WORKING METHODS FOR COMPUTERS

(c) yd Approximation

471

Multipliers.
Products.
Total of
Squares.

I-OO
•7216
•5800
-4900
•3000
-3800
•4400

75
'435°
-4763
•3450
-0750
•0975
•'575

•63
•3087
-2898
•2390
•0567
•0756
•1575

•54
•l620
*°54°
•0486
•2157
•2700
•1512

•6cS
.2584
•0884
•0816
-3400
•4261
•2448

-65
•2860
•1365
•1625
•1820
-2340
•2551

Divide by 2-1717)2-1717
1-6250
1-3667
1-1694
1-4832
1-4061

I -0000
7483
•6293
•5385
•6830
•6475

Squares

Multiply-J->7

V 2*1717

Sat, Coeflfs.
Squares .

1*0000 *5599 -3960 ^2900 -4665 -4193

-5651 -4229 -3556 -3043 -3859 -3659

•8328 -6232 -5240 -4484 -5687 '5392

-6935 '3884 -2746 -201i ^3234 -2907

3-1317=1-76972

2-1717=1-47372

Verification (Saturation Coefficients as Weights).

Multipliers,
Products.
Total.

•833
•60 1 1
-4831
•4082
•2499
-3165
•3665

•623
•3613
•3957
-2866
•0623
•O8l0
•1308

•524
•2567
•2410
•1988
•0472
•0629
•1310

-448
'*344
•0448
-0403
•1789
*2240
•1254

•5&)
'2162
•0740
-0683
-2845
•3565
•2048

*539
•2371
•1132
•1347
-1509
-1940
•2115

Totals 3-536
i -8068
1-3518
1-1369
'9737
1-2349
I-I700
7-6741

Sat. Coeffs,* -8325
•6228
•5238
•4486
-5690
•5391

Squares , »693°
-3879
•2743
-2012
*3237
-2906
2-1707

-. ~. .

* Divisor ^

roughly estimate the changes they in turn will produce. Thus, with
the second column, the first computation gave 73, the second
735 ; the increase will give a large weight to a large figure; hence
we guess that the third or fourth computation will give -75. In
the third column this increased weighting will probably raise -621