|
|
Tuesday, February 11, 2003 - 01:04 PM
 Ever wonder how to determine the square root of a number without the aid of a calculator? Believe it or not, people used to do this. Here's one method for doing so.
If you're good with long division, here's a quick way to find pretty accurate square roots without the aid of a calculator. Let's try 24.6.
- Make a guess. It can be a very bad guess. It doesn't matter. You can even guess 1. Let's try 5 since 52 is 25, which is pretty close to 24.6.
- Divide 24.6 by 5. 24.6 / 5 = 4.92.
- Now, comes the trick: Pick a new guess between 5 and 4.92 and divide it into 24.6 again. Let's try 4.95. 24.6 / 4.95 = 4.96. 4.96 is pretty close to 4.9598 which is the actual square root of 24.6.
- Repeat steps 2 and 3 to any desired level of accuracy. The further you go, the harder the long division becomes. But the first few cycles yield a pretty close answer.
The reason this works is because n*n = 24.6 and n = 24.6 / n. Therefore, the real square root will always be somewhere between 24.6 / n and n.
Click here to see another method described by Dr. Math at the Math Forum.
If you like CuriousMath.com, here's a book you'll love:
Learn more about Concepts of Modern Mathematics
Rating
|
Other Stories
- A fast way to sum a series of numbers
(Nov 11, 2009)
- Summing Cubes to Produce a Cube
(Oct 26, 2009)
- unusual whole number sequence
(Oct 04, 2009)
- Numbers Having Two Positive Cubes
(Aug 31, 2009)
- Speculation about Prime Number Space
(Jul 30, 2009)
- In Pursuit of Pi
(Mar 14, 2008)
- Simple Summation Formula
(Feb 27, 2008)
- 0.999...=1 ?
(Nov 23, 2007)
- Large Special Products Factorization Algorithm
(Jun 24, 2007)
- the golden number
(May 22, 2007)
|
Comments
Very detailed instructions and explanation for calculating square roots by hand.
For those of you who are over 50 years old, you probably remember the paper method of calculating square roots as described by Dr. Math and Yet another method referenced by Clay.
To find square roots without a scientific calculator, Clay's method works very well. He'll just have to repeat the same step a few times to get a better answer. Like he said, eventually, it will be right. This is the method by iterations, as higher mathematics call it. In fact, there is a method devised by Newton, so we call it Newton's method of calculating square roots.
It is almost identical to Clay's method, except that to get a better answer, we do not simply divide the original number N by the estimate, but we add the residual divided by twice the estimate, E. If it all sounds complicated, an example is in order.
Let's try to calculate the square root of 24.6. As Clay did, the initial estimate is 5. So N=24.6, and
E=5.
The Newton's method says a better estimate, E1 is
E1=E+(N-E*E)/(2E)
=5-(24.6-5*5)/(2*5)
=5-0.4/10
=4.96 (error = 0.0033%)
This can easily be handled using mental arithmetic, and is pretty close to the real value of 4.9598.
Continuing to use mental arithmetic (see Clay's article on multiplication tricks), we can mentally calculate
4.96*4.96=5*5-2*0.04*5+0.0016=24.6016
This time the residue is 0.0016, divided by 2*4.96=9.92 gives the next estimate,
E2=E1-(N-E1*E1)/(2*E1)
=4.96-.0016/9.92¸
Now, we can look at how we can handle the division. The approximation
0.0016/10=0.00016 is a very close. In fact, it gives E2=4.95984, the error is 0.0000261%!
However, if you want to push your mental arithmetic, the division mentally is possible! You notice that dividing by 9.92 instead of 10 gives 0.8% more to the result, which we can simply add to 0.00016+0.00000128=0.00016128
Therefore
E2=E1-0.00016128=4.95983872, compared to the actual value of 4.959838707 (error of 0.0000003%). I think we fare better than calculators displaying up to 8 digits, using just our brains, not even pencil and paper!
By the way, Clay, this may answer one of your reader's question about a colleague being able to do square roots mentally, and amazingly accurately. Now you can too! Feel free to write me for more explanations if I didn't make it clear above.
To be consistent with the calculations, the formulas should read
E1=E-(E*E-N)/(2E), and
E2=E1-(E1*E1-N)/(2E1).
Notice that I have switched around the E*E and N to change the sign to a minus.
Hope you enjoy this new Square Root Power!
E1=1/2 (E + N/E)
Using your example
E1 = 1/2 (5 + 24.6/5)
E1 = 1/2 (5 + 4.92)
E1 = 1/2 (9.92)
E1 = 4.96
square root of 24.6 is 4.9598 --- pretty close
The math WORKS in this formula!
=5-(24.6-5*5)/(2*5)
=5-0.4/10
=4.96 (error = 0.0033%)
Check your math...
5 + (24.6 - 5*5) / (2*5)
5 + (- 0.4) /10
4.6/10
0.46 (not even close!)
E1=1/2 (E + N/E)
I have a quick and dirty version, it's really just a refinement of Clay's version.
At step 3, after you have 4.92, instead of picking at random a number between 4.92 and 5, split the difference in half and add it to 4.92. 5 - 4.92 = .08, .08 / 2 = .04, 4.92 + .04 = 4.96
Now repeat steps 2 and 3 with your new guess, and it almost nails it. Plus, using splitting the difference as a guide and you can keep boosting your accuracy the most efficient way possible. On paper it's more work, but it's easy as pie when you just do it in your head.
Now I just need to find a trick like the X multiplication trick to make dividing these numbers easier. :P
This site rocks.
it's (-0.4)/10 first, which is -0.04 then add 5
so:
5 + (24.6 - 5*5)/(2*5) = 5 + (-0.4)/10 = 5 - 0.04 = 4.96, which is very close
Only logged in users are allowed to comment. register/log in