Elastic Collisions

Lift one ball of a Newton's cradle and let it fall. Click — the four balls in the middle never visibly move, and the far ball leaps out with the full speed of the first. Not two balls at half speed, not five balls sharing the motion: exactly one ball, exactly the incoming speed. The desk toy is executing, five times over in a few milliseconds, the most constrained interaction in mechanics — the elastic collision, in which the total kinetic energy survives the crash along with the momentum that every collision conserves. That extra conservation law is a powerful straitjacket. In one dimension it leaves no freedom at all: two equations, two unknown final velocities, one outcome. Equal masses are forced to swap velocities — that is the cradle's secret. A ping-pong ball bounces off a bowling ball because the mass ratio forces a reversal; a bowling ball plows through a ping-pong ball and kicks it forward at nearly twice the bowling ball's speed, because the algebra says so. All of it is packaged in one elegant rule: the speed at which two bodies approach equals the speed at which they separate. The same straitjacket runs serious machinery. Nuclear reactors slow their neutrons by letting them collide elastically with light nuclei — the equal-mass exchange rule is why water works so well. Rutherford discovered the atomic nucleus by watching alpha particles scatter elastically off gold. And in 2026, the most sensitive experiments on Earth sit a mile underground waiting for a dark-matter particle to collide elastically with a xenon nucleus — physics still betting on the same two conservation laws Huygens wrote down in 1669.

Key takeaways

An elastic collision conserves both total momentum (like every collision) and total kinetic energy (the special, defining extra) — nothing leaks into heat, sound, or permanent deformation.
In 1D the two laws collapse into one linear rule: relative velocity reverses, $v_1 - v_2 = -(v_1' - v_2')$ — approach speed equals separation speed, whatever the masses.
Closed forms follow: $v_1' = \frac{m_1 - m_2}{m_1 + m_2}v_1 + \frac{2m_2}{m_1 + m_2}v_2$ ; equal masses swap velocities; a light ball bounces off a heavy one; a heavy ball kicks a light one to nearly $2v$ .
Maximum energy transfer happens at equal masses — one head-on elastic collision can hand over 100% of the kinetic energy, the principle behind neutron moderation.
In 2D the conservation laws leave one angle free (the impact geometry); equal masses with one at rest always separate at 90°.
The coefficient of restitution $e =$ separation speed / approach speed grades real collisions: $e = 1$ elastic, $e = 0$ perfectly sticky.
Key idea: momentum conservation always holds in a free collision; elastic means kinetic energy survives too — equivalent in 1D to the relative velocity exactly reversing, which fixes the outcome completely.

The core insight

Momentum conservation always holds in a free collision; elastic means kinetic energy is also conserved — in 1D equivalent to the relative velocity reversing sign, $v_1 - v_2 = -(v_1' - v_2')$ , which together with momentum fixes both final velocities completely.

Where this lives today

The equal-mass exchange rule is load-bearing infrastructure. Every pressurized-water reactor on Earth thermalizes its neutrons by elastic collisions with the hydrogen in water — the Worked Example 2 calculation, run at industrial scale — and moderator choice (light water, heavy water, graphite) is still a trade of moderating power against neutron absorption. The same kinematics runs in reverse in neutron scattering facilities, where measured recoil energies decode the masses and motions of atoms in materials. The frontier application is stranger: listening for dark matter. Experiments like LZ and XENONnT sit kilometers underground watching tonnes of liquid xenon for the flash of a single nucleus recoiling from an elastic collision with a passing dark-matter particle. The expected recoil spectrum is computed with exactly this article's two conservation laws — the energy-transfer fraction $4 m_1 m_2/(m_1+m_2)^2$ dictates that a WIMP kicks a xenon nucleus hardest when their masses match, which is why detector materials are chosen with the unknown mass in mind. And every interplanetary mission still books its gravitational slingshots as elastic collisions with planets: in Jupiter's frame the spacecraft's speed is unchanged (a "bounce" off the gravity well), but transformed back to the Sun's frame it can steal tens of km/s — the heavy-on-light $2v$ kick, with a planet as the clubhead.

The physical picture

Every elastic collision is a spring story told too fast to watch. As the bodies meet, they deform; kinetic energy pours into elastic strain energy — for steel balls, real compression over microseconds — and then the deformation unwinds completely, paying every joule back into motion. "Elastic" is precisely the promise of full repayment. A clay ball breaks the promise (the dent stays; the energy stays in it as heat), a superball nearly keeps it, and two colliding atoms — with no internal friction to feed and their excited states energetically out of reach — keep it exactly. That is worth savoring: perfect elastic collisions are not an idealization of the microworld, they are its routine. What outcome does the promise force? Three regimes, one formula, and you can read all three off the mass ratio before doing any algebra:

Figure

Three mass ratios, three destinies — all from the same two conservation laws

The tidiest viewpoint sits in the center-of-mass frame — ride along with the point $V_{\mathrm{cm}} = (m_1 v_1 + m_2 v_2)/(m_1 + m_2)$ , and the two bodies always approach with equal and opposite momenta. An elastic collision, seen from there, is almost nothing: each body keeps its speed and reverses (in 1D) or rotates (in 2D) its direction. All the apparent complexity of the lab-frame outcomes is just this symmetric bounce viewed from a moving walkway. Huygens discovered the trick three and a half centuries ago, imagining a collision performed on a smoothly gliding boat while an observer watches from the bank — the same experiment, two stories, and the boat's speed can be chosen to make the story simple. The one-line summary of elasticity, and the tool you will actually use: whatever speed the gap was closing at, it opens at the same speed afterward. A superball meets the floor at 8 m/s and leaves at 8 m/s; a tennis ball meets an oncoming racket and the relative speed of ball and strings is preserved — which, as we will compute, is how a moving heavy object can fling a light one forward at up to twice its own speed.

The math: two laws, one straitjacket

Set up the standard 1D encounter: masses $m_1$ and $m_2$ moving along a line with signed velocities $v_1$ and $v_2$ before the collision, $v_1'$ and $v_2'$ after. Every free collision, elastic or not, conserves the vector total of momentum — the internal contact forces cancel in third-law pairs:

m_1 v_1 + m_2 v_2 = m_1 v_1' + m_2 v_2' \qquad (1)

where: all velocities are signed numbers along the collision line — the bookkeeping fails silently if a leftward velocity forgets its minus sign. One equation, two unknowns: momentum alone cannot predict the outcome, which is why a sticky crash and a bouncy one can share the same total $p$ . Elasticity is the additional promise that the kinetic-energy ledger also balances — the strain energy stored in the microsecond of compression is repaid in full:

\tfrac12 m_1 v_1^2 + \tfrac12 m_2 v_2^2 = \tfrac12 m_1 v_1'^2 + \tfrac12 m_2 v_2'^2 \qquad (2)

where: only translational kinetic energy appears — no heat, no sound, no dents. Two equations now, two unknowns: the outcome is fully determined. But (2) is quadratic, and solving quadratics invites pain. The classic trick converts the pair into something linear — and the result is the single most useful fact about elastic collisions.

Derivation AThe relative-velocity reversal rule

We are deriving equation (3), the signature of 1D elasticity: the relative velocity of the two bodies exactly reverses, $v_1 - v_2 = -(v_1' - v_2')$ — approach speed equals separation speed. The strategy is to rearrange both conservation laws so that body 1's quantities sit on the left and body 2's on the right, then divide the energy equation by the momentum equation; the quadratics factor as differences of squares and the masses cancel wholesale. Watch the division step: it silently assumes $v_1 \neq v_1'$ , which is exactly the assumption that a collision actually occurred — the "nothing happened" solution is a genuine algebraic root that must be discarded by hand.

Step 1. Rearrange momentum (1) with each body on its own side:

m_1 (v_1 - v_1') = m_2 (v_2' - v_2)

Why: this reads "momentum lost by 1 equals momentum gained by 2" — conservation as a transfer statement.

Step 2. Rearrange energy (2) the same way (the halves cancel) and factor each side as a difference of squares:

m_1 (v_1 - v_1')(v_1 + v_1') = m_2 (v_2' - v_2)(v_2' + v_2)

Why: the factoring is the whole trick — it exposes the momentum-transfer factors of Step 1 sitting inside the energy equation.

Step 3. Divide Step 2 by Step 1. The mass-weighted transfer factors cancel on both sides, leaving

v_1 + v_1' = v_2 + v_2'

Why the division is legal: it requires $v_1 \neq v_1'$ — a genuine collision. The excluded case $v_1' = v_1,\, v_2' = v_2$ ("the bodies miss each other") satisfies both conservation laws trivially and always lurks as the second root of the quadratic.

Step 4. Rearrange into the memorable form:

v_1 - v_2 = -(v_1' - v_2') \qquad (3)

where: $v_1 - v_2$ — the approach velocity (rate the gap closes); $v_1' - v_2'$ — the separation velocity (rate it opens). Every mass has vanished: the rule is universal for 1D elastic collisions, from billiard balls to neutrons on nuclei. Verified: for a ball on an immovable wall ( $v_2 = v_2' = 0$ ) the rule gives $v_1' = -v_1$ , the perfect bounce; cross-checked against Morin ch. 5 and Taylor ch. 14.

The reversal rule replaces the quadratic (2) with a linear equation, and two linear equations solve by substitution in three lines:

Go deeper · Derivation B · 1 minThe final velocities in closed form

Derivation BThe final velocities in closed form

The goal is the general solution of the 1D elastic collision — the outgoing velocities as explicit functions of the masses and incoming velocities, equations (4) and (5). The strategy is to solve the linear pair consisting of momentum conservation (1) and the reversal rule (3), which is legitimate because Derivation A showed (3) carries the full content of energy conservation for a genuine collision. The payoff to watch for: the answer is a weighted mixing of the incoming velocities, and the three panels of the hero figure fall out of the weights.

Step 1. Solve the reversal rule (3) for $v_2'$ : $v_2' = v_1' + v_1 - v_2$ . Why: eliminating one unknown reduces the system to a single equation in $v_1'$ .

Step 2. Substitute into momentum (1): $m_1 v_1 + m_2 v_2 = m_1 v_1' + m_2 (v_1' + v_1 - v_2)$ , and collect the $v_1'$ terms: $(m_1 + m_2)\,v_1' = (m_1 - m_2)\,v_1 + 2 m_2 v_2$ .

Step 3. Divide, and write the symmetric partner (swap labels $1 \leftrightarrow 2$ ):

v_1' = \frac{m_1 - m_2}{m_1 + m_2}\,v_1 + \frac{2 m_2}{m_1 + m_2}\,v_2 \qquad (4)

v_2' = \frac{2 m_1}{m_1 + m_2}\,v_1 + \frac{m_2 - m_1}{m_1 + m_2}\,v_2 \qquad (5)

where: the coefficients are pure mass ratios. Read (4) physically: the first weight $(m_1 - m_2)/(m_1 + m_2)$ is a reflection coefficient — positive (keep going) when you outweigh the target, negative (bounce back) when it outweighs you, zero at equal masses; the second weight is how much of the target's motion you inherit.

Step 4. Sanity-check the three panels. Equal masses, target at rest: $v_1' = 0$ , $v_2' = v_1$ — complete exchange. Light on heavy ( $m_1 \ll m_2$ , $v_2 = 0$ ): $v_1' \to -v_1$ , a wall bounce. Heavy on light ( $m_1 \gg m_2$ , $v_2 = 0$ ): $v_1' \to v_1$ (plows on), $v_2' \to 2 v_1$ — the light target is flung at up to twice the projectile's speed. Why the 2: in the heavy body's frame the light one approaches at $v_1$ and must depart at $v_1$ (reversal rule); adding the frame velocity back doubles it. *Verified: substituting (4) and (5) back into both (1) and (2) balances each ledger identically; the $m_1 = m_2$ exchange and the $m_2 \to \infty$ bounce match Morin ch. 5.*

One more result belongs in the permanent toolkit, because it turns 2D collisions from a nightmare into a protractor exercise. In two dimensions the conservation laws (three scalar equations) cannot fix four unknown velocity components — the missing input is how off-center the hit was. But for the special case that rules billiards, the geometry is pinned anyway:

Go deeper · Derivation C · 1 minEqual masses separate at ninety degrees

Derivation CEqual masses separate at ninety degrees

We are proving the billiard player's theorem: when two equal masses collide elastically and one starts at rest, the outgoing velocity vectors are always perpendicular — whatever the impact geometry. The strategy is vector algebra rather than components: write momentum conservation as a vector triangle, square it, and subtract the energy condition; the cross term that survives is exactly the dot product of the final velocities. This proof is worth internalizing as a template — "square the momentum equation and compare with energy" resolves many collision problems without ever touching angles.

Step 1. With $m_1 = m_2$ and target at rest, momentum conservation (1) reads, as vectors, $\mathbf{v}_1 = \mathbf{v}_1' + \mathbf{v}_2'$ . Why the masses vanish: they are equal and divide out of every term.

Step 2. Square it (dot each side with itself):

v_1^2 = v_1'^2 + v_2'^2 + 2\,\mathbf{v}_1' \cdot \mathbf{v}_2'

Step 3. Energy conservation (2), with equal masses, says $v_1^2 = v_1'^2 + v_2'^2$ . Subtract:

\mathbf{v}_1' \cdot \mathbf{v}_2' = 0

Why this ends the proof: a vanishing dot product between two nonzero vectors means a $90°$ angle — cue ball and object ball leave along perpendicular lines. The degenerate head-on case gives $\mathbf{v}_1' = 0$ (full exchange), where the angle is undefined because one ball stops. *Verified: visible on any pool table and in cloud-chamber photographs of equal-mass particle scattering — proton on proton tracks meet at right angles, a classic identification signature.*

Figure

The billiard theorem — equal masses, one at rest, always part at a right angle

Which formula answers which question

you want	use	notes
any 1D elastic outcome	eqs. (4)–(5)	signed velocities in, signed velocities out
a quick check or a shortcut	reversal rule, eq. (3)	approach speed = separation speed, always
how elastic a real bounce is	$e = \dfrac{\|v_2' - v_1'\|}{\|v_1 - v_2\|}$	$e = 1$ elastic, $e = 0$ sticky; superball $\approx 0.9$
energy handed to a resting target	fraction $\dfrac{4 m_1 m_2}{(m_1 + m_2)^2}$	maximal ( $= 1$ ) at $m_1 = m_2$
2D equal-mass geometry	$90°$ separation	one ball at rest initially; head-on is the exception
symmetric viewpoint	CM frame: speeds unchanged, directions rotate	the collision reduced to one angle

Limiting cases

Case	Result	Meaning
equal masses, target at rest	$v_1' = 0$ , $v_2' = v_1$	complete velocity exchange — Newton's cradle, neutron on proton
light projectile, heavy target	$v_1' \to -v_1 + 2 v_2$ , $v_2' \approx v_2$	target is a moving wall; ball reverses relative to it
heavy projectile, light target	$v_1' \to v_1$ , $v_2' \to 2 v_1$	projectile unbothered; target flung at nearly double speed
wall at rest ( $m_2 \to \infty$ , $v_2 = 0$ )	$v_1' = -v_1$	perfect bounce, zero energy lost to the wall
any masses	$v_1 - v_2 = -(v_1' - v_2')$	the universal elastic signature — masses cancel
restitution $e < 1$	separation $= e \times$ approach	real collisions interpolate toward the sticky limit $e = 0$

Common trap“Elastic means the collision barely affects things”

The word suggests gentleness — as if an elastic collision is one where the incoming ball shrugs and continues. The truth is nearly opposite: maximum energy transfer happens in an elastic collision. Hit a resting equal mass head-on and the elastic outcome is the most violent possible for the projectile: it stops dead, donating 100% of its kinetic energy — that is the equal-mass panel of the hero figure, and the reason a single head-on bounce off a proton can thermalize a fast neutron. The energy-transfer fraction $4m_1 m_2/(m_1+m_2)^2$ peaks at exactly $m_1 = m_2$ . What elasticity forbids is not change but loss: every joule stays in visible motion instead of leaking into heat. Two companion errors to purge: (i) forgetting that the trivial root " $v_1' = v_1$ , nothing happened" always solves the equations and must be discarded; (ii) using the speed-preserving wall-bounce intuition on comparable masses, where it badly fails — the reversal rule preserves relative speed, not either body's own speed.

Worked examples

1 · Driver meets golf ball (confidence builder)

A driver clubhead of mass $200\ \mathrm{g}$ arrives at $40\ \mathrm{m/s}$ and strikes a $46\ \mathrm{g}$ golf ball at rest, head-on. Treating the impact as elastic and the clubhead as free (the shaft's grip force is negligible during the half-millisecond of contact), find both outgoing speeds.

Step 1. Identify the inputs: $m_1 = 0.200\ \mathrm{kg}$ , $v_1 = 40\ \mathrm{m/s}$ ; $m_2 = 0.046\ \mathrm{kg}$ , $v_2 = 0$ . Mass sum: $m_1 + m_2 = 0.246\ \mathrm{kg}$ .

Step 2. Ball speed from (5): $v_2' = \dfrac{2 m_1}{m_1 + m_2} v_1 = \dfrac{2 \times 0.200}{0.246} \times 40 = 1.626 \times 40 = \boxed{65\ \mathrm{m/s}}$ — the ball leaves faster than the club ever moved, at $1.63$ times the clubhead speed, chasing the heavy-on-light limit of $2v_1$ .

Step 3. Clubhead speed from (4): $v_1' = \dfrac{m_1 - m_2}{m_1 + m_2} v_1 = \dfrac{0.154}{0.246} \times 40 = \boxed{25\ \mathrm{m/s}}$ — the club keeps most of its speed, feeling only a firm deceleration.

Step 4. Audit both ledgers. Momentum: before $0.200 \times 40 = 8.0\ \mathrm{kg\,m/s}$ ; after $0.200 \times 25.0 + 0.046 \times 65.0 = 5.01 + 2.99 = 8.0$ ✓. Energy: before $\tfrac12 (0.200)(40)^2 = 160\ \mathrm{J}$ ; after $\tfrac12(0.200)(25.0)^2 + \tfrac12(0.046)(65.0)^2 = 62.6 + 97.3 \approx 160\ \mathrm{J}$ ✓. (Real drives lose about 20% — clubface restitution $e \approx 0.83$ — but the elastic calculation sets the ceiling, and golf's governing bodies literally regulate $e$ .)

2 · Slowing a reactor neutron (exam level)

A fission neutron (mass $1$ u) leaves uranium with kinetic energy $2\ \mathrm{MeV}$ and must be moderated to thermal energy, $0.025\ \mathrm{eV}$ . For head-on elastic collisions with nuclei at rest, how many collisions does this take if the moderator is carbon-12? Why is hydrogen dramatically better?

Step 1. Find the energy the neutron keeps in one head-on bounce. With target at rest, its outgoing speed from (4) is $v_1' = \frac{m_1 - m_2}{m_1 + m_2} v_1$ , so the retained energy fraction is the square: $\dfrac{K'}{K} = \left( \dfrac{m_1 - m_2}{m_1 + m_2} \right)^2$ .

Step 2. Carbon-12 target ( $m_2 = 12$ u): $\dfrac{K'}{K} = \left( \dfrac{1 - 12}{13} \right)^2 = \left( \dfrac{11}{13} \right)^2 = 0.716$ . Each head-on bounce keeps 71.6% — carbon is a grudging moderator, shaving only 28% per hit.

Step 3. Count the collisions: need $0.716^{\,n} = \dfrac{0.025\ \mathrm{eV}}{2 \times 10^6\ \mathrm{eV}} = 1.25 \times 10^{-8}$ . Take logs: $n = \dfrac{\ln(8 \times 10^7)}{\ln(1/0.716)} = \dfrac{18.2}{0.334} = \boxed{\approx 55\ \text{head-on collisions}}$ . (Real collisions are mostly glancing, transferring less; the angle-averaged reactor-physics answer for carbon is about 115 collisions — same physics, geometry included.)

Step 4. Hydrogen target ( $m_2 \approx m_1 = 1$ u): the retained fraction is $\left( \dfrac{1-1}{2} \right)^2 = 0$ — a single head-on collision stops the neutron completely, donating all $2\ \mathrm{MeV}$ to the proton. This is the equal-mass exchange rule wearing a hard hat. Angle-averaged, water still needs only $\sim 20$ collisions — five times fewer than carbon.

Step 5. Read the design consequence: moderators must be light — water, heavy water, graphite. A lead moderator ( $m_2 = 207$ ) would keep $\left(\frac{206}{208}\right)^2 = 98.1\%$ per bounce: the neutron would ricochet thousands of times, which is precisely why heavy materials make good neutron reflectors and terrible moderators.

Origins · 1 minChristiaan Huygens and how this was discovered

Christiaan Huygens

1629–1695 · The Hague, Paris, London

Huygens solved elastic collisions a generation before Newton's Principia, and he did it with a thought experiment of enduring brilliance. Descartes — Huygens' childhood tutor's idol, and a family friend — had published collision rules that were confidently wrong, because his conserved "quantity of motion" carried no sign. Huygens found the correct laws by imagining a collision performed on a canal boat gliding at constant speed while an observer watches from the bank: the same event, seen from two frames, must obey the same physics, and by choosing the boat's speed cleverly he could turn any collision into a symmetric one he already understood. That maneuver — solve it in the convenient frame, transform back — is precisely the center-of-mass method of this article, invented before frames had a name. When the Royal Society posed the collision problem as a challenge in 1668, Huygens, Wallis, and Wren returned the correct answer nearly simultaneously; Huygens' unpublished manuscript De Motu Corporum ex Percussione had held the complete solution for over a decade, including the theorem that equal masses exchange velocities and the observation that $\sum mv^2$ survives elastic impacts — the first conservation-of-kinetic-energy statement in history, half a century before the term "energy" existed.

History

1644

— Descartes publishes collision rules based on unsigned "quantity of motion"; most are experimentally wrong, but the dream of a conservation law is planted.

1656

— Huygens works out the correct elastic-collision laws, including equal-mass velocity exchange, in De Motu Corporum ex Percussione — circulated privately, published only in 1703.

1668

— The Royal Society challenges Europe's mathematicians on collisions; Wallis (inelastic), Wren, and Huygens (elastic) deliver the correct, sign-respecting answers.

1687

— Newton's Principia embeds the collision laws in the general framework of the three laws of motion, crediting the 1668 trio.

1911

— Rutherford reads the atom's structure from elastic scattering of alpha particles — large-angle bounces reveal a tiny, massive nucleus.

1932

— Chadwick weighs the neutron by letting it collide elastically with hydrogen and nitrogen, measuring recoil speeds and inverting the collision formulas.

Going further

Texts. Morin, Introduction to Classical Mechanics, Ch. 5 (Momentum) — 1D and 2D elastic collisions, CM frame · Goldstein, Poole & Safko, Classical Mechanics, 3rd ed., §3.10–3.11 — two-body scattering and the lab/CM transformation · Landau & Lifshitz, Mechanics, §16–17 · Taylor, Classical Mechanics, Ch. 14 — collision theory and cross sections · Kleppner & Kolenkow, Ch. 4.

Provenance. All formulas are standard textbook results. A verifier should: (1) re-derive the 1D final velocities by solving momentum + energy and confirm the boxed expressions (Morin Ch.5; Taylor Ch.14); (2) check every limiting case in the 'limits' section by substituting mass ratios; (3) verify the equal-mass 90° result via the dot-product argument (Goldstein); (4) confirm the lab/CM angle relation $\tan\theta_{\mathrm{lab}}=\sin\theta^*/(\cos\theta^*+m_1/m_2)$ against Goldstein §3.11. No empirical numbers are asserted beyond the illustrative worked examples, whose arithmetic (momentum and KE balances) is self-checking. Worked-example numbers on this page (golf: $v_2' = 65\ \mathrm{m/s}$ , $v_1' = 25\ \mathrm{m/s}$ , both ledgers rebalanced to $8.0\ \mathrm{kg\,m/s}$ and $160\ \mathrm{J}$ ; neutron: retained fraction $(11/13)^2 = 0.716$ , $\approx 55$ head-on collisions for carbon vs one for hydrogen) are illustrative arithmetic recomputed from the boxed formulas.

Builds on: Newton's Third Law · Newton's Second Law · Relative Motion and Galilean Transformations · Vectors in Kinematics and Coordinate Systems · One-Dimensional Kinematics. Leads to: Coefficient of Restitution · Two-Body Problem and Reduced Mass · Rutherford Scattering · Kinetic Theory of Gases.

Do perfectly elastic collisions actually exist, or are they a textbook fiction?

At the atomic scale they are routine reality. Two colliding gas atoms at room temperature cannot dissipate energy: heat "loss" means exciting internal motion, and the first excited state of an atom sits electron-volts above the ground state while thermal collisions carry hundredths of an electron-volt — the energy leak is quantum-mechanically forbidden. That is why the kinetic theory of gases can treat molecular collisions as perfectly elastic. Macroscopically, elasticity is only approached: a superball manages $e \approx 0.9$ , hardened steel balls $\approx 0.95$ , and every real "click" of a Newton's cradle leaks a little energy into sound — which is precisely why the cradle eventually stops.

Why does maximum energy transfer happen at equal masses? Intuitively a heavier projectile should hit harder

A heavier projectile carries more energy but donates a smaller fraction. The transfer fraction to a resting target is $4 m_1 m_2/(m_1 + m_2)^2$ , which peaks at exactly 1 when $m_1 = m_2$ and falls off in both directions. The two failure modes are instructive: a light projectile bounces back, keeping most of its energy; a heavy one plows through, and although it flings the target fast ( $\approx 2v_1$ ), the target's tiny mass means that speed carries little energy. Equal masses are the impedance-matched case — the collision analog of matching resistances to move maximum power. Neutron moderation (Worked Example 2) is this formula turned into reactor engineering.

The reversal rule says approach speed equals separation speed. Doesn't that mean nothing really changed?

The relative speed is preserved; each body's own velocity can change drastically — that is the entire content of the collision. Equal masses: relative speed before is $v$ , after is $v$ , yet the projectile went from $v$ to 0 and the target from 0 to $v$ . The rule is a constraint, not a description of stasis. It is also the fastest practical solver: combined with momentum conservation it gives two linear equations, sidestepping the quadratic energy condition entirely. And it generalizes gracefully to real collisions: separation speed $= e \times$ approach speed, with the coefficient of restitution $e$ grading the bounce from 1 (elastic) to 0 (stuck together).

How can the golf ball leave faster than the clubhead was ever moving? Where does the extra speed come from?

No ledger is violated — speed is not a conserved quantity; momentum and energy are. Watch it from the clubhead's frame: the ball approaches at $40\ \mathrm{m/s}$ , and the reversal rule sends it back out at $40\ \mathrm{m/s}$ relative to the club. Transform back to the ground frame and those $40$ add to the club's own $\approx 25\ \mathrm{m/s}$ : about $65\ \mathrm{m/s}$ . The heavy-light limit makes the bound clean: a very heavy moving object can fling a light resting one at up to twice its own speed, never more. The same frame argument powers the gravitational slingshot — Jupiter is the clubhead, the spacecraft is the ball.

Why does Newton's cradle eject exactly one ball instead of two balls at half speed?

Two balls at half speed would conserve momentum perfectly — but not energy. Check: one ball arriving at $v$ carries $p = mv$ and $K = \tfrac12 mv^2$ . Two balls leaving at $v/2$ carry the same $p = mv$ but only $K = \tfrac12 (2m)(v/2)^2 = \tfrac14 mv^2$ — half the energy has vanished, which elasticity forbids. The only outcome satisfying both ledgers (for sequential pairwise collisions of equal masses) is complete velocity exchange propagating down the chain: each ball hands its motion to the next, and one ball exits with the full $v$ . Lift two balls and the same logic ejects exactly two. The cradle is a mechanical proof that momentum conservation alone under-determines collisions.

In 2D the conservation laws don't fix the outcome. What actually decides the scattering angle?

The impact parameter — how off-center the encounter is — plus the actual force law between the bodies. Conservation laws are bouncers, not choreographers: they rule out illegal outcomes but leave a one-parameter family of legal ones, labeled by the CM-frame scattering angle. Aim dead-center and you recover the 1D head-on results; graze and both bodies barely deflect. This "missing" freedom is not a defect — it is the information channel that scattering experiments read. Rutherford inferred the nuclear atom precisely by measuring how many alpha particles came out at each angle and comparing against the angle-vs-impact-parameter map of the Coulomb force.

Is the 90-degree rule why pool players can predict where balls go?

Yes — for any non-head-on hit, cue ball and object ball separate along perpendicular lines (equal masses, target at rest, elastic). Players learn it as the "tangent line": after contact the cue ball departs along the tangent to the two balls at the moment of touch, at right angles to the object ball's line. Real tables bend the rule in known ways: follow or draw (topspin/backspin) makes the cue ball curve off the tangent line after contact, because spin couples in energy the point-mass analysis ignores, and the balls' contact is only $e \approx 0.95$ elastic. The right angle is the honest baseline that spin adjustments are measured against.

What exactly does the coefficient of restitution measure, and is it a property of one object?

$e$ is the separation-to-approach speed ratio of a specific colliding pair: $e = |v_2' - v_1'| / |v_1 - v_2|$ . It grades the bounce between the elastic ideal ( $e = 1$ , energy fully repaid) and perfect stickiness ( $e = 0$ , bodies merge — the subject of the inelastic-collisions article). It is a property of the pair and the conditions, not one body: a golf ball on a clubface, a basketball on hardwood, and the same basketball on grass all have different $e$ . Temperature and speed matter too — squash players warm the ball up precisely to raise its $e$ . Kinetic-energy loss follows directly: for a resting target, the fraction $(1 - e^2)$ of the CM-frame kinetic energy is dissipated.