(PC-1211) Multiple Dice Throw

[SlideRule]

An excerpt from Science and Engineering Sourcebook, page 87:

Description
Many games, e.g. the popular YAH-T-ZEE, and even computer modeling simulations, are based on the outcome of the multiple dice throw. This program permits a throw of 1 to 9 dice and displays the results of each multiple throw. The pseudo random generator shown in line 70, which does the "throwing," works well in the applications to which I submitted it; however, if you have your own favorite feel free to replace that line.
Program Listing
  10: "Z"CLEAR :PAUSE "MULTIPLE DICE":USING
  20: INPUT "STARTER= ";B
  30: INPUT "HOW MANY DICE?";A
  40: IF (A>0)*(A<10)*(A=INT A)THEN 60
  50: BEEP 1:PAUSE "ERROR":GOTO 30
  60: FOR I=1TO A
  70: B=997B+π :B=B-INT B:C=INT 6B+1
  80: D=C*10^(I-1)+D:NEXT I
  90: PRINT A;" DICE=";D
100: D=0:GOTO 30

INSTRUCTIONS
Shift Z initializes the program and displays a prompt for a seed to start the pseudo random sequence. For best results, enter a number between 0 and 1. Next comes the prompt HOW MANY DICE for this throw. After a number n between 1 and 9 is entered, the program displays a n-digit number where each digit is between 1 and 6 and corresponds to a die. Pressing ENTER starts the sequence for another dice throw.

BEST!
SlideRule

correction made

[rprosperi]

It appears there is a missing ")" on line 80.

Also, does this flavor of BASIC allow implied multiplication, look at line 70??

[SlideRule]

Thanks for the 'eyeball', my error.

Line 70 listing as published;
[attachment=9257]

BEST!
SlideRule

[Valentin Albillo]

(03-20-2021 01:00 PM)rprosperi Wrote:  Also, does this flavor of BASIC allow implied multiplication, look at line 70??

Yes, it does, i.e. 2ABC -> 2*A*B*C , SIN 3A -> SIN(3*A), etc.

Have a nice weekend.
V.

[C.Ret]

Hello cheer PC-1211 enthusiasts !

Yes, a good flavor of BASIC with a lot of tricks and short-cuts which make this pocket a useful calculator...
Except for PRINT capabilities and, due to it's great age, speed of running (walking ?) programs !

As in most of applications (i.e.; Yahtzee, poker dice, ...), we need to display or print the listing of values of the dices we have just throwed.
Here is my solution for this delicate implementation on the venerable PC-1211 machine:

1:PRINT T;USING "##";A:RETURN
2:PRINT T;USING "##";A;B:RETURN
3:PRINT T;USING "##";A;B;C:RETURN
4:PRINT T;USING "##";A;B;C;D:RETURN
5:PRINT T;USING "##";A;B;C;D;E:RETURN
6:PRINT T;USING "##";A;B;C;D;E;F:RETURN
7:PRINT T;USING "##";A;B;C;D;E;F;G:RETURN
8:PRINT T;USING "##";A;B;C;D;E;F;G;H:RETURN
9:PRINT T;USING "##";A;B;C;D;E;F;G;H;I:RETURN
10:PRINT T;USING "##";A;B;C;D;E;F;G;H;I;J:RETURN

12:"Z"AREAD Z:BEEP (Z<=0)+(Z>=1):END
13:"N"AREAD N:IF (N<1)+(N>10)+(N<>INT N)BEEP 1:END
14:T=0:FOR K=1TO N:Z=(∏+Z)^5,Z=Z-INT Z,A(K)=1+INT 6Z,T=T+A(K):NEXT K:USING "###.":GOSUB N:GOTO 14

This code is conveniently use in DEF mode.
Enter a fractional seed and press [shift]+[ Z ] to initiate randomizing seed. The pocket beeps when seed if out of the ] 0 ; 1 [ range.
Enter the number of dices to throw and press [shit]+[ N ]. The pocket will display the total and all the dice’s values. It beeps when user ask for less than one dice or more than ten dices.
Pressing the [ ENTER ] key re-throw a full set of dices.

P.S.#1: To enter lines from 1 to 10, start with complete line 10 and go down to 1 by erasing extra characters with space, you will spare time and efforts...

P.S.#2: For the laziest of us, but limited to height dices throws:
1:"Z"AREAD Z:BEEP (Z<=0)+(Z>=1:END
2:"N"AREAD N:IF (N<1)+(N>8)+(N<>INT NBEEP 1:END
3:T=0:USING :FOR K=1TO N:Z=(∏+Z)^5,Z=Z-INT Z,T=T+(1+INT 6Z)*(1+10^-K:NEXT K:PRINT T:GOTO 3

[robve]

(03-20-2021 11:48 AM)SlideRule Wrote:    70: B=997B+π :B=B-INT B:C=INT 6B+1

Thank you for sharing this code!

Line 70 uses an interesting pseudo-random number generator. The PRNG is very simple, yet seems to be good enough to reliably simulate multiple dice throws, perhaps. It passes the spectral test when I tested this with a simple program on a PC-1350 (select image to view):

   

10 "A" CLS: WAIT 0: B=0
20 FOR I=1 TO 32*150
30 B=997*B+π,B=B-INT B:X=150*B
40 B=997*B+π,B=B-INT B:Y=32*B
50 PSET (X,Y)
60 NEXT I
70 WAIT: GPRINT: END
Disclaimer: this spectral test code may not meet official test requirements.

By comparison, the PC-1350 RND function spectacularly fails the spectral test, so I would not use it to throw dice for example (select image to view):

   

10 "A" CLS: WAIT 0: FOR I=1 TO 32*150: PSET (RND 150-1,RND 32-1): NEXT I: WAIT: GPRINT: END

The number 997 is used by other PRNGs. See also previous HP Forum threads on this subject:
https://hpmuseum.org/forum/thread-7196.html
https://hpmuseum.org/forum/thread-2041.html

- Rob

[Massimo Gnerucci]

(03-23-2021 02:43 PM)robve Wrote:  

(03-20-2021 11:48 AM)SlideRule Wrote:    70: B=997B+π :B=B-INT B:C=INT 6B+1

Thank you for sharing this code!

Line 70 uses an interesting pseudo-random number generator. The PRNG is very simple, yet seems to be good enough to reliably simulate multiple dice throws, perhaps. It passes the spectral test when I tested this with a simple program on a PC-1350 (select image to view):



10 "A" CLS: WAIT 0: B=0
20 FOR I=1 TO 32*150
30 B=997*B+π,B=B-INT B:X=150*B
40 B=997*B+π,B=B-INT B:Y=32*B
50 PSET (X,Y)
60 NEXT I
70 WAIT: GPRINT: END
Disclaimer: this spectral test code may not meet official test requirements.

By comparison, the PC-1350 RND function spectacularly fails the spectral test, so I would not use it to throw dice for example (select image to view):



10 "A" CLS: WAIT 0: FOR I=1 TO 32*150: PSET (RND 150-1,RND 32-1): NEXT I: WAIT: GPRINT: END

The number 997 is used by other PRNGs. See also previous HP Forum threads on this subject:
https://hpmuseum.org/forum/thread-7196.html
https://hpmuseum.org/forum/thread-2041.html

- Rob

Shouldn't you use RANDOM before invoking RND, unless you want to use the same seed each time?

[robve]

(03-23-2021 03:34 PM)Massimo Gnerucci Wrote:  Shouldn't you use RANDOM before invoking RND, unless you want to use the same seed each time?

RANDOM and RANDOMIZE set the seed of the sequence. Testing a PRNG is independent of its seeding. In other words, when a PRNG is periodic (all LCGs are but periods differ and large periods are preferable) it does not matter where you start. Also, most RANDOM implementations that take no seed value simply use TIME or some other system clock value.

Many mistakes in the past made cryptographic operations insecure because the PRNG was predicable and the seed was simply TIME, which is also predictable see e.g. random number generator attacks such as the successful attack on Netscape's secure HTTPS page browsing. By contrast, modern CPUs offer RDRAND or similar instructions for "true" randomness required by secure cryptographic operations.

EDIT: answer: setting B=0 (the seed) or not using RANDOM in the other code does not help or hinder. Running the programs multiple times with or without seeding produces very similar results.

- Rob

[pyedog]

(03-20-2021 11:48 AM)SlideRule Wrote:    70: B=997B+π :B=B-INT B:C=INT 6B+1

I think this random number generator might come from page 241 of "119 Practical Programs For The TRS-80 Pocket Computer", although the book's author probably got it from somewhere else.

A very similar slightly faster version from https://www.aldweb.com/articles.php?lng=en&pg=51

Code:

R=221R + 0.2113:R=R-INT R

although I'm not sure how well distributed it winds up being.

BSD used to use something like

Code:

R=1103515245+12345 MOD 2^31

which was horrible.

I think Microsoft Windows uses (or used to use) something like

Code:

R=214013R+2531011 MOD 2^31

which seems to be slightly better.

But on a PC-1211 the BSD and Microsoft variations are significantly slower.