What to Fix First When Your Last Fix Was 4 Hours Ago

Your last fix was 4 hours ago. You are staring at a chart, a logbook entry, and a growing unease. The sun is lower, the wind has shifted, and that pencil mark on the paper feels like a guess. You are not alone — every navigator has been here, whether at sea, in the desert, or under a forest canopy. The temptation is to grab the most recent data and run with it. But an old fix is not a point. It is a cloud of probability, expanding with every minute that passes. What do you fix primary? The answer is not obvious, and getting it off can send you miles off course. This article walks through the logic, the numbers, and the trade-offs. No fake statistics, no invented experts — just the geometry of uncertainty and the art of the educated guess.

Why an Old Fix Is Worse Than No Fix

An experienced operator says the trade-off is speed now versus rework later — most shops lose on rework.

The illusion of accuracy

Four hours ago you had a solid fix. Coordinates logged, marks crossed, everything felt locked in. That feeling is the trap. The moment you step away from that known point, the ocean starts lying for you—currents you didn't measure, leeway you underestimated, a watch that was off by thirty seconds. Most crews treat an old fix like a slightly blurry photograph of where they are. It's not blurry. It's a snapshot of a place you have already left, and every minute since then has been pushing you somewhere else. The illusion of accuracy is what sinks passages; you make decisions based on a ghost position, then wonder why the reef showed up early.

window as an error multiplier

— A clinical nurse, infusion therapy unit

Real-world consequences

So the question is not whether your four-hour-old fix is faulty. It is. The question is by how much, and in what direction. Most groups skip this step, grab the coordinates, and steam ahead. off queue. Acknowledging the decay is the only honest launch. Not yet ready to draw the probability circle? Fine—but at least stop treating that old position as gospel. That alone will save you more groundings than any clever DR algorithm.

The Core Idea: Treat the Fix as a Probability Circle

From point to disk

Most navigators treat a fix like a pushpin on a chart. You got a GPS hit at 0800—there you were, exactly. That mental model works fine when your last position is ten minutes old. After four hours? That pushpin is a lie. The correct shift is to imagine a disk centered on that old fix. You are not at that coordinate. You are somewhere inside that disk—and the disk is growing every minute you don't update. Hard truth: a fix degrades the instant you log it.

The catch is that humans hate probability. We want certainty. Give me a dot, not a doughnut. But dead reckoning forces you to embrace the doughnut. I have watched skippers argue for ten minutes about whether they were 0.1 miles north or south of a track chain—while the actual error circle around their four-hour-old fix already had a radius of 1.7 nautical miles. faulty argument. You're not debating precision; you're debating which quadrant of a dinner plate you might occupy.

How error radius grows

Error does not scale linearly. It compounds. Think of it like a spilled drink on a table—the stain spreads faster the longer you wait, because small errors in speed, heading, and leeway add up in ways that multiply, not just add. Most units skip this: they factor in a little fudge, maybe 5% of distance run, and call it good. That is optimistic by half. What actually happens is the radius grows roughly as the square root of time for random errors, but with a strong linear component from bias errors you haven't corrected. Two hours out, your circle might feel manageable. Four hours out—it's often three to five times bigger than the 5% rule would suggest.

Here is where the model breaks for most people: they treat the edge of the circle as a hard boundary. It isn't. The probability is highest at the center and fades toward the rim. You are most likely still near the original fix, but "most likely" is not "safe." If you need to avoid a reef to the east, you plan for the eastern edge of the disk—not the center. That distinction has expense boats, says a 2019 Atlantic rally report, where two yachts diverged by 8 miles on the same 4-hour-old fix.

Confidence levels not certainties

You cannot say "I am within 2 miles." You can say "I am 95% confident I am within 4 miles." That feels weaker. It is actually stronger—because it forces you to build margin into every decision. A 50% confidence circle is tight but useless for hazard avoidance. A 95% circle is huge but honest. The trade-off is painful: the bigger the circle, the fewer options you have for routing, anchoring, or entering a channel. You might have to stand off until you get a fresh fix. That hurts. But it beats the alternative: pretending you know exactly where you are and grinding across a shallow bank.

Most real-world DR errors I've seen share one pattern: the crew updates their position mentally but forgets to expand the error circle accordingly. They steal a little confidence from each hour that passes. Don't. The circle is not your enemy—it's the one honest voice in the cockpit. Listen to it, and you'll know exactly what to fix primary: the size of the disk you're willing to accept.

Under the Hood: Error Sources and Their Growth Rates

According to published workflow guidance, skipping the calibration log is the pitfall that shows up on audit day.

Compass Drift and Steering Wobble

Heading error is the opening thief. It's not one big mistake but a thousand tiny ones. The compass itself drifts—magnetic variation changes slowly, deviation shifts when you transition a winch handle too close, and the fluxgate compass on a steel hull develops its own personality over four hours. Worse: your helmsman. Even an autopilot hunting in a swell introduces a 2–3° oscillation. On a 4-knot boat, one degree of heading error shoves you 0.07 nautical miles off track every hour. Multiply by four—nearly a third of a mile. Now add steering instability: if your pilot swings 5° port to starboard every 30 seconds, that's a sine wave eating your straight row. The catch is we treat heading as a one-off number, but it's a band of probability. Most groups skip this: they log the compass reading and call it done. That band widens with fatigue, cross-seas, and that third cup of coffee at 0300.

Log and Speed Errors

Speed error grows faster than heading error over time—because it multiplies distance directly. A knot of over-reading on your paddlewheel log? You think you've covered 4 miles when you've done 3. The probability circle stretches asymmetrically. I have seen a log fouled by weed report 5.2 knots for two hours straight; the actual speed was 3.8. That's 2.8 miles of phantom progress. You'll chase an intersection that never existed. The trade-off is real: a Doppler log is accurate to ±0.1 knots in calm water, but in a following sea it can read 20% low. GPS speed-over-ground you can trust—but you have no GPS fix, remember? That's the cruel irony. You have precise speed data from four hours ago and a log that's been slowly accumulating barnacles since. Not yet a disaster. But after 4 hours at 6 knots, a 0.5-knot error alone opens a 2-mile radius of uncertainty. That hurts.

Current and Leeway Unknowns

The silent killer. Heading and speed you can measure with instruments; current you guess. Tidal streams flip direction on a 6-hour cycle—four hours means you've crossed at least one ebb-to-flood boundary. Did it push you 1 mile south before turning? Or 0.5 miles north after? You don't know. Leeway—the sideways slip from wind on your hull and rig—adds another vector. A 30-knot breeze on a beam reach can shove a 40-foot sloop 8–10° off your heading row. That's not a wobble; that's a persistent offset. Over four hours at 6 knots, leeway alone can drift you 0.8 miles downwind. Combine it with an unknown 0.5-knot cross-current, and your probability circle isn't round anymore—it's an ellipse tilted 20° leeward. The limit: you cannot measure these with onboard sensors. You infer them from tide tables, your last fix, and feel. flawed batch and the circle becomes a blob. One anecdote: a delivery skipper I know ran a DR track 3 miles east of the rhumb chain for five hours because he forgot to apply the flood set from the last charted arrow. He found the buoy by luck, not skill. Don't bet on luck.

'The sea does not care about your logbook. It moves while you sleep.'

— old watch captain, after we overshot the channel by two miles

So what breaks primary? Heading and speed errors accumulate linearly—predictable, accountable. Current and leeway inject nonlinear chaos. The probability circle after 4 hours is not a neat 2-mile radius; it's a stretched, skewed, slightly drunk oval. You fix this not by polishing your compass swing but by acknowledging which error source dominates right now. In a calm anchorage before departure? Leeway is negligible—focus on the log calibration. In a 25-knot beam reach with a spring tide? Current and leeway swallow your heading precision whole. That dictates how you rebuild the fix—not with math alone, but with judgment about which input has rotted fastest.

A mentor explained however confident beginners feel, the pitfall is skipping the failure rehearsal; says the quiet part out loud — most rework traces back to one undocumented assumption that looked obvious on day one.

Walkthrough: Fixing the Fix on a 4-Hour Passage

Starting data: last GPS fix and log

You're four hours past the last known position. The GPS went dark at 0900, and the log shows 24.8 nautical miles run since then, course 045° magnetic. That fix is a dead point—it tells you where you were, not where you are. I've watched crews stare at that old fix like it still holds magic. It doesn't.

The log is your first ally, but only if you recorded it correctly. Did you reset the trip meter at the fix? If not, you're guessing at distance. Assume you did—24.8 miles is solid data, but it's raw. No current, no leeway, no steering wobble factored in yet.

According to practitioners we interviewed, the trade-off is rarely about talent — it is about handoffs. The pitfall shows up when someone else repeats your shortcut without the same context.

Applying DR: course and speed

Plot the DR position straight: from the 0900 fix, draw a row 045° for 24.8 miles. That lands you at a point—call it DR 1300. off order to stop here. Most people circle that dot and say "we're close." The catch is you've ignored everything that pushed you sideways. Speed through water is not speed over ground.

A 1-knot crosscurrent over four hours shoves you 4 miles off. That hurts. So your DR position is a guess, not a fix. Write it down, but label it "DR only" in the log. The real work starts when you draw the error circle around it.

'The DR position is where you would be if the ocean were a parking lot. It isn't. The error circle is where you actually are.'

— old skipper's note, scratched into a chart table

Plotting the error ellipse

Here's where the math bites. Your uncertainty grows in two directions: along the course row (speed error) and perpendicular (heading error and crosscurrent). Assume your compass uncertainty is ±3° and your log accuracy is ±5%. Over 24.8 miles, along-track error is about 1.2 miles. Cross-track? 3° of arc over 24.8 miles is roughly 1.3 miles.

That's a circle with radius ~1.7 miles—not a problem. But add a 1-knot unknown current, and cross-track error jumps to 4 miles plus the 1.3. Now you've got an ellipse: 1.2 miles long, 5.3 miles wide. That's not a circle. That's a blob with a skinny waist. Plot it on the chart with a soft pencil—make the long axis perpendicular to your course. I've seen people draw perfect circles and miss the whole point: the dangerous direction is to the side, not ahead.

What to do with the result

The ellipse shows your probability zone. The center is the DR position, but you might be anywhere inside the blob. What breaks first is the assumption that you're on course. Stop the boat. Look around. If you see a buoy, a headland, a depth contour that matches—sail toward it. If not, adjust your plan. Set a new course to a point that clears the ellipse's edge, not its center. That sounds conservative until you're threading a reef at night. The trade-off: smaller circles mean more confidence but less caution. A 2-mile radius feels safe until the bottom rises 1.5 miles off your port bow. This process won't save you from a faulty log reading or a misread compass—it's just the best guess from bad data. What you do with that guess is seamanship, not math. Now go fix the fix.

Edge Cases: When the Circle Becomes a Blob

A community mentor says however confident you feel, rehearse the failure case once before you ship the change.

Crossing a Current Boundary

The simple probability circle assumes you're drifting uniformly in one soup. That's fine until your boat slides from a 0.5-knot flood into a 2-knot ebb along a tide chain. I have seen this wreck a four-hour DR in under forty minutes. The circle doesn't stretch—it shears. The fix you plotted at hour two still looked plausible until the actual set bent your track into a banana. The fix: split your passage into separate DR legs at the boundary. Treat each side of the current row as its own error budget. Otherwise your 'circle' is really a skewed ellipse with no center worth trusting.

'A tide row is not a suggestion. It is a wall. Cross it without updating your DR and you are navigating by hope.'

— skipper during a failed delivery, as he pulled out a paper chart

Nighttime Without Landmarks

Darkness strips away your visual fixes entirely. The probability circle inflates silently. But here's the trap: your compass and log keep feeding you data, so the DR plot looks clean. That's a lie. Without at least a depth contour or a lone light to pinch your error, the circle doesn't grow—it bleeds. Most teams skip this: they keep steering a perfect course while the real position wanders. The partial fix of a depth sounder reading can halve your uncertainty radius in one ping. One depth line, one radar bearing to a distant headland—grab anything. Sitting in clean DR data while the blob expands is the fastest way to lose the game.

Multiple Course Changes

Each turn adds an independent error vector. Two course changes in four hours? Your probability circle just doubled its radius—not linearly, but nearly in proportion to the sum of the turn angles. flawed order. You'd think small adjustments are harmless, but they compound helm error, compass lag, and timing uncertainty. The fix: log each turn as a separate mini-fix event. Recompute your error circle from that point, not from the original fix. Most skippers treat course changes as trivial—they aren't. Every wheel spin adds a kink that the simple model cannot smooth.

Partial Fixes (Depth, Bearing, one-off Line)

A partial fix feels like progress. It's not. A single depth contour gives you a line of position—now your probability circle is squashed into a long oval along that line. A single bearing does the same. You haven't fixed your position; you've just shrunk one axis of uncertainty. The catch is that people stop updating the DR after a partial fix, assuming they 'know where they are.' You don't. You know a corridor. The probability blob remains wide on the unmeasured axis. What fixes this? Treat any partial fix as a constraint, not a solution. Keep your original DR alive as the primary track until you cross a second line. That hurts—especially when you're tired and four hours overdue—but it's the only move that prevents the blob from swallowing your arrival.

The Limits: What DR Cannot Fix

Unknown Currents Persist

Dead reckoning cannot see what it cannot feel. You can log every course change, factor in leeway, and interpolate tidal streams from a dog-eared almanac — but an unrecorded eddy or a slant of current that didn't exist on the last chart will shove you sideways without leaving a footprint. That's the real treachery of an old fix: the longer you steer by it, the more likely some invisible force has already bent your track. I've watched a 0.1-knot discrepancy, sustained over four hours, land a boat half a mile off the intended line. Not catastrophic — until the water shallows faster than expected. The danger isn't that DR is wrong; it's that DR can't tell you it is wrong. The only signal you get is the absence of a signal.

'DR gives you a direction to look, not a place to stand. When the circle grows wider than your depth sounder's margin, start treating the chart as a suggestion.'

— old coastwise hand, watching a radar display flicker over a 2-hour-old fix

No Substitute for a Fresh Fix

You've cleaned the compass, tweaked the log calibration, rechecked the GPS offset. None of that matters if the ocean has moved under you. The hard truth: dead reckoning can't self-validate. It extrapolates, always. Every mile you cover without a reset point compounds the previous error — not linearly, but with the quiet acceleration of a fuse. Most crews skip this: they assume that because they've plotted carefully, the dot on the chart still means something. It doesn't. After four hours without a fresh fix, that dot is a memory. The only cure is a new one, and sometimes that means accepting the risk of getting one — heaving to in a lumpy sea to let a sextant sight clear, or punching the waypoint button on a handheld GPS that's been living in the drybag for three weeks. It's uncomfortable. So is hitting charted rocks.

When to Stop DR and Take a Risk

There comes a point where continuing dead reckoning becomes an exercise in faith, not navigation. The circle is a blob. The depths don't match. You've been guessing at leeway for two hours. That's the moment to change strategy — stop the forward motion, heave-to, and run a deliberate search pattern. It feels backward. You were making miles; now you're making circles. But losing ground is better than losing the boat. I once spent forty minutes doing a back-and-forth box search off an unlit coast because the DR position disagreed with the echo sounder by five fathoms. We found the channel by finding the bottom first. That cost little. Running aground would have cost everything. The limits of DR aren't a failure of technique — they're a signal to switch from calculation to caution. Stop. Recalculate. Fix. Only then move. That's not giving up; it's giving yourself a chance to start fresh.

A community mentor says however confident you feel, rehearse the failure case once before you ship the change.