Home / Coordinates and Location Formats

How to Find Coordinates When the Midpoint Is Given: Work Backwards to the Endpoint

Precision drafting tools including compass and set square on a sketchpad

Photo: Liezl Wilken / Pexels

The forward midpoint problem is a warm-up: average the x's, average the y's, done. The backwards problem looks like the same exercise and quietly is not. You are handed one endpoint and the midpoint, and the second endpoint, the thing the average was taken with, has to be recovered from the average itself. Students who try to run the forward routine a second time, averaging the endpoint and the midpoint, get a point halfway to the answer and lose the mark. The backwards problem has its own one-line method, and it is genuinely one line: double the midpoint, subtract the known endpoint. Better still, the method carries its own proof, because the geometry underneath it is a reflection: the midpoint sits exactly halfway, so the missing endpoint is as far past the midpoint as the known endpoint is before it. This guide builds that picture, runs the arithmetic on clean numbers and on the negative-coordinate cases where marks actually die, extends it to 3D, and ends with the check that makes a wrong answer impossible to hand in.

Interactive tool

Work Backwards From the Middle: The Checklist

Check off each step as you finish it. Your progress stays on this page.

0 of 7 complete

0% done

The Formula, Read in Reverse

The forward formula says the midpoint is the average: mx = (x1 + x2) / 2 and my = (y1 + y2) / 2. Solve the first equation for the unknown x2 and the backwards method falls out: multiply both sides by 2 to get 2mx = x1 + x2, then subtract the known endpoint: x2 = 2mx - x1. The same for y: y2 = 2my - y1. That is the whole technique. Double the midpoint coordinate, subtract the known endpoint coordinate, repeat for the other axis. The picture behind it is worth holding onto: on a number line, if the known endpoint is 3 units left of the midpoint, the missing endpoint is 3 units right of it. The formula is that sentence in algebra, doubling places you at the mirror point and subtracting removes the known endpoint's contribution. When a coordinate is negative, nothing about the method changes; the arithmetic just needs its signs respected, which is why the worked examples below deliberately include the negative cases instead of only the friendly ones.

Worked Example 1: The Friendly Case

One endpoint is A = (2, 4) and the midpoint is M = (5, 7). Find the other endpoint B. Apply the formula coordinate by coordinate. For x: x2 = 2mx - x1 = 2(5) - 2 = 10 - 2 = 8. For y: y2 = 2my - y1 = 2(7) - 4 = 14 - 4 = 10. So B = (8, 10). Now the check, which is not optional: run the forward formula on A and B and demand the given midpoint back. Midpoint x: (2 + 8) / 2 = 5. Midpoint y: (4 + 10) / 2 = 7. The midpoint is (5, 7), exactly as given. Notice the reflection picture confirming itself: from A at (2, 4) to M at (5, 7) is a step of (+3, +3), and from M to B is the same step again, (+3, +3), landing on (8, 10). When the numbers are kind, the step view and the formula agree at a glance, and either can be used to audit the other.

Worked Example 2: Negative Coordinates

One endpoint is A = (-3, 1) and the midpoint is M = (4, -2). Find B. For x: x2 = 2(4) - (-3) = 8 + 3 = 11. Watch the double negative do its work: subtracting a negative endpoint adds, and this is precisely the step where hurried work writes 8 - 3 = 5 and loses everything downstream. For y: y2 = 2(-2) - 1 = -4 - 1 = -5. So B = (11, -5). Check by re-averaging: midpoint x is (-3 + 11) / 2 = 8 / 2 = 4. Midpoint y is (1 + (-5)) / 2 = -4 / 2 = -2. The midpoint (4, -2) comes back exactly. The reflection view agrees too: A to M is a step of (+7, -3), and M to B is (+7, -3) again. Two habits protect you in this terrain. Write the subtraction with the negative number in parentheses, exactly as above, so the sign operation is visible. And always re-average; the check is four operations and it converts a guess into a certainty.

Architect sketching on blueprints with ruler at desk, designing a building layout
Architect sketching on blueprints with ruler at desk, designing a building layout

The 3D Version

Nothing about the method is two-dimensional. With a third coordinate, the same reversal applies axis by axis: z2 = 2mz - z1, alongside x2 = 2mx - x1 and y2 = 2my - y1. Suppose one endpoint of a segment in space is A = (1, 0, 5) and the midpoint is M = (3, 2, 1). Then x2 = 2(3) - 1 = 5, y2 = 2(2) - 0 = 4, and z2 = 2(1) - 5 = -3, giving B = (5, 4, -3). Re-average all three: ((1+5)/2, (0+4)/2, (5-3)/2) = (3, 2, 1), the given midpoint. Surveyors, game developers, and anyone placing a point symmetric to another across a center in 3D uses this constantly, because "the point opposite A through center M" and "the endpoint whose midpoint with A is M" are the same request wearing two coats.

When the Problem Adds a Distance

Some problems dress the backwards midpoint in a second question: find the endpoint, then the length of the segment, or the distance from the midpoint to an endpoint. Take A = (1, 3) with midpoint M = (3, 5). The endpoint: x2 = 2(3) - 1 = 5 and y2 = 2(5) - 3 = 7, so B = (5, 7). The full segment length AB is the distance between (1, 3) and (5, 7): sqrt((5-1)^2 + (7-3)^2) = sqrt(16 + 16) = sqrt(32), about 5.66. And the distance from M to either endpoint is exactly half of that, sqrt(8), about 2.83, because the midpoint is halfway by definition. That halving is itself a check with real power: if a problem gives you the midpoint-to-endpoint distance, doubling it must reproduce the segment length your coordinates imply, and if a problem gives the full length first, the endpoint you find must sit exactly half of it from M. Our guide to finding coordinates between two points covers the forward family this inverts, the midpoint as the special case of the section formula and the geographic midpoint of two places, and the backwards skill here is its mirror image: where that guide divides a segment at a point, this one extends a segment through a point.

Three Drills, With Answers to Check Against

Skill on this problem is repetitions with feedback, so here are three drills in ascending trickiness. Cover the answers, work each one with the formula, then check by re-averaging before you peek. Drill one: known endpoint (6, -2), midpoint (3, 1). Doubling and subtracting: x2 = 2(3) - 6 = 0, y2 = 2(1) - (-2) = 4, so the endpoint is (0, 4); re-averaging gives ((6+0)/2, (-2+4)/2) = (3, 1), correct. Drill two: known endpoint (-5, -5), midpoint (-1, 2). Then x2 = 2(-1) - (-5) = 3 and y2 = 2(2) - (-5) = 9, giving (3, 9); re-averaging returns (-1, 2), correct, and notice both subtractions added, because both known coordinates were negative. Drill three, in 3D: known endpoint (2, 7, -4), midpoint (0, 3, -1). Then x2 = 2(0) - 2 = -2, y2 = 2(3) - 7 = -1, z2 = 2(-1) - (-4) = 2, giving (-2, -1, 2); re-averaging all three coordinates returns (0, 3, -1). If a drill fights back, the re-average tells you which coordinate carries the error, and nine times in ten it is a negative that lost its sign on the way through the subtraction. That diagnostic loop, attempt, re-average, locate, repair, is the whole study method for this topic compressed into four verbs.

The Errors That Cost the Marks

Three errors account for nearly every lost mark on this problem type. The first is averaging again: taking the midpoint of the known endpoint and the midpoint, which lands halfway to the true endpoint. It feels like progress because it moves in the right direction, and it is wrong by exactly half the remaining distance. The cure is the picture: the midpoint is the middle, so the answer must be as far beyond it as the known point is behind it, never between the two given points. The second error is the sign slip in the subtraction, 2mx - x1 computed with x1 negative as if it were positive; parentheses around negative numbers and the re-averaging check together make this error nearly impossible to submit. The third is mixing the points: subtracting the midpoint from twice the endpoint, the formula inverted, which produces a point on the wrong side entirely. Label every number with the point it belongs to, A, M, or B, before substituting, and the formula has no way to be assembled backwards.

Related Guides

Readers working through this topic usually need its closest companions too. These coordinates guides cover the neighboring ground:

Frequently Asked Questions

What is the formula to find an endpoint when the midpoint is given?

Run the midpoint formula in reverse: x2 = 2mx - x1 and y2 = 2my - y1, where (x1, y1) is the known endpoint and (mx, my) is the midpoint. Double each midpoint coordinate and subtract the corresponding known endpoint coordinate. In 3D, add z2 = 2mz - z1.

Why does doubling the midpoint and subtracting work?

Because the midpoint is the average of the endpoints. If mx = (x1 + x2) / 2, then 2mx = x1 + x2, so x2 = 2mx - x1. Geometrically, the midpoint is a mirror: the unknown endpoint lies as far past the midpoint as the known endpoint lies before it, and the formula computes exactly that reflection.

How do I check my endpoint answer?

Re-average. Take the midpoint of your new endpoint and the known endpoint using the ordinary forward formula. It must return the given midpoint exactly, coordinate for coordinate. If it does not, the error is in your subtraction, most often a mishandled negative sign.

Can this work with negative coordinates?

Yes, and the method does not change at all. The one place to be careful is subtracting a negative known coordinate: 2(4) - (-3) is 8 + 3, not 8 - 3. Writing negative numbers in parentheses during substitution, and re-averaging to check, keeps negative cases as safe as positive ones.

Is the missing endpoint always on the far side of the midpoint?

Yes. The midpoint is by definition halfway between the endpoints, so the unknown endpoint is the reflection of the known endpoint through the midpoint: same distance, opposite direction. If your answer falls between the known endpoint and the midpoint, you have averaged again instead of reflecting, which is the classic error on this problem.

Photo credits: Liezl Wilken (Pexels); Tima Miroshnichenko (Pexels). Photos are served from this site; follow the credit to the library page for the original.