Build the Bridge

The maths behind the bridge

Build the Bridge is a real — if simplified — structural simulation. Every beam carries a force you could compute by hand, and it breaks for a physical reason. Here is the engineering it runs on, so you can play it like an engineer.

1. A pin-jointed truss

The structure is modelled as a truss: rigid joints (the dots) connected by straight members (the beams). Each member is a two-force member — it carries only an axial force along its own length, either tension (being pulled apart) or compression (being squeezed). Loads act at the joints; the banks are fixed supports that push back with reaction forces.

2. Member force — Hooke's law

Each beam behaves like a stiff spring. Stretch it by ΔL and it pulls back with a force proportional to that stretch:

E is the material's stiffness (Young's modulus), A the cross-section area, L the length and L₀ its natural (unloaded) length. A positive ΔL means tension; negative means compression. Because F = (E·A/L)·ΔL = E·A·ε, the force is just the material stiffness times the strain ε = ΔL/L — which is why the beams visibly stretch under load before they break.

3. Balancing every joint

A truss is in equilibrium when the forces at every joint cancel out — the method of joints:

For a statically determinate truss — members m, reactions r and joints j satisfying m + r = 2j — these equations have a single exact solution that gives the force in every member. This game reaches the same equilibrium a different way: it relaxes the structure, nudging each joint along its members until the forces balance. Relaxation has a bonus — it also handles redundant bracing (extra, stronger) and under-braced structures (which simply sag and collapse, exactly as a real mechanism would).

4. When a beam breaks

A member fails when its force passes the capacity of the material — and tension and compression fail in completely different ways.

In tension, it snaps once the pull exceeds the tensile strength — independent of length:

In compression, a slender strut doesn't get crushed — it bows sideways and buckles. Euler's critical load governs it:

I is the second moment of area and K the effective-length factor. The decisive part is 1/L²: halving a strut's length quadruples how much compression it can take. That's why long, unbraced struts are always the first to buckle — and why short triangulating members are so effective.

A beam's utilization is its force over its capacity; it breaks when that ratio passes 1, and one failure dumps its load onto its neighbours — which can trigger a progressive (cascade) collapse:

The beam colours are this number: green is a relaxed member (u < 0.5), amber is working hard (u < 0.85), and red means u is approaching 1 — about to snap.

5. Why triangles win

A triangle is the only polygon that is rigid with pinned corners — you can't change its shape without changing the length of a side. A four-bar square panel is a mechanism: it shears flat with no resistance. Triangulating a span converts bending (which slender members resist poorly) into pure axial tension and compression (which they carry efficiently). Almost all of bridge design is finding a path of triangles from the load down to the supports.

6. Why a flat road can't hold a truck

Picture a weight W hanging at mid-span from two almost-horizontal beams meeting the load at an angle θ. Vertical equilibrium needs 2·F·sin θ = W, so the force in each beam is:

As the road flattens, θ shrinks and the required force explodes — a nearly flat deck has to develop enormous tension to support even a light truck, so it tears apart. Give the load a path through steep diagonals (a deep truss) and θ is large, keeping every member force small. That single equation is the whole reason this game is hard.

7. From physics to the game

The simulation lumps the textbook quantities into a handful of constants. Same equations, simpler units:

So when you watch a beam flush from green to red as the truck rolls on, you're watching its demand-to-capacity ratio climb toward 1 in real time — the same calculation a structural engineer does before anyone drives across.

8. A worked example

Put numbers on it. Suppose a wheel drops a load of W = 20 kN (about a 2-tonne axle) onto a joint carried by two members that meet the load at an angle θ. From F = W / (2·sin θ):

The same 20 kN load asks eight times more force of a flat deck than of a 45° truss. Steeper triangles are not a style choice — they are the difference between a member that survives and one that snaps.

9. Dead load vs. live load

Engineers split the load in two. The dead load is the permanent self-weight of the structure — constant, modelled here as a small downward force at every joint. The live load is the moving truck: it is transient and, crucially, it travels. A member that is relaxed when the truck is on the bank can be the most-stressed member when the truck reaches mid-span. Real design checks the live load at every position (an influence line); in the game, just watch the colours sweep as the truck crosses to spot the worst case.

10. Common truss patterns

Centuries of bridges converged on a few triangulation patterns. Any of them work here — try them:

• Warren — a row of equal triangles with alternating diagonals and no verticals. Simple and efficient; diagonals alternate tension and compression.
• Pratt — verticals plus diagonals sloping down toward the centre. Under gravity the long diagonals are in tension — ideal when ties are cheaper or stronger than struts.
• Howe — the mirror of Pratt: diagonals slope up toward the centre and go into compression. Historically used for timber, which is strong in compression.
• K-truss — verticals split each diagonal into two shorter pieces. Because buckling capacity scales as 1/L², halving a strut's length quadruples its strength — so K-bracing is the move when compression members are failing.

11. Assumptions & limitations

This is a faithful teaching model, not a design tool. To stay simple it assumes:

• Pin joints — frictionless hinges, so members carry only axial force and no bending moment. Real welded or bolted joints carry some moment.
• Linear-elastic members — force is proportional to stretch right up to a single breaking threshold. Real materials yield, work-harden, and fatigue along a full stress–strain curve.
• Two dimensions — in-plane only; no torsion, out-of-plane buckling, wind or lateral load.
• Quasi-static relaxation — it captures sag and progressive collapse but not true dynamics, resonance, or impact.
• One idealised material, lumped constants — no real units (kN, MPa) and, deliberately, no safety factors — a real bridge is designed to carry several times its rated load.
• Simplified buckling — Euler's 1/L² law without slenderness limits or inelastic effects.

In short: it is right about why bridges stand or fall, but you should not build a real one from it.

Glossary

Joint (node)

—

a connection point where members meet; loads and supports act here.

Member (beam)

—

a straight bar between two joints carrying a single axial force.

Axial force

—

force acting along a member's length; the only force a pin-jointed member carries.

Tension / compression

—

being pulled apart (tension) versus being pushed together (compression).

Buckling

—

sudden sideways bowing of a compression member before it is crushed; governed by Euler's load.

Statically determinate

—

a truss whose forces follow from equilibrium alone, when m + r = 2j.

Utilization

—

force divided by capacity (the demand-to-capacity ratio); failure at 1.

Span

—

the clear distance the bridge has to bridge between supports.

Frequently asked questions

Further reading

• Truss — structure and analysis of triangulated frames.
• Euler's critical load — the buckling formula (Leonhard Euler, 1757).
• Hooke's law — force proportional to extension (Robert Hooke, 1678).
• Method of joints — solving truss forces from equilibrium.