[VA] SRC #012a - Then and Now: Probability

[Albert Chan]

(10-16-2022 02:10 PM)J-F Garnier Wrote:  HP-75 result is: 9.51234350244E-6
...
The Saturn (from the HP-71) introduces a very small improvement, known as the "round-to-even"

My code of scaling by P by 6^S take account of round-to-even rule.
For machine that round-away-from-zero, I would scale by gcd(2,4,6) = 12 instead.
Or, scale by 1.2 for decimal machine, to keep denominator (1.2^S) from overflow, even with big S

< 40 FOR I=1 TO R @ FOR J=1 TO C @ W(I,J)=3/(3-(J=1)-(I=J)-(I=R)) @ NEXT J @ NEXT I
> 40 FOR I=1 TO R @ FOR J=1 TO C @ W(I,J)=0.6/(3-(J=1)-(I=J)-(I=R)) @ NEXT J @ NEXT I

< 180 DISP TIME-T0;R;S;T/6^S
> 180 DISP TIME-T0;R;S;T/1.2^S

With above changes, on emu71, I get: (don't worry about error of few ULP's)

>RUN
[VA]SRC012A R,S= 30,60
...
70.6      30 60      9.51234350204E-6