XB/BASIC Seeds

[Opry99er]

Procedurally generated game maps done in BASIC vs. XB using RANDOMIZE with a seed do not yield results even remotely similar. Below are two screen shots... The first is a game map generated in BASIC, the second is the exact same code, generating the same seeded map, but in XB.

 

This would change the game completely!

 

I tried multiple seeds with the same code in both BASIC and XB, and it is an entirely different game. I painstakingly manipulated the code in BASIC to have a very specific, reproducible pattern. I have been coding this entire game on the real console. It wasn't until I pulled it over onto my PC and tried it on Classic99 that I loaded the XB cartridge to check the SIZE of my code. I thought, "what the hell, I'll go ahead and run the map generator in XB to see if it runs any faster". I was then greeted with quite a surprise.

 

I know the topic of the BASIC vs. XB RND functionality has been discussed ad nauseum, but it wasn't until I started really counting on seeds that I encountered the true difference for myself. It is really quite fascinating.

 

Anyway, I just thought I'd share.

 

 

 

[+Lee Stewart]

Most PRNGs are designed to reproduce the same pattern every time they start with the same seed so that code using them can be adequately debugged. This is the case with both TI Basic and XB. As you noted, however, they generate different patterns with the same seed. The obvious conclusion is that they are using different PRNGs, which is, indeed, the case. TIB’s PRNG calls GPL RAND at least once for each of the 7 radix-100 digits of the floating point number that gets assembled. Whereas, XB uses a completely different algorithm that Rich posted in at least one thread you so eloquently characterized as discussing the topic ad nauseam.

 

...lee

[Opry99er]

Well, I meant nothing derogatory with that term... I, myself, have participated in those threads and even started one or two.

 

Just interesting seeing an entirely different world created by the differences.

 

Almost like using different starting tiles in Conways Game of Life...

[senior_falcon]

Try RXB and see if the pattern is the same as TI BASIC. I understand that RXB uses the random number generator from TI BASIC.

[Opry99er]

I will try that out when I get back home. Currently traveling to Tennessee to pick up my folks... dads health is declining and I am moving them up to Wisconsin permanently.

 

I have some test code on my phone if anyone cares to give RXB a try.

[+Lee Stewart]

Well, I meant nothing derogatory with that term...

 

I did not mean to imply that you did. Perhaps I should have said “aptly” instead. In this context, “eloquently” is too easily taken as sarcasm. I do agree that the topic was pretty much pounded into the ground. Or, as @Willsy might say, “that horse is well and truly dead!”

 

...lee

[+OLD CS1]

Whatever, Lee. You can refrain from sugar coating. I will say what I know we are all thinking, as I alone can speak with full authority for all users here at AtariAge.

 

We all know Rich posts extended, long-winded, no spoiler-tagged, overly in-depth technical IEEE-like articles about anything to do with GPL and Extended BASIC.

 

And we geeks eat it up. Thanks, Rich!

[RXB]

Ok thanks all.

I only delved into GPL RND of TI Basic and XB as when I posted the insanely complicated XB RND as being slower for a reason.

 

Which is also why I dumped XB RND in favor of GPL TI Basic RND in RXB 2015