Colonels it might not satisfy you but seble has said:

"The new code uses the built-in C# RNG and the old code uses a third party RNG."

So while it isn't the TRUE random numbers like some want it is a change from the third party RNG they are using now.

It'll be interesting to see whether this is a good or bad thing. I'm kind of surprised that the opposite wasn't the case to be honest.

help me understand the statistical stuff - are y'all saying that different random number generators produce results that can be seen to be different? then they aint random right?

its a complete guess on my part but i think both these guys have no clue what they're talking about....

"I think" that the difference from one RNG to the next is infinitesimal. like 0.000001....

now i'm probably the idiot for making an assumption on something i have no clue about, but whatever...

question to the stat guys. . . is there REALLY any difference from one program to the next? in the HD sense, that is.

i'm sure there's a difference when calculating 100 decimal points. but thats not the case in HD.

the way i look at it, this is like the scientists that argue over who can calculate pi to the largest decimal. who cares if its calculated to a million or two million decimal points. its really not that big a deal.

Colonels has wanted true random that is harder to get and more expensive for a program like this to use.

Pseudo-Random is where a computer program picks out a number by use of a math formula. It usually is faster and cheaper to get this randomizer but is not truly random as the computer is picking it through the formula.

Ya as long as the number generator over the course of the long haul can produce what is seen as random I am fine with it. We shall see, really the only way to truly tell how big of an effect the two randomness models would have is to use both of them over the course of a few thousand games at least to get a feel for it.

okay, if that is what this is about it is clear to me that these differences are going to be waaaay too small ever to matter - lost in the noise

Quote: Originally posted by schroedess26 on 1/19/2010Ya as long as the number generator over the course of the long haul can produce what is seen as random I am fine with it. We shall see, really the only way to truly tell how big of an effect the two randomness models would have is to use both of them over the course of a few thousand games at least to get a feel for it.

this has nothing to do with game play. thats has to do with the forumla's that WIS uses once the numbers are generated. it has nothing to do with the RNG.

a RNG works fine. if you were to generate a million digits (1-10) the results would come out ~10% for each digit. plus or minus a couple tenths.

any issue a coach has with gameplay would be with the forumla used to compute the random numbers that were generated. not the RNG itself

Being pseudo-random (instead of purely random) simply means that if one were to look really, really hard at a long sequence of these numbers there may be some complex patterns that can be found, since the sequence was generated by a formula.

So, if pseudo-random numbers are used for say something like roulette, or dice rolling, where the generated number itself is so close to the surface, that could be bad.

But here, each generated number is so far imbedded in the code, so as mrpolo says, it is not the RNG that needs to be questioned, it is how the code takes those numbers and intertwines them with the player ratings to yield the outcomes.

For an extreme example (not that I necessarily claim whether it is true or false), as others have previously said, maybe there is a built-in upset function. So before the game, with a certain probability the underdog gets a 20% ratings boost. So even if the RNG was purely random, there would be these built-in upsets.

The "randomness" in the HD engine is not based on pure random numbers. There is no reason for it to be. Who really cares if it becomes predictable after the 10,000th number. No one is going to track it anyway.

There are so many different places in the formula where a random number is generated that the C# code is just as useful as any 3rd party source. If he is more comfortable with that then I say go for it.

One major use of the RNG, as I understand it, is to generate the various ranges of NCAA statistics within the game. So if say Free Throw success in the NCAA is 72.3% (not the real number) then the engine will generate results that ultimately will produce that number in the overall. What we see in our boxscores is simply one game's generation and we also have to remember that it is produced using a standard deviation of +/- 3.

As has been pointed out above, the real key here is the formula used not the RNG itself. The reason for that is that is that the code calls for a minimum and maximum range to generate a number within. That min/max may be a percentage, a decimal, or a whole number depending on where it is used. It is those numbers that should be questioned if there is an issue. And before anyone asks, you will stand a better chance of being invited into Fort Knox and told to help yourself, than of ever seeing those ranges.

Quote: Originally Posted By schroedess26 on 1/19/2010Colonels has wanted true random that is harder to get and more expensive for a program like this to use.

Pseudo-Random is where a computer program picks out a number by use of a math formula. It usually is faster and cheaper to get this randomizer but is not truly random as the computer is picking it through the formula

Schro, even the web site that colonels tried to point to to back his claim actually said that using true random numbers would be a negative (thanks to billyg for uncovering this). So this is really all pretty silly and a moot point.

So if the randomness isn't/wasn't a problem, then why is seble changing where it originates from?

True random is still true random, with pseudo random you always run the risk of clustering and other problems. True random is generated by unpredictable factors, and that's what I really like about it, and I will always try to use it for my piddly games.

Quote: Originally posted by colonels19 on 1/19/2010So if the randomness isn't/wasn't a problem, then why is seble changing where it originates from?True random is still true random, with pseudo random you always run the risk of clustering and other problems.  True random is generated by unpredictable factors, and that's what I really like about it, and I will always try to use it for my piddly games.

I would assume they are switching to save money. This way the program itself will create the random number and they don't have to license a third party RNG.

I am not saying its a big deal. I was just bringing it up over here since I had asked a question about it in the beta. Personally it really only matters if it effects the game negatively.

A computer can't be random. It has to be programmed so it can't be real. Close, but no cigar.