Dicebot question

[pward]

One question/suggestion I thought up from the latest dicebot rant in one of the other folders:

Is each die a separate instance of a random number generator, or is the same object used for 2 sequential die rolls?

If it's the same random generator, can an enhancement/modification be made to make this work that way, with the first die also being used for single drs?

From a programming standpoint, I know that this is possible, just not how reasonable a change to the VASL/VASSAL engine it would be. This should reduce the number of times a string of 6s was rolled on both dice, to produce the unfortunate sequences of bad in-game effects. Separating the cdr from the wdr as separate pseudo-random number generators should reduce the number of sequential repeats of the same result. Provided of course that the initial seed is different for each die object (constant doubles anyone???). Using the cdr object for the single die rolls would also get them further "out of sync" and less of a repeatable series of numbers when comparing between games.

 

[zgrose]

It is one Random used over and over.

Do you have some documentation the benefits of 2 Randoms over 1? I'd be interested in checking it out.

 

[pward]

Sorry no documentation, just the random (pun intended) musings of a computer science major. Maybe there's a masters thesis in here somewhere.

I would think that 2 separate random number generators (pseudo-random anyhow) would behave more like 2 separate dice. Since we have object oriented programming in Java, why not have one object instance for each die? The current model is more like rolling the same die twice and just counting the first throw for things like rate, something most players would look down upon in FtF play. More resources for another instance seems to be the only serious drawback. I'm not conversant on the internal working of random number generation in Java, so an additional instance of the RNG may not really get separate set of results. This may also vary by OS implementation.

The thread in the other forum had a lot of complaints about "abnormal" results like a string of 5 boxcars. With a single random generator this is really a string of nothing but 10 6's in a row. (From 5 boxcars.) Like I said in the other thread, dice don't have memories, nor concerns for the results, and the same should hold true for the random number generator. It just sucks when these are the MC's you need to keep GO MMC's in the victory hexes, or you just broke every gun you had on the board...

 

[andy]

:gas: The Dicebot thread :scream:

Actually this has been debated at serious length. However, I have suggested it as a research paper topic. I'd be happy to help you take it on.

 

[pward]

Thanks Andy, but no. I just finished my BS and I won't be going back anytime too soon for the masters. (School intruded too much on my gaming time).

 

[Jeff Leslie]

I was playing a game the other day and out of 10 dice rolls, 8 of them were 11's (6,5 or 5,6), and the first 6 in a row were 11's. Of the other two rolls out of the ten that weren't 11's, one was a 10 (6,4) and the other was a 9 (5,4).

Did I mention that was with real dice? Dicebot shmicebot.

 

[Doc Ozzie]

But Jeff, you were not playing with real balanced dice that were properly agitated. You might as well have been using the VASL Bot than your heathen non-balanced squares you call dice.

 

[Robin Reeve]

But Jeff, you were not playing with real balanced dice that were properly agitated.

And don't forget sweat, grease, an air draft, air density modulation, undetected sismic activity, unequal surface of the table or dicetower, parkinson or drinking have some effect on dice... :laugh:

 

[countermanCX]

undetected seismic activity - that must be it.

mangles my DRs ev'ry time.