Chapter map

1. 1. System
2. 2. Initial state
3. 3. Step size
4. 4. Momentum update
5. 5. Position update
6. 6. Energy
7. 7. Explicit Euler
8. 8. Jacobian
9. 9. Repeat
10. 10. Phase portrait
11. 11. Leapfrog
12. 12. Variable h
13. 13. Model check
14. 14. Conclusion

Visit the How Mathematics Works hub for the wider series and its connections to education, reasoning, science, computing, design and careers.

---

1. System

This is where modelling choices begin to affect evidence. Use q prime=p,p prime=-q.

A concrete situation

H=(q^2+p^2)/2.

What the mathematics does not license

Canonical coordinates. That distinction matters because a correct formula can still answer the wrong operational question.

Try it, check it, transfer it

Record equations. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

2. Initial state

The central mathematical idea is concrete. Set q0=1,p0=0.

A concrete situation

Initial H=.5.

What the mathematics does not license

Keep exact values. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Record state. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

3. Step size

Here the geometry becomes operational. Set h=.1.

A concrete situation

One small step.

What the mathematics does not license

Long-time claims need many steps. The same reasoning helps students move from an exercise to a real dataset.

Try it, check it, transfer it

Record h. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

4. Momentum update

This step deserves a careful pause. Compute p1=p0-hq0.

A concrete situation

p1=-.1.

What the mathematics does not license

Momentum-first ordering. That distinction matters because a correct formula can still answer the wrong operational question.

Try it, check it, transfer it

Record p1. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

5. Position update

The useful question is not only how to calculate, but what the calculation assumes. Compute q1=q0+hp1.

A concrete situation

q1=.99.

What the mathematics does not license

Use updated momentum. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Record q1. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

6. Energy

A small example makes the mechanism visible. Evaluate H1.

A concrete situation

H1=.49505.

What the mathematics does not license

Not exactly conserved. The same reasoning helps students move from an exercise to a real dataset.

Try it, check it, transfer it

Record value. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

7. Explicit Euler

This is where modelling choices begin to affect evidence. Use q1=1,p1=-.1.

A concrete situation

H1=.505.

What the mathematics does not license

Compare same h. That distinction matters because a correct formula can still answer the wrong operational question.

Try it, check it, transfer it

Record baseline. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

8. Jacobian

The central mathematical idea is concrete. Form symplectic Euler map.

A concrete situation

Determinant=1.

What the mathematics does not license

Area preservation is necessary here. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Record check. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

9. Repeat

Here the geometry becomes operational. Advance many periods.

A concrete situation

Energy error should oscillate boundedly for suitable h.

What the mathematics does not license

Test rather than assert. The same reasoning helps students move from an exercise to a real dataset.

Try it, check it, transfer it

Plot H. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

10. Phase portrait

This step deserves a careful pause. Plot q versus p.

A concrete situation

Compare circle distortion.

What the mathematics does not license

Use equal axes. That distinction matters because a correct formula can still answer the wrong operational question.

Try it, check it, transfer it

Record plot. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

11. Leapfrog

The useful question is not only how to calculate, but what the calculation assumes. Compose half and full substeps.

A concrete situation

Second-order symmetric option.

What the mathematics does not license

Specify ordering. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Record scheme. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

12. Variable h

A small example makes the mechanism visible. Change step adaptively.

A concrete situation

Naive variation can break structure.

What the mathematics does not license

Use justified method. The same reasoning helps students move from an exercise to a real dataset.

Try it, check it, transfer it

Record warning. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

13. Model check

This is where modelling choices begin to affect evidence. Perturb potential.

A concrete situation

Dynamics change physically.

What the mathematics does not license

Integrator cannot validate force law. That distinction matters because a correct formula can still answer the wrong operational question.

Try it, check it, transfer it

Record limitation. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

14. Conclusion

The central mathematical idea is concrete. State Hamiltonian,map,step,invariants and comparison.

A concrete situation

One-step H=.49505.

What the mathematics does not license

Structure differs from exact energy. The habit of checking assumptions is one of the practical benefits of learning mathematics.

Try it, check it, transfer it

Write conclusion. Then explain the result to someone who has not seen the formula. If the explanation needs a hidden assumption, add that assumption to the model record.

---

A seven-day practice lab

1. Day 1: define the decision, observation unit, X variables and Y or reference condition.
2. Day 2: reproduce the smallest worked example and check every dimension, unit and sign.
3. Day 3: build a transparent baseline and write what it assumes.
4. Day 4: fit the main method inside a training-only pipeline; record every parameter.
5. Day 5: challenge one assumption with a deliberately difficult dataset or operating regime.
6. Day 6: compare validation metrics with the practical workload created by errors.
7. Day 7: present one supported conclusion, one unresolved uncertainty and one safe next experiment.

Questions for a student, parent or mentor

• Which part of the model is a mathematical definition, and which part is a choice?
• What small calculation would catch a sign, scale or dimension error?
• What evidence comes from training, validation and the untouched test?
• What benign condition could look like a fault or prediction failure?
• How would the answer change for another response, asset, school project or time period?
• What should a human review before acting on the score?

A method-specific audit before deployment

Symplectic integrators are time-stepping methods designed for Hamiltonian systems. They preserve the geometric two-form of phase space, which helps reproduce long-term qualitative behaviour of conservative dynamics.

For the harmonic oscillator q prime=p and p prime=-q, the Hamiltonian energy is H=(q squared+p squared)/2. The exact trajectory is a circle in the q-p plane.

Start at q0=1, p0=0 with h=0.1. One symplectic Euler ordering updates momentum first: p1=p0-hq0=-0.1, then q1=q0+hp1=0.99.

The new energy is [0.99 squared+0.1 squared]/2=0.49505, compared with the exact 0.5. One step does not preserve this Hamiltonian exactly, yet the method preserves the symplectic structure.

Ordinary explicit Euler gives q1=1 and p1=-0.1, with energy 0.505. Repeated explicit Euler steps spiral outward for this oscillator, while symplectic Euler usually keeps energy error bounded and oscillatory over long runs.

The word symplectic refers to preservation of oriented phase-space area in the canonical coordinates. It is a geometric condition on the update map, not a synonym for small local error.

Did You Know? A method can have a visibly larger one-step error yet reproduce long-term orbital geometry better than a higher-looking method that systematically drifts in energy.

Ruth’s 1983 canonical integration technique built explicit higher-order compositions for Hamiltonian motion. Modern geometric integration develops this idea through splitting, generating functions and backward error analysis.

Hamiltonians that split into kinetic and potential parts allow alternating exact subflows. The leapfrog or Störmer-Verlet method is a widely used symmetric, second-order example.

Symplecticity does not automatically preserve every invariant. Momentum may be preserved when symmetry and discretisation align; energy is commonly nearly conserved through a modified Hamiltonian rather than exactly fixed.

Step size still matters. Resonances, close encounters or rapidly varying forces can defeat an otherwise appropriate integrator, and careless variable step sizes can destroy symplectic structure.

Students can plot q against p for explicit Euler, symplectic Euler and an accurate reference. Geometry makes accumulated bias easier to see than a short table of local errors.

Applications include celestial mechanics, accelerator physics, molecular dynamics and plasma models. Long-time numerical fidelity does not validate a force law, potential or omitted interaction.

Dissipative systems need different reasoning because true phase-space volume contracts. Applying a symplectic method merely because it sounds stable can preserve the wrong qualitative feature.

For parents and students, the transferable lesson is that repeated small decisions should respect the structure of the whole system, especially when the horizon is much longer than one step.

A responsible report states the Hamiltonian, canonical variables, splitting, update order, step size, invariant diagnostics, phase portraits, convergence tests, comparison method and model limitations.

---

Frequently asked questions

What is a symplectic integrator?

A time-stepping map that preserves the symplectic geometric structure of a Hamiltonian system.

Does it conserve energy exactly?

Usually not; it often keeps energy error bounded by following a nearby modified Hamiltonian.

Why is long-term behaviour important?

Small systematic bias can overwhelm orbital or molecular simulations after many steps.

Is symplectic Euler the same as explicit Euler?

No. It uses an updated variable in the second substep and has different geometry.

Can the step size vary freely?

Naive adaptation can break symplecticity and requires special design.

What must be reported?

Hamiltonian, coordinates, splitting, update order, step size, invariants, phase plots, convergence and model limits.

---

Primary source and fact-check note

Ronald D. Ruth, “A Canonical Integration Technique,” IEEE Transactions on Nuclear Science 30, 1983. Definitions and core claims in this article were checked against this primary publication. Calculations shown here were independently recomputed. Examples are educational illustrations; they are not reported industrial performance results.

Useful next reading

• Spectral methods
• Finite difference method
• Computational mathematics hub
• How Mathematics Works
• Study and learning methods

Keep the options open. Mathematics education supports clearer questions, transferable problem-solving skills and informed choices across science, engineering, computing, AI, statistics, design and many other careers. It does not guarantee intelligence, income, admission or employment; those outcomes depend on far more than one subject or model.

Method audit: equations, structure and evidence

Geometry audit

Check the update map rather than relying on an energy table alone.

Invariant audit

Name which quantities are exact, near-conserved or unconstrained.

Horizon audit

Test across the actual number of periods required.

Model audit

Separate numerical drift from an incorrect Hamiltonian.

Mechanism workshop: calculate, challenge and transfer

Map build

Write the momentum-first update as a matrix.

Determinant check

Verify phase-area preservation for the oscillator.

Euler comparison

Run both maps for one hundred periods.

Energy plot

Graph H_n-H_0 rather than quoting one endpoint.

Phase portrait

Use equal axes and compare orbital drift.

Leapfrog build

Compose half momentum, full position and half momentum steps.

Step study

Repeat with h, h/2 and h/4.

Perturbed potential

Test whether the qualitative advantage persists.

Student explanation

Explain structure preservation with a phase-space drawing.

Parent explanation

Connect repeated small bias with long-horizon decisions.

Evidence notebook: reconstruct every step

Evidence notebook 1: System

Step 1 isolates system. Use q prime=p,p prime=-q. The checked result is: H=(q^2+p^2)/2. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Canonical coordinates. Record equations. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 2: Initial state

Step 2 isolates initial state. Set q0=1,p0=0. The checked result is: Initial H=.5. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Keep exact values. Record state. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 3: Step size

Step 3 isolates step size. Set h=.1. The checked result is: One small step. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Long-time claims need many steps. Record h. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 4: Momentum update

Step 4 isolates momentum update. Compute p1=p0-hq0. The checked result is: p1=-.1. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Momentum-first ordering. Record p1. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 5: Position update

Step 5 isolates position update. Compute q1=q0+hp1. The checked result is: q1=.99. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Use updated momentum. Record q1. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 6: Energy

Step 6 isolates energy. Evaluate H1. The checked result is: H1=.49505. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Not exactly conserved. Record value. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 7: Explicit Euler

Step 7 isolates explicit euler. Use q1=1,p1=-.1. The checked result is: H1=.505. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Compare same h. Record baseline. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 8: Jacobian

Step 8 isolates jacobian. Form symplectic Euler map. The checked result is: Determinant=1. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Area preservation is necessary here. Record check. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 9: Repeat

Step 9 isolates repeat. Advance many periods. The checked result is: Energy error should oscillate boundedly for suitable h. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Test rather than assert. Plot H. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 10: Phase portrait

Step 10 isolates phase portrait. Plot q versus p. The checked result is: Compare circle distortion. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Use equal axes. Record plot. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 11: Leapfrog

Step 11 isolates leapfrog. Compose half and full substeps. The checked result is: Second-order symmetric option. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Specify ordering. Record scheme. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 12: Variable h

Step 12 isolates variable h. Change step adaptively. The checked result is: Naive variation can break structure. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Use justified method. Record warning. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 13: Model check

Step 13 isolates model check. Perturb potential. The checked result is: Dynamics change physically. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Integrator cannot validate force law. Record limitation. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Evidence notebook 14: Conclusion

Step 14 isolates conclusion. State Hamiltonian,map,step,invariants and comparison. The checked result is: One-step H=.49505. Keep the equation, units, node or element order, face orientation, basis convention, time level and precision beside every value. The safeguard is: Structure differs from exact energy. Write conclusion. Recalculate after changing exactly one parameter and explain whether the change arises from spatial approximation, time integration, geometry preservation, refinement logic, numerical flux, rounding or a different physical model. Ask a second reader to reproduce the arithmetic and finish with the narrowest claim the evidence supports.

Failure cases and responsible recovery

Energy-equals-symplectic

These are different properties. Build a small counterexample, restore the missing condition and recalculate before making the claim.

One-step ranking

Long-time structure is the main question. Build a small counterexample, restore the missing condition and recalculate before making the claim.

Variable-step shortcut

Naive adaptation can break the geometry. Build a small counterexample, restore the missing condition and recalculate before making the claim.

Dissipative misuse

Volume preservation may model the wrong physics. Build a small counterexample, restore the missing condition and recalculate before making the claim.

Invariant overclaim

Not every conserved quantity is automatic. Build a small counterexample, restore the missing condition and recalculate before making the claim.

Phase plot omitted

Geometry can reveal drift hidden by summaries. Build a small counterexample, restore the missing condition and recalculate before making the claim.

Large step

Resonance and instability remain possible. Build a small counterexample, restore the missing condition and recalculate before making the claim.

Force-law trust

An integrator cannot validate the model. Build a small counterexample, restore the missing condition and recalculate before making the claim.

Verified case study: fourteen steps from equation to decision

1. Define the continuous model

Write the equation, variables, units, domain and initial or boundary data.

2. Identify preserved structure

Name conservation, dissipation, symplectic geometry, smoothness or localisation that should guide the discretisation.

3. Choose representation

Justify nodes, elements, local polynomials, nested grids or phase variables.

4. Derive one local rule

Connect a matrix row, canonical substep, refinement decision or face flux to the governing mathematics.

5. Calculate by hand

Use the smallest nontrivial example and keep exact values where practical.

6. Check signs and units

Audit orientation, dimensions, time levels and indexing.

7. Test an invariant

Measure mass, phase area, energy behaviour, boundary values or another eligible diagnostic.

8. Trigger a known failure

Use a bad time step, escaped feature, wrong trace or broken interface transfer.

9. Refine systematically

Vary one of space, time, polynomial degree or threshold while controlling the others.

10. Compare a baseline

Use an exact, manufactured, uniform-grid or independently implemented reference.

11. Count cost

Record degrees of freedom, cell-updates, matrix solves, memory and elapsed time.

12. Validate the model

Compare the application output with independent observations or accepted benchmarks.

13. Communicate limits

State what the result supports and what remains untested.

14. Preserve evidence

Save equations, data, code, meshes, parameters, logs and corrected calculations.

Learner lab: ten checks for transferable understanding

Lab question 1

Create a reproducible mini-investigation of canonical map, phase area, modified energy, resonance and long-horizon diagnostics. For task 1, state one prediction, calculate or simulate a transparent case, record units and conventions, compare an independent baseline, identify one failure mode, and explain the result once for a student and once for a parent. End with one practical next step and one claim the evidence does not support.

Lab question 2

Create a reproducible mini-investigation of canonical map, phase area, modified energy, resonance and long-horizon diagnostics. For task 2, state one prediction, calculate or simulate a transparent case, record units and conventions, compare an independent baseline, identify one failure mode, and explain the result once for a student and once for a parent. End with one practical next step and one claim the evidence does not support.

Lab question 3

Create a reproducible mini-investigation of canonical map, phase area, modified energy, resonance and long-horizon diagnostics. For task 3, state one prediction, calculate or simulate a transparent case, record units and conventions, compare an independent baseline, identify one failure mode, and explain the result once for a student and once for a parent. End with one practical next step and one claim the evidence does not support.

Lab question 4

Create a reproducible mini-investigation of canonical map, phase area, modified energy, resonance and long-horizon diagnostics. For task 4, state one prediction, calculate or simulate a transparent case, record units and conventions, compare an independent baseline, identify one failure mode, and explain the result once for a student and once for a parent. End with one practical next step and one claim the evidence does not support.

Lab question 5

Create a reproducible mini-investigation of canonical map, phase area, modified energy, resonance and long-horizon diagnostics. For task 5, state one prediction, calculate or simulate a transparent case, record units and conventions, compare an independent baseline, identify one failure mode, and explain the result once for a student and once for a parent. End with one practical next step and one claim the evidence does not support.

Lab question 6

Create a reproducible mini-investigation of canonical map, phase area, modified energy, resonance and long-horizon diagnostics. For task 6, state one prediction, calculate or simulate a transparent case, record units and conventions, compare an independent baseline, identify one failure mode, and explain the result once for a student and once for a parent. End with one practical next step and one claim the evidence does not support.

Lab question 7

Create a reproducible mini-investigation of canonical map, phase area, modified energy, resonance and long-horizon diagnostics. For task 7, state one prediction, calculate or simulate a transparent case, record units and conventions, compare an independent baseline, identify one failure mode, and explain the result once for a student and once for a parent. End with one practical next step and one claim the evidence does not support.

Lab question 8

Create a reproducible mini-investigation of canonical map, phase area, modified energy, resonance and long-horizon diagnostics. For task 8, state one prediction, calculate or simulate a transparent case, record units and conventions, compare an independent baseline, identify one failure mode, and explain the result once for a student and once for a parent. End with one practical next step and one claim the evidence does not support.

Lab question 9

Create a reproducible mini-investigation of canonical map, phase area, modified energy, resonance and long-horizon diagnostics. For task 9, state one prediction, calculate or simulate a transparent case, record units and conventions, compare an independent baseline, identify one failure mode, and explain the result once for a student and once for a parent. End with one practical next step and one claim the evidence does not support.

Lab question 10

Create a reproducible mini-investigation of canonical map, phase area, modified energy, resonance and long-horizon diagnostics. For task 10, state one prediction, calculate or simulate a transparent case, record units and conventions, compare an independent baseline, identify one failure mode, and explain the result once for a student and once for a parent. End with one practical next step and one claim the evidence does not support.