I think the RNG is broken....

[dantiesilva]

So I am in two separate games and for the life of me I can not roll anything but a 1 when I attack. Even on one attack I not only rolled a 1 on the attack but on 3 separate damage rolls. The count is up to 10 in a row....Just wondering if this is only happening to me or what.

[SiuiS]

The RNG has been working fine on my end for the last week, including two hours ago.

[Togath]

works fine for me as well;

testing(since I heard something about a possible bug):
[roll0]
[roll1]
[roll2]
[roll3]
[roll4]
[roll5]

[roll6]
[roll7]
[roll8]
[roll9]
[roll10]
[roll11]

[dantiesilva]

So I am just the unlucky one who gets to eat all the natural 1's. Yay.

[Strawberries]

So I am just the unlucky one who gets to eat all the natural 1's. Yay.

It's called "having angered the Dice Gods". Solution is the sacrifice of a virgin at midnight.:smalltongue:

Jokes aside, I've been using the RNG for a long time, and it seems to be working fine for me (well, it rolls low when I need high and vice versa, but that's normal behaviour. :smalltongue:)

[Neftren]

Just to chime in, when something is random, it's not uncommon to see sprees of the same number over and over again. That's the whole deal with random. :smallbiggrin:

[The Bushranger]

The general consensus of the people I play with is that the GitP dice roller is loaded low, and I have to agree (at one point it got to where my DM got furious at it and ruled not to use it anymore, period - none of us argued). InvisibleCastle (http://invisiblecastle.com/roller/) is an alternative if the dice hate you, too.

[dantiesilva]

I like this castle first roll was a natural 20. Thank you, now just to get my DMs to allow it.

[Lateral]

The general consensus of the people I play with is that the GitP dice roller is loaded low, and I have to agree (at one point it got to where my DM got furious at it and ruled not to use it anymore, period - none of us argued). InvisibleCastle (http://invisiblecastle.com/roller/) is an alternative if the dice hate you, too.
Just to put in, it might feel that way, but it really isn't. I did a test not that long ago, and it came out to average 10.68. Yes, that's not perfectly 10.5, but it demonstrates that there's almost certainly no significant, objective bias.

[Khedrac]

Just to put in, it might feel that way, but it really isn't. I did a test not that long ago, and it came out to average 10.68. Yes, that's not perfectly 10.5, but it demonstrates that there's almost certainly no significant, objective bias.
Sorry but while your test may have shown that there is no significant bias, the average being near the correct average does no such thing. Ten 1s and ten 20s has the same average as one of each number 1 through 20, but would suggest a clear bias.

One problem with computer RNGs is that they cannot be truly random, and their result distributions can appear surprisingly clumpy to the users, even if they pass all tests of function. Now real random numbers are clumpy, it's part of the randomness, but something about the computers makes them seem worse. Good case in point it World of Warcraft's RNG, most players will curse the RNG and say it is far too clumpy - but is it really just being a better RNG than people are used to?

[FujinAkari]

Sorry but while your test may have shown that there is no significant bias, the average being near the correct average does no such thing. Ten 1s and ten 20s has the same average as one of each number 1 through 20, but would suggest a clear bias.

I have also tested it and no value registered a significant deviation to ANOVA testing. While it is true that computers cannot truly be random, it is also true that dice are also not truly random. Frankly, computers do a better job :P

[Brother Oni]

It could be that because multiple people are taking separate samples of the RNG, individual samples may be skewed while the whole distribution is sufficiently randomised.

[Neftren]

Sorry but while your test may have shown that there is no significant bias, the average being near the correct average does no such thing. Ten 1s and ten 20s has the same average as one of each number 1 through 20, but would suggest a clear bias.

One problem with computer RNGs is that they cannot be truly random, and their result distributions can appear surprisingly clumpy to the users, even if they pass all tests of function. Now real random numbers are clumpy, it's part of the randomness, but something about the computers makes them seem worse. Good case in point it World of Warcraft's RNG, most players will curse the RNG and say it is far too clumpy - but is it really just being a better RNG than people are used to?

Even the most basic of computer RNGs are random enough for most purposes. Nobody writes a RNG that is insufficiently random for the purposes of simulating dice rolls. We're not doing Monte Carlo simulations here... seeing clumps of numbers is to be expected, and perfectly acceptable within a range of random numbers.

[Destro_Yersul]

The general consensus of the people I play with is that the GitP dice roller is loaded low, and I have to agree (at one point it got to where my DM got furious at it and ruled not to use it anymore, period - none of us argued). InvisibleCastle (http://invisiblecastle.com/roller/) is an alternative if the dice hate you, too.

I disagree, BR. Dark Heresy calls for low numbers, and the roller loves giving out high numbers for that game. It's not that it's loaded, it's just that it hates everyone. :smalltongue:

[SiuiS]

Gentlefolk, I believe there is a better place for discussion?

[Roland St. Jude]

Gentlefolk, I believe there is a better place for discussion?Sheriff: Agreed.