ARCHETYPE: River–Boat Crossing

Watch a ferryman work a wide river and you will see a small piece of applied genius. He never points his boat at the jetty he wants to reach — he aims upstream of it, at what looks like the wrong target, holds that heading, and the current quietly carries him sideways onto a perfectly straight track to the far bank. Ask him why and he will shrug: everyone who works on water learns it. What he has internalized is vector addition — the fact that his boat's velocity over the ground is the sum of what his engine does relative to the water and what the water does relative to the Earth. Examiners have loved this setup for decades, and with reason. One river, one boat, two speeds — and out of that tiny cast come at least four genuinely different questions: cross in the least time, cross by the shortest path, find the drift, survive the case where the river is faster than you. CBSE and ISC pose the direct versions; JEE-Main flips them into inverse problems (here are two crossing times, find the river); JEE-Advanced brings the sly one where zero drift is impossible and you must minimize it by pure geometry; NEET keeps the 3-4-5 classics on rotation. This article treats the river crossing as what it is — a problem archetype worth mastering once, completely. One frame equation, one axis choice, and every variant reduces to reading off components. Learn the master pattern here, and the sibling setups — Rain & the Moving Observer, aircraft in wind — become the same problem wearing different clothes.

Key takeaways

Everything follows from one frame equation: $\vec{v}_{b,g} = \vec{v}_{b,w} + \vec{v}_{w,g}$ — boat over ground = boat relative to water + water over ground.
Choose axes across and along the river. Crossing time depends only on the across component: $t = \dfrac{d}{v_b\cos\theta}$ ; the current cannot speed up or delay the crossing.
Minimum time: aim straight across ( $\theta = 0$ ), $t_{\min} = d/v_b$ , and accept the drift $x = v_r d/v_b$ .
Zero drift (shortest path): aim upstream at $\sin\theta = v_r/v_b$ — possible only when $v_b > v_r$ ; then $t = d/\sqrt{v_b^2 - v_r^2}$ .
If the river is faster ( $v_r > v_b$ ), zero drift is impossible; drift is least when $\sin\theta = v_b/v_r$ , giving $x_{\min} = \dfrac{d\sqrt{v_r^2 - v_b^2}}{v_b}$ .
Key idea: split the boat's water-relative velocity into across and along components — the across component alone fights the width, the along component alone fights the current.

The core insight

A boat's velocity over the ground is the vector sum of its velocity relative to the water and the water's velocity over the ground; resolving that sum across and along the river turns every crossing question into one-line component arithmetic.

Where this lives today

Watch any aircraft land in a crosswind and you are watching Example 2 at 250 km/h: the nose points visibly into the wind — pilots call it the crab angle — so that the velocity over the ground tracks the runway centreline. Flight computers solve $\sin\theta = v_{\text{wind}}/v_{\text{air}}$ continuously; on gusty days pilots fly it by feel. Ferries across tidal straits run the same correction against water, and GPS units display the very distinction this article is built on: heading (where you point) versus course over ground (where you actually go). Delivery drones inherit the full archetype: route planners choose between minimum-time and minimum-drift crossings of a windy corridor exactly as the swimmer of Example 4 does, and when the wind exceeds the drone's airspeed, the tangent construction decides the least-bad track.

The setup

Geometry first, because in an archetype the geometry is the physics. A river of width $d$ flows with uniform velocity $v_r$ parallel to its straight banks. A boat (or swimmer) moves with speed $v_b$ relative to the water — the still-water speed, what the engine or the swimmer's strokes actually produce. The boat leaves one bank aiming at an angle $\theta$ measured from the normal to the banks (the straight-across direction), tilted upstream.

Figure

The river crossing — geometry, aim angle, and the three classic strategies

Set up axes once and never again: call the straight-across direction the $y$ -axis and the downstream direction the $x$ -axis. The frame equation of the whole chapter — the one you met in Relative Velocity & Reference Frames — chains the subscripts:

\vec{v}_{b,g} = \vec{v}_{b,w} + \vec{v}_{w,g} \qquad (1)

where: $\vec{v}_{b,g}$ — the boat's velocity over the ground (what an observer on the bank sees, and what actually carries the boat to the far side); $\vec{v}_{b,w}$ — the boat's velocity relative to the water, magnitude $v_b$ , direction chosen by the boatman; $\vec{v}_{w,g}$ — the current, magnitude $v_r$ , always downstream. Read it physically: the water is a moving carpet. The engine decides how you move on the carpet; the carpet decides how the carpet itself moves; the ground sees the sum. Two facts make the archetype solvable in your head. The current has no across component. It flows parallel to the banks, so it can never help you toward the far bank nor hold you back — it only slides you sideways. The components separate completely. With the boat aimed upstream at angle $\theta$ from the normal, the ground velocity has components

v_y = v_b\cos\theta, \qquad v_x = v_r - v_b\sin\theta \qquad (2)

where: $v_y$ — the across component, the only thing eating the width $d$ ; $v_x$ — the net downstream component, the only thing producing drift. Every river-boat question ever set is a question about equation (2): the crossing time interrogates $v_y$ , the drift interrogates $v_x$ , and the optimizations ask you to trade one against the other.

The master solution pattern

The crossing ends when the boat has covered the width $d$ across. Only $v_y$ contributes, so the crossing time is

t = \frac{d}{v_b\cos\theta} \qquad (3)

where: $\theta$ — the aim angle from the normal. Notice what is absent: $v_r$ . The current does not appear, because it has nothing to offer in the across direction. Aiming upstream ( $\theta > 0$ ) always lengthens the crossing — you are spending some of your $v_b$ budget on fighting the current instead of on crossing. While the boat is in the water for time $t$ , the net downstream component carries it sideways by the drift

x = \left(v_r - v_b\sin\theta\right)\frac{d}{v_b\cos\theta} \qquad (4)

where: $x$ — how far downstream of the start-opposite point the boat lands ( $x<0$ would mean landing upstream — the boatman over-corrected). Equations (3) and (4) are the whole archetype; the named cases are just choices of $\theta$ .

strategy	aim	crossing time	drift
minimum time	straight across, $\theta = 0$	$t_{\min} = \dfrac{d}{v_b}$	$x = \dfrac{v_r d}{v_b}$
zero drift (shortest path), needs $v_b > v_r$	upstream, $\sin\theta = \dfrac{v_r}{v_b}$	$t = \dfrac{d}{\sqrt{v_b^2 - v_r^2}}$	$0$
minimum drift when $v_r > v_b$	upstream, $\sin\theta = \dfrac{v_b}{v_r}$	$t = \dfrac{d}{v_b\cos\theta}$	$x_{\min} = \dfrac{d\sqrt{v_r^2 - v_b^2}}{v_b}$
straight up / down the river	along the bank	speeds $v_b \mp v_r$	—

Derivation AFastest crossing versus shortest path

We derive the two headline strategies — minimum crossing time and zero drift — directly from the component pair in equation (2). The strategy is to notice that time and drift are controlled by different components, so each optimization is a one-component problem. Watch the logic of the zero-drift case: we are not minimizing anything by calculus; we are imposing a condition ( $v_x = 0$ ) and asking what aim angle achieves it.

Step 1. Crossing time. The width $d$ is consumed at rate $v_y = v_b\cos\theta$ , so $t = d/(v_b\cos\theta)$ . The factor $\cos\theta$ is largest at $\theta = 0$ : any tilt of the aim, upstream or downstream, slows the crossing. Why: the current contributes nothing across the river, so the fastest crossing simply devotes the entire still-water speed to the across direction. Aim straight at the far bank and go.

Step 2. Minimum time and its price. With $\theta = 0$ : $t_{\min} = d/v_b$ , and the full current acts for the whole crossing, giving drift $x = v_r t_{\min} = v_r d / v_b$ . Fast, but you land downstream.

Step 3. Zero drift. Demand $v_x = 0$ in equation (2): $v_r = v_b \sin\theta$ , so $\sin\theta = v_r / v_b$ . The upstream component of your own velocity exactly cancels the current, the ground track is the straight perpendicular line — the shortest possible path, length $d$ . Why sine and not tangent: the cancellation happens between the current $v_r$ and the along-river component of the boat's velocity, which is $v_b\sin\theta$ because $\theta$ is measured from the normal. Draw the velocity triangle: $v_b$ is the hypotenuse, $v_r$ the side opposite $\theta$ .

Step 4. Zero-drift crossing time. The surviving across component is $v_b\cos\theta = v_b\sqrt{1 - v_r^2/v_b^2} = \sqrt{v_b^2 - v_r^2}$ , so $t = d/\sqrt{v_b^2 - v_r^2}$ — always longer than $t_{\min}$ . The condition $\sin\theta \le 1$ demands $v_b \ge v_r$ : you cannot cancel a current faster than yourself. Verified: $v_b = 5$ m/s, $v_r = 3$ m/s, $d = 200$ m. Minimum time: $t = 40$ s with drift $x = 3 \times 40 = 120$ m. Zero drift: $\sin\theta = 3/5$ , $\theta = 37°$ , $t = 200/\sqrt{25-9} = 200/4 = 50$ s. Shortest path costs 10 extra seconds — both numbers reappear in the worked examples.

Go deeper · Derivation B · 2 minMinimum drift when the river is faster than you

Derivation BMinimum drift when the river is faster than you

We find the best possible landing point when $v_r > v_b$ , the JEE-Advanced twist where zero drift is unattainable. The tool is a velocity diagram, not calculus: the boatman's choice of heading makes the tip of the ground-velocity vector trace a circle, and the best heading is a tangent line. This geometric construction is worth owning — it recurs in rain problems and in projectile optimizations.

Step 1. Draw the current $\vec{v}_{w,g}$ as a fixed horizontal arrow $\vec{OC}$ of length $v_r$ . The boat adds to it a vector of fixed length $v_b$ in any direction, so the tip $G$ of the total $\vec{v}_{b,g} = \vec{OG}$ lies somewhere on a circle of radius $v_b$ centred at $C$ . Because $v_b < v_r$ , the point $O$ lies outside this circle.

Step 2. Drift is smallest when the ground track points as nearly across the river as possible — when $\vec{OG}$ makes the largest angle with the downstream direction $\vec{OC}$ . Swing the line $OG$ around: the extreme is the tangent from $O$ to the circle. Any steeper line misses the circle entirely — no heading can produce it.

Step 3. Read the tangent triangle. In right triangle $OGC$ (right angle at $G$ , the tangency point): $\sin\alpha = v_b / v_r$ , where $\alpha$ is the angle between the ground track and the downstream direction. The boat's heading at tangency is perpendicular to $OG$ — and its own aim angle from the normal satisfies the twin relation $\sin\theta = v_b / v_r$ . Why the right angle: a tangent to a circle is perpendicular to the radius at the point of contact, and the radius at $G$ is exactly the boat's velocity relative to the water. So at minimum drift the boat's water-velocity is perpendicular to its ground track — a striking, examinable fact.

Step 4. Extract the minimum drift. With $\sin\theta = v_b/v_r$ : the across component is $v_b\cos\theta = v_b\sqrt{v_r^2 - v_b^2}/v_r$ , giving $t = \dfrac{d\,v_r}{v_b\sqrt{v_r^2 - v_b^2}}$ , and the residual downstream component is $v_r - v_b\sin\theta = (v_r^2 - v_b^2)/v_r$ . Multiply:

x_{\min} = \frac{v_r^2 - v_b^2}{v_r}\cdot\frac{d\,v_r}{v_b\sqrt{v_r^2 - v_b^2}} = \frac{d\sqrt{v_r^2 - v_b^2}}{v_b}

Verified: swimmer $v_b = 3$ m/s in a $v_r = 5$ m/s river, $d = 120$ m. $\sin\theta = 3/5$ , $\theta = 37°$ upstream of the normal; across component $3(4/5) = 2.4$ m/s, so $t = 50$ s; residual drift rate $(25-9)/5 = 3.2$ m/s, so $x_{\min} = 160$ m — and the formula gives $120 \times 4 / 3 = 160$ m. Consistent.

Figure

Minimum drift by tangent construction — the tip of the ground velocity lives on a circle

The variants examiners actually pose

Up and down the river. Point along the bank and the vectors collapse to one dimension: ground speed $v_b + v_r$ downstream, $v_b - v_r$ upstream (motion possible only if $v_b > v_r$ ). A there-and-back trip of length $L$ each way takes

t_{\updownarrow} = \frac{L}{v_b - v_r} + \frac{L}{v_b + v_r} = \frac{2Lv_b}{v_b^2 - v_r^2} \qquad (5)

where: the denominator $v_b^2 - v_r^2 < v_b^2$ guarantees the round trip is always slower than in still water ( $2L/v_b$ ) — the current's help downstream never repays the time lost crawling upstream, because you spend longer on the slow leg. The same harmonic-mean logic runs the average-speed traps in Speed & Velocity.

The swim-then-walk hybrid. A JEE favourite: cross in minimum time, land downstream, then walk back along the bank at walking speed $u$ to the point opposite. Total time $= d/v_b + v_r d/(v_b u)$ . Whether this beats the zero-drift swim depends on $u$ — a lovely "compare two strategies" question that is pure equation (3)-and-(4) bookkeeping.

Boat and raft. A raft is a free-floating object: it moves with the water, $\vec{v}_{\text{raft},g} = \vec{v}_{w,g}$ . Relative to the water the raft is at rest, so a boat leaving a raft and returning to it takes equal times out and back in the water frame — the classic "boatman drops a hat" problem solves in one line if you think in the water's frame.

Limiting cases

case	result	meaning
$v_r = 0$	$t = d/v_b$ , zero drift for $\theta = 0$	still water: aim where you want to go
$v_r \to v_b$	zero-drift aim $\theta \to 90°$ , $t \to \infty$	all effort fights the current, none crosses
$v_b \gg v_r$	drift $\to 0$ even aiming straight	fast boat barely notices the river
$v_r > v_b$	zero drift impossible; $x_{\min} = d\sqrt{v_r^2-v_b^2}/v_b$	you cannot cancel a current faster than yourself
$\theta = 90°$ (aim dead upstream)	$v_y = 0$ : never crosses	ground speed $v_r - v_b$ , pure drift

Common trap“Shortest time and shortest path are the same crossing”

They are different crossings with different aims, different times, and different landings — and mixing them is the single most reliable way to lose river-boat marks. Shortest time means aim straight across: $t_{\min} = d/v_b$ , and you accept the drift. Shortest path means land exactly opposite: aim upstream at $\sin\theta = v_r/v_b$ , and you accept a longer time $d/\sqrt{v_b^2 - v_r^2}$ . For $v_b = 5$ , $v_r = 3$ , $d = 200$ m: fastest crossing $40$ s landing $120$ m downstream; shortest path $50$ s landing on target. If a question says "reaches the opposite point", it has chosen the second for you; if it says "in minimum time", the first. Read the sentence before you pick the formula.

Worked examples

1 · Board level — aim straight, find the landing (CBSE/ISC pattern, 3 marks)

A boat can move at $4.0$ m/s in still water. It heads straight across a river $240$ m wide flowing at $3.0$ m/s. Find the time to cross, the distance downstream at which it lands, and its resultant speed.

Step 1. Set axes: across ( $y$ ) and downstream ( $x$ ). Aimed straight: $v_y = 4$ m/s, and the current alone supplies $v_x = 3$ m/s.

Step 2. Crossing time uses only the across component: $t = d/v_y = 240/4 = \boxed{60\ \mathrm{s}}$ . The current is irrelevant to this part — say so, it earns the method mark.

Step 3. Drift: $x = v_x t = 3 \times 60 = \boxed{180\ \mathrm{m}}$ downstream.

Step 4. Resultant speed: $\sqrt{4^2 + 3^2} = \boxed{5.0\ \mathrm{m/s}}$ , directed $\tan^{-1}(3/4) = 37°$ downstream of the straight-across line. Sanity: the track length is $5 \times 60 = 300$ m — and $\sqrt{240^2 + 180^2} = 300$ m ✓.

2 · Board level — land exactly opposite (ISC/CBSE pattern)

A swimmer's still-water speed is $5.0$ m/s; the river flows at $3.0$ m/s and is $200$ m wide. At what angle must she aim to reach the point directly opposite, and how long does the crossing take? Compare with her fastest possible crossing.

Step 1. Zero drift demands the upstream component cancel the current: $v_b\sin\theta = v_r$ , so $\sin\theta = 3/5$ and $\theta = \boxed{37°\ \text{upstream of straight across}}$ .

Step 2. Surviving across component: $v_b\cos\theta = 5 \times \tfrac{4}{5} = 4.0$ m/s.

Step 3. Time: $t = 200/4 = \boxed{50\ \mathrm{s}}$ .

Step 4. Fastest crossing (aim straight): $t_{\min} = 200/5 = 40$ s, but she lands $3 \times 40 = 120$ m downstream. The straight path costs $10$ s of "insurance premium".

3 · Competitive level — two times, find the river (JEE-Main pattern, ~2.5 minutes)

A man crosses a $300$ m river in $10$ minutes when he swims to minimize time, and in $12.5$ minutes when he swims to reach the point directly opposite. Find his still-water speed and the river's speed.

Step 1. Minimum time is width over still-water speed: $v_b = d/t_1 = 300/600 = \boxed{0.50\ \mathrm{m/s}}$ .

Step 2. The zero-drift crossing uses the reduced across speed: $\sqrt{v_b^2 - v_r^2} = d/t_2 = 300/750 = 0.40$ m/s.

Step 3. Solve: $v_r^2 = 0.5^2 - 0.4^2 = 0.09$ , so $v_r = \boxed{0.30\ \mathrm{m/s}}$ .

Step 4. Free check: his zero-drift aim was $\sin\theta = 0.3/0.5 = 3/5$ — the 3-4-5 triangle again, $\theta = 37°$ . Timing note: budget ~2 minutes. The entire question is recognizing which time corresponds to which formula — $t_{\min} \leftrightarrow v_b$ , $t_{\text{opposite}} \leftrightarrow \sqrt{v_b^2 - v_r^2}$ . Inverse problems like this are why you memorize the pair, not re-derive it.

4 · Competitive level — the river wins (JEE-Advanced stretch, ~4 minutes)

A swimmer with still-water speed $3.0$ m/s attempts a river flowing at $5.0$ m/s, width $120$ m. Show she cannot land opposite her start, find the aim that minimizes her drift, and compute that minimum drift and her crossing time.

Step 1. Zero drift would need $\sin\theta = v_r/v_b = 5/3 > 1$ — impossible. No aim cancels a current faster than the swimmer.

Step 2. Tangent construction (or the memorized twin): minimum drift at $\sin\theta = v_b/v_r = 3/5$ , i.e. $\theta = \boxed{37°}$ upstream of the normal. At this aim her water-velocity is perpendicular to her ground track.

Step 3. Across component: $3\cos 37° = 3 \times \tfrac{4}{5} = 2.4$ m/s, so $t = 120/2.4 = \boxed{50\ \mathrm{s}}$ .

Step 4. Residual downstream rate: $v_r - v_b\sin\theta = 5 - 3\times\tfrac{3}{5} = 3.2$ m/s. Minimum drift: $x_{\min} = 3.2 \times 50 = \boxed{160\ \mathrm{m}}$ — matching the formula $d\sqrt{v_r^2 - v_b^2}/v_b = 120 \times 4/3 = 160$ m. Timing note: ~3.5 minutes if you know $\sin\theta = v_b/v_r$ cold; do NOT differentiate the drift expression in the exam hall — the tangent picture replaces the calculus. Spot the giveaway "river faster than swimmer" and switch formulas immediately.

Exam traps

Dividing the width by the wrong speed. The crossing time is $d$ over the across component — $d/v_b$ when aimed straight, $d/(v_b\cos\theta)$ in general. Students divide by the resultant $5$ m/s in a 3-4-5 problem and get $40$ s where $60$ s is right. The resultant governs the track length, never the crossing time.

Measuring $\theta$ from the bank instead of the normal. Every formula here takes $\theta$ from the straight-across direction. Some textbooks quote the same physics with the angle from the bank ( $90° - \theta$ ), turning $\sin$ into $\cos$ . Anchor yourself: zero drift, aim upstream at $\sin\theta = v_r/v_b$ from the normal — and re-derive in the velocity triangle if a question uses the other convention.

Assuming zero drift is always available. The condition $\sin\theta = v_r/v_b$ requires $v_b \ge v_r$ . When the river is faster, the honest answers are "impossible" plus the minimum-drift aim $\sin\theta = v_b/v_r$ — the ratio flips. Writing $\sin\theta = 5/3$ and pressing on is an automatic zero.

Sine–tangent confusion in the aim condition. The zero-drift aim satisfies $\sin\theta = v_r/v_b$ (components of the boat's own velocity), while the straight-aim drift direction satisfies $\tan\phi = v_r/v_b$ (components of the ground velocity). Two different triangles. If your $\theta$ comes out equal to your $\phi$ , you have mixed them.

Sign slips in the drift. Drift rate is $v_r - v_b\sin\theta$ : positive means landing downstream, negative upstream. In "reaches a point $x$ m downstream" inverse problems, keep the sign and solve; guessing which term is bigger costs more time than the algebra.

Round-trip averaging. Upstream-downstream trips obey the harmonic combination of $v_b - v_r$ and $v_b + v_r$ , never the arithmetic mean $v_b$ . If your round-trip time equals the still-water time, you have averaged speeds — redo with equation (5).

Origins · 1 minChristiaan Huygens and how this was discovered

Christiaan Huygens

1629–1695 · The Hague and Paris

Christiaan Huygens — pendulum-clock inventor, discoverer of Titan, wave theorist of light — was also the first physicist to make relativity of motion do real work. In his collision treatise De motu corporum ex percussione he imagined an experimenter performing a collision on a smoothly gliding Dutch canal boat while a colleague watched from the bank: one experiment, two frames, two sets of velocities — and any correct law of collisions had to satisfy both observers at once. With that boat-and-bank argument he derived the collision rules that had defeated Descartes, effectively wielding the velocity-addition equation (1) as a theoretical instrument. The examiner's river crossing is Huygens' canal boat with the roles reversed: same water, same frames, same vector sum.

History

1632

— Galileo's Dialogue argues motion below deck of a smooth ship is undetectable: the principle behind adding frame velocities.

1656

— Huygens drafts De motu corporum ex percussione, deriving collision laws by comparing canal-boat and bank observers (published 1703).

1687

— Newton's Principia, Corollary V: the laws of motion hold identically in any uniformly moving frame.

1905

— Einstein's special relativity replaces Galilean velocity addition near light speed; at river speeds the correction is one part in

10^{17}

1920

s — aircraft navigators formalize the "wind triangle" — the river-boat diagram drawn in the sky — as standard dead-reckoning practice.

Going further

Texts. NCERT Physics Class 11, Ch. 4 (Motion in a Plane, relative velocity in two dimensions) · HC Verma, Concepts of Physics Vol. 1, Ch. 3 (the river figures) · Irodov, Problems in General Physics, §1.1 for the minimum-drift stretch.

Provenance. All results are standard kinematics. A verifier should recompute: the table's four cases from equations (2)–(4); Example 1's $60$ s / $180$ m / $5$ m/s; Example 2's $37°$ and $50$ s vs $40$ s; Example 3's $v_b = 0.5$ , $v_r = 0.3$ m/s; Example 4's $37°$ , $50$ s, $160$ m (both routes); the round-trip formula (5).

Prerequisites. Relative Velocity & Reference Frames · Components & Resolution of Vectors. Leads to. Rain & the Moving Observer · Motion in a Plane (aircraft-in-wind problems) · the water-frame trick returns in Man on Boat (momentum).

Why doesn't the current change the crossing time at all?

Because the current has no component across the river — it is parallel to the banks by definition of the problem. Time to cross is width divided by the across component of velocity, and the only vector with an across component is the boat's own water-relative velocity. The current decides where you land, never when. This clean separation is the whole reason the axes are chosen along and across the flow.

Is θ measured from the bank or from the line straight across?

In this article — and in most Indian textbooks — from the line straight across (the normal to the banks). With that convention the across component is $v_b\cos\theta$ and the zero-drift condition is $\sin\theta = v_r/v_b$ . Some books measure from the bank, swapping sines and cosines. Don't memorize both: memorize one convention plus the velocity triangle, and re-derive in ten seconds if a diagram uses the other.

Why is the zero-drift condition a sine and not a tangent?

Because you are cancelling the current with a component of the boat's own velocity. The boat's water-speed $v_b$ is the hypotenuse of its component triangle; the upstream part is $v_b\sin\theta$ ; setting $v_b\sin\theta = v_r$ gives $\sin\theta = v_r/v_b$ . A tangent appears in a different triangle — the drift angle of the ground velocity when you aim straight across, $\tan\phi = v_r/v_b$ . Same letters, different triangles; drawing them is the cure.

What happens to the shortest-path crossing as the current approaches my swimming speed?

It degenerates gracefully. As $v_r \to v_b$ the aim angle $\sin\theta = v_r/v_b \to 1$ , so $\theta \to 90°$ : you point almost dead upstream, the across component $\sqrt{v_b^2 - v_r^2}$ shrinks toward zero, and the crossing time blows up. At exactly $v_r = v_b$ you would hold position forever. The formula is warning you the strategy is running out of physics.

Do swimmer problems and boat problems ever differ?

Physically they are identical — "speed in still water" plays the same role for both. The only difference is dressing: swimmer problems add walking back along the bank, boat problems add return trips. A raft is the special case $v = 0$ relative to water: it rides the current, which is why "boat meets raft" questions collapse in the water's frame.

Why is the upstream-downstream round trip slower than still water even though the current helps for half the distance?

It helps for half the distance but far less than half the time. You crawl upstream at $v_b - v_r$ for a long while and fly downstream at $v_b + v_r$ only briefly, so the slow leg dominates the average. Algebraically the round-trip average speed is $(v_b^2 - v_r^2)/v_b$ , which is below $v_b$ for any current — the same harmonic-mean penalty behind the "40 and 60 average 48" trap in Speed & Velocity.

In an exam, how do I decide which case a question is asking for?

Scan for the objective phrase. "In minimum/least time" → aim straight, $t = d/v_b$ , compute drift. "Reaches the point directly opposite" or "shortest path" → aim upstream, $\sin\theta = v_r/v_b$ , time $d/\sqrt{v_b^2 - v_r^2}$ . "River flows faster than…" → minimum drift, flip to $\sin\theta = v_b/v_r$ . "Rows upstream/downstream" → one-dimensional $v_b \mp v_r$ . Four phrases, four formulas — the reading is the physics.