Squaring a number is easy to do forward and surprisingly hard to undo — multiplying 7 by itself takes seconds, but starting from 49 and working backward to "7" has no shortcut arithmetic operation the way addition has subtraction. That backward direction is what a square root asks for: which value, squared, lands on the number you started with? For perfect squares — 1, 4, 9, 16, 25, 36... — someone has effectively already done the backward work for you, and recognizing the pattern is enough. For everything else, there's no clean backward path at all, only a process of educated guessing and checking that closes in on the answer.
**A quick note before starting: this method is an estimation technique, not a shortcut to an exact answer** — for non-perfect squares, no amount of refinement by hand produces the *exact* irrational value, only an increasingly close approximation, which is completely sufficient for virtually every practical use.
**Step 1: bracket the number between two perfect squares.** Find the two perfect squares your target number falls between. For √50: 49 (7²) and 64 (8²) bracket it, so √50 is somewhere between 7 and 8.
**Step 2: estimate the position within that bracket.** 50 is much closer to 49 than to 64 (50 − 49 = 1, versus 64 − 50 = 14), so √50 should land much closer to 7 than to 8 — a first estimate might be around 7.1.
**Step 3: refine by squaring your guess and adjusting.** Square your estimate and compare to the target. 7.1² = 50.41 — slightly too high, so try a bit lower. 7.07² = 49.98 — very close. 7.071² = 50.0004 — extremely close. √50 ≈ 7.071.
**Worked example: estimate √85.** Bracket: 81 (9²) and 100 (10²), so √85 is between 9 and 10, closer to 9 (85−81=4, versus 100−85=15). First guess: about 9.2. Check: 9.2² = 84.64, slightly low. Try 9.22² = 85.01, very close. √85 ≈ 9.22.
**Worked example: estimate √200.** Bracket: 196 (14²) and 225 (15²), so between 14 and 15, much closer to 14 (200−196=4, versus 225−200=25). First guess: about 14.1. Check: 14.1² = 198.81, a bit low. Try 14.14² = 199.94, very close. √200 ≈ 14.14.
**A more systematic refinement method (the "divide and average" or Babylonian method), for higher precision.** Start with any reasonable guess g. Compute a better guess as (g + n/g) ÷ 2, and repeat with the new guess until it stops changing meaningfully. For √50, starting with g=7: next guess = (7 + 50/7) ÷ 2 = (7 + 7.143) ÷ 2 = 7.071. Checking 7.071² ≈ 50.0004 — already extremely accurate after just one iteration. This method converges very fast and is genuinely usable by hand with a bit of practice at the division step.
**Perfect squares worth memorizing for faster bracketing.** Knowing the perfect squares from 1² through 20² (1, 4, 9, 16, 25, 36, 49, 64, 81, 100, 121, 144, 169, 196, 225, 256, 289, 324, 361, 400) covers bracketing for any number up to 400 instantly, without needing to hunt for the right pair each time.
**Common mistakes.** The most frequent error is assuming every number has a "nice" square root and stopping at a rough bracket instead of refining — reporting √50 as "about 7" loses meaningful precision compared to the more accurate 7.071. A second common mistake is confusing squaring with doubling: squaring a number multiplies it by itself, not by 2, and this confusion often shows up specifically when checking an estimate by squaring it back.
**A note on negative numbers.** No real number squared produces a negative result (a negative times a negative is always positive), so negative numbers have no real square root — this is different from cube roots, which do exist for negative numbers.
**Worked example: estimate √120 using the Babylonian refinement method, for a faster route to high precision.** Bracket: 100 (10²) and 121 (11²), so √120 is between 10 and 11, very close to 11 (120 is only 1 less than 121). First guess: g=11. Refine: (11 + 120/11) ÷ 2 = (11 + 10.909) ÷ 2 = 10.95. Check: 10.95² = 119.9, extremely close already after one refinement step. A second refinement would push it even closer, but one step is usually sufficient for practical purposes.
**Worked example: estimate √30, a number without a nearby perfect square.** Bracket: 25 (5²) and 36 (6²), so between 5 and 6, closer to 5 (30−25=5, versus 36−30=6, roughly the midpoint but very slightly toward 5). First guess: 5.5. Check: 5.5²=30.25, very close but slightly high. Try 5.48²=30.03, closer. √30 ≈ 5.48.
**Why estimation skill matters even with calculators everywhere.** Being able to quickly bracket a square root between two whole numbers is a fast sanity check against a mistyped calculator entry — if you compute √50 on a calculator and get something outside the range 7-8, you immediately know something was entered wrong, without needing to redo the full calculation.
**Worked example: estimate √150, one more full bracket-and-refine pass for practice.** Bracket: 144 (12²) and 169 (13²), so between 12 and 13, closer to 12 (150−144=6, versus 169−150=19). First guess: 12.2. Check: 12.2²=148.84, a bit low. Try 12.25²=150.06, very close. √150 ≈ 12.25.
**Worked example: apply the Babylonian method to double-check the √300 estimate above with more precision.** Starting guess g=17.32 (from the bracket-and-refine result). Refine: (17.32 + 300/17.32) ÷ 2 = (17.32 + 17.32) ÷ 2 — since 300/17.32 ≈ 17.32 already, the refined guess barely changes, confirming 17.32 was already an excellent estimate before this extra step.
**Worked example: estimate √300, a number requiring bracketing further up the perfect-square list.** Bracket: 289 (17²) and 324 (18²), so between 17 and 18, closer to 17 (300−289=11, versus 324−300=24). First guess: 17.3. Check: 17.3²=299.29, close but slightly low. Try 17.32²≈299.99, essentially exact. √300 ≈ 17.32.
**Perfect squares worth having memorized, extended a bit further for larger everyday numbers.** Beyond 1² through 20² (covered above), knowing 25²=625 and 30²=900 extends quick bracketing up to 900 without needing the full list in between — useful since square roots of three-digit numbers come up often enough in practical contexts (room areas, garden plots) to be worth this small extra investment of memory.
Once you're comfortable with square roots, cube roots use a related but distinct estimation approach, covered in How to Find a Cube Root, and the underlying exponent relationship between roots and fractional powers is covered in Understanding Exponents and Powers.