do not predict the next challenge.
A strong science note separates observation from interpretation. Estimate estimation creates an early reality check before using detailed calculation or construction. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Estimate estimation creates an early reality check before using detailed calculation or construction. Keep this boundary visible: Past episodes do not predict the next challenge.
When teammates disagree, compare assumptions, run a quick test or inspect the relevant principle. A team culture that welcomes correction is faster than one that protects status. Useful disagreement names the claim and the evidence needed to decide.
For evidence should outrank the loudest voice, speed improves when the team makes assumptions visible. When teammates disagree, compare assumptions, run a quick test or inspect the relevant principle. A team culture that welcomes correction is faster than one that protects status. Useful disagreement names the claim and the evidence needed to decide. Preserve this guardrail: Competition placement is not a complete measure of scientific ability.
Try this
Create a role card for evidence should outrank the loudest voice. Review evidence should outrank the loudest voice and identify the clue that should transfer to a new problem. Rotate roles in the next practice. Preserve this boundary: Competition placement is not a complete measure of scientific ability.
A strong science note separates observation from interpretation. Review evidence should outrank the loudest voice and identify the clue that should transfer to a new problem. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Review evidence should outrank the loudest voice and identify the clue that should transfer to a new problem. Keep this boundary visible: Competition placement is not a complete measure of scientific ability.
If a build or experiment is permitted, change one condition at a time where possible and record the result. A quick trial still needs a baseline. Otherwise the team may mistake motion, noise or visual drama for successful performance.
A strong solution to testing must preserve the comparison is easy for a teammate to inspect. If a build or experiment is permitted, change one condition at a time where possible and record the result. A quick trial still needs a baseline. Otherwise the team may mistake motion, noise or visual drama for successful performance. Remember: Current official instructions control unpublished tasks and judging.
Try this
Explain testing must preserve the comparison in ninety seconds. Write the goal, constraints and evidence needed for testing must preserve the comparison. Include units, uncertainty and one check. Respect this limit: Current official instructions control unpublished tasks and judging.
A strong science note separates observation from interpretation. Write the goal, constraints and evidence needed for testing must preserve the comparison. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Write the goal, constraints and evidence needed for testing must preserve the comparison. Keep this boundary visible: Current official instructions control unpublished tasks and judging.
CHAPTER 10 OF 17 · Solve with evidence
10. A failed attempt contains directional information
Failure can show that a material bends, a model omits a force or an assumption is wrong. State what failed and why the next change follows. Random redesign wastes time; evidence-led revision converts failure into progress.
The energy in a failed attempt contains directional information comes from reasoning together while the problem is still unfamiliar. Failure can show that a material bends, a model omits a force or an assumption is wrong. State what failed and why the next change follows. Random redesign wastes time; evidence-led revision converts failure into progress. Keep this boundary: A fast answer is useful only when its assumptions and units survive checking.
Try this
Run a timed challenge for a failed attempt contains directional information. Assign a solver, checker, recorder and explainer for a failed attempt contains directional information. Save the team’s first assumption and final correction. Keep this limit: A fast answer is useful only when its assumptions and units survive checking.
A strong science note separates observation from interpretation. Assign a solver, checker, recorder and explainer for a failed attempt contains directional information. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Assign a solver, checker, recorder and explainer for a failed attempt contains directional information. Keep this boundary visible: A fast answer is useful only when its assumptions and units survive checking.
A complete answer identifies the question, method, result, uncertainty and recommendation. Clear diagrams, units and short causal sentences let judges follow the reasoning. Presentation polish should never conceal an unsupported leap.
For communication is part of the solution, speed improves when the team makes assumptions visible. A complete answer identifies the question, method, result, uncertainty and recommendation. Clear diagrams, units and short causal sentences let judges follow the reasoning. Presentation polish should never conceal an unsupported leap. Preserve this guardrail: Past episodes do not predict the next challenge.
Try this
Create a role card for communication is part of the solution. Estimate communication is part of the solution before using detailed calculation or construction. Rotate roles in the next practice. Preserve this boundary: Past episodes do not predict the next challenge.
A strong science note separates observation from interpretation. Estimate communication is part of the solution before using detailed calculation or construction. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Estimate communication is part of the solution before using detailed calculation or construction. Keep this boundary visible: Past episodes do not predict the next challenge.
Set moments to confirm the target, inspect intermediate values and reserve time for the final explanation. Teams should decide in advance when to abandon a dead end. A checkpoint prevents sunk effort from consuming the entire round.
A strong solution to time pressure needs checkpoints is easy for a teammate to inspect. Set moments to confirm the target, inspect intermediate values and reserve time for the final explanation. Teams should decide in advance when to abandon a dead end. A checkpoint prevents sunk effort from consuming the entire round. Remember: Competition placement is not a complete measure of scientific ability.
Try this
Explain time pressure needs checkpoints in ninety seconds. Review time pressure needs checkpoints and identify the clue that should transfer to a new problem. Include units, uncertainty and one check. Respect this limit: Competition placement is not a complete measure of scientific ability.
A strong science note separates observation from interpretation. Review time pressure needs checkpoints and identify the clue that should transfer to a new problem. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Review time pressure needs checkpoints and identify the clue that should transfer to a new problem. Keep this boundary visible: Competition placement is not a complete measure of scientific ability.
CHAPTER 13 OF 17 · Learn after the clock
13. Practice should mix known and unfamiliar tasks
Topic revision builds foundations, while mixed challenges train transfer. Use household phenomena, public data and design prompts. After each practice, discuss which clue revealed the governing idea and which knowledge gap actually blocked progress.
The energy in practice should mix known and unfamiliar tasks comes from reasoning together while the problem is still unfamiliar. Topic revision builds foundations, while mixed challenges train transfer. Use household phenomena, public data and design prompts. After each practice, discuss which clue revealed the governing idea and which knowledge gap actually blocked progress. Keep this boundary: Current official instructions control unpublished tasks and judging.
Try this
Run a timed challenge for practice should mix known and unfamiliar tasks. Write the goal, constraints and evidence needed for practice should mix known and unfamiliar tasks. Save the team’s first assumption and final correction. Keep this limit: Current official instructions control unpublished tasks and judging.
A strong science note separates observation from interpretation. Write the goal, constraints and evidence needed for practice should mix known and unfamiliar tasks. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Write the goal, constraints and evidence needed for practice should mix known and unfamiliar tasks. Keep this boundary visible: Current official instructions control unpublished tasks and judging.
CHAPTER 14 OF 17 · Learn after the clock
14. Past episodes are examples, not forecasts
The official page links previous championship episodes. They can reveal pacing, teamwork and communication, but they do not promise the same 2026 tasks. Copying surface features is less valuable than analysing the problem-solving moves underneath.
For past episodes are examples, not forecasts, speed improves when the team makes assumptions visible. The official page links previous championship episodes. They can reveal pacing, teamwork and communication, but they do not promise the same 2026 tasks. Copying surface features is less valuable than analysing the problem-solving moves underneath. Preserve this guardrail: A fast answer is useful only when its assumptions and units survive checking.
Try this
Create a role card for past episodes are examples, not forecasts. Assign a solver, checker, recorder and explainer for past episodes are examples, not forecasts. Rotate roles in the next practice. Preserve this boundary: A fast answer is useful only when its assumptions and units survive checking.
A strong science note separates observation from interpretation. Assign a solver, checker, recorder and explainer for past episodes are examples, not forecasts. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Assign a solver, checker, recorder and explainer for past episodes are examples, not forecasts. Keep this boundary visible: A fast answer is useful only when its assumptions and units survive checking.
Sleep, hydration, calm transitions and respectful communication affect working memory and judgement. Last-minute cramming can make a team less flexible. Preparation should include routines that keep attention available when the challenge becomes surprising.
A strong solution to wellbeing supports reasoning is easy for a teammate to inspect. Sleep, hydration, calm transitions and respectful communication affect working memory and judgement. Last-minute cramming can make a team less flexible. Preparation should include routines that keep attention available when the challenge becomes surprising. Remember: Past episodes do not predict the next challenge.
Try this
Explain wellbeing supports reasoning in ninety seconds. Estimate wellbeing supports reasoning before using detailed calculation or construction. Include units, uncertainty and one check. Respect this limit: Past episodes do not predict the next challenge.
A strong science note separates observation from interpretation. Estimate wellbeing supports reasoning before using detailed calculation or construction. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Estimate wellbeing supports reasoning before using detailed calculation or construction. Keep this boundary visible: Past episodes do not predict the next challenge.
Placement reflects performance in one field under one edition’s rules. It does not measure every student’s scientific potential. The review should preserve strengths, misunderstandings and next questions regardless of the scoreboard.
The energy in a result is one data point comes from reasoning together while the problem is still unfamiliar. Placement reflects performance in one field under one edition’s rules. It does not measure every student’s scientific potential. The review should preserve strengths, misunderstandings and next questions regardless of the scoreboard. Keep this boundary: Competition placement is not a complete measure of scientific ability.
Try this
Run a timed challenge for a result is one data point. Review a result is one data point and identify the clue that should transfer to a new problem. Save the team’s first assumption and final correction. Keep this limit: Competition placement is not a complete measure of scientific ability.
A strong science note separates observation from interpretation. Review a result is one data point and identify the clue that should transfer to a new problem. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Review a result is one data point and identify the clue that should transfer to a new problem. Keep this boundary visible: Competition placement is not a complete measure of scientific ability.
CHAPTER 17 OF 17 · Learn after the clock
17. A complete NSTEMC receipt supports the next challenge
Record edition, official brief, eligibility, team roles, preparation tasks, phase instructions, assumptions, calculations, tests, submitted answers, judge feedback, outcome and personal reflection. Separate confirmed organiser facts from team interpretation.
For a complete nstemc receipt supports the next challenge, speed improves when the team makes assumptions visible. Record edition, official brief, eligibility, team roles, preparation tasks, phase instructions, assumptions, calculations, tests, submitted answers, judge feedback, outcome and personal reflection. Separate confirmed organiser facts from team interpretation. Preserve this guardrail: Current official instructions control unpublished tasks and judging.
Try this
Create a role card for a complete nstemc receipt supports the next challenge. Write the goal, constraints and evidence needed for a complete nstemc receipt supports the next challenge. Rotate roles in the next practice. Preserve this boundary: Current official instructions control unpublished tasks and judging.
A strong science note separates observation from interpretation. Write the goal, constraints and evidence needed for a complete nstemc receipt supports the next challenge. Record the source, version or date, population or device context, method, result and one uncertainty. If the official record does not show a field, write “not shown” instead of filling the gap from memory.
Evidence checkpoint
Write the goal, constraints and evidence needed for a complete nstemc receipt supports the next challenge. Keep this boundary visible: Current official instructions control unpublished tasks and judging.
Championship problem-solving map
| Question | Evidence to capture | Boundary to retain |
|---|---|---|
| Frame | Goal, inputs, constraints and output | Agree on the same problem |
| Build | Principle, model, test and units | Make assumptions visible |
| Check | Bounds, repeat, counterexample and peer review | Speed does not remove verification |
| Explain | Result, evidence, uncertainty and recommendation | Keep claims within the work |
Science learning becomes useful when a student can show where a claim came from, why it applies and where it stops. Continue with the eduKate Science Learning Hub, compare choices in the Education Hub, and connect these habits to career pathways built from subject capabilities.
NSTEMC:GO makes science feel wonderfully alive: four minds meet one unfamiliar problem and discover that calm teamwork can make knowledge move. The lasting win is a group that can listen, test, correct itself and explain why the answer deserves trust.
Primary sources and verification date
Source status checked 27 October 2026. Programme rules, dates, eligibility and participation routes can change; use the linked official owner for the current edition.
- Science Centre Singapore: National STEM Championship — Official 2026 Game On identity, timeline, participation route and past episodes.
- Science Centre Singapore: Competitions — Official school competition directory.
- Science Centre Singapore: About — Official public-science mission and institutional context.