I am quitting Arelith forever

[Diegovog]

I swear I roll more 1's than the enemies roll 20's... It's really easy to notice when you're a low lvl character with 1 attack per round.

[AlauraElizabeth]

I’ve been rolling high in everything for too long now. I’m gonna get absolutely fooked later when I actually need it

[Elaetheus]

If you go to another server, you'll get the same thing, promise!

On a serious note, I think it might be interesting to go and discuss that with Beamdog. Maybe they can actually rewrite the RNG for a more modern, much more (pseudo)-random.

[flower]

go with potions. always getting high increase on third to fourth

[Aodh Lazuli]

I'm at 33 Masterly Damask Rapiers sacrificed for a chance at keen (and counting), if it makes you feel better.

I exploded 49 addy helms.

[Rabbid]

I got triple crit to death today by the wharftown boys leader.

And this is a common occurrence.

By triple crit I mean he rolled 3 20s in a row and the crits landed. My spellsword was not happy.

[-XXX-]

I believe that it was mentioned some time ago somewhere that the engine doesn't actually generate random numbers, but a stream of numbers that approximates randomness. As a result the variance allows more often for instances of incredible luck streaks and vice versa. [quotation needed]

[Marsi]

Here's a thread on Ravenloft's forum discussing the phenomenon:
http://www.nwnravenloft.com/forum/index ... ic=36024.0

It's a pretty interesting topic. Basically, computers struggle to generate truly random numbers (pseudo-random). Some RNGers are better than others (like C++ rand(), which is considered awful), but frankly, the difference is really only of importance for sensitive fields such as cryptography. In any other application the quality of RNG would be improbable to pick up by human observation, so however bad NWN's RNG facilities are (in the linked thread it mentions it just calls your OS's RNG iirc), it would never meaningfully affect you. As argued in the thread, streaking is statistically normal and we fall victim to human perception bias.

[Party in the forest at midnight]

I believe that it was mentioned some time ago somewhere that the engine doesn't actually generate random numbers, but a stream of numbers that approximates randomness. As a result the variance allows more often for instances of incredible luck streaks and vice versa. [quotation needed]

That's what I've noticed, sometimes the number generator gets caught within ranges. It often doesn't matter, but when it starts to be in a low range (or even high range) it's very observable. My guess was it was something with how it drew the randomness seed, that the seed occasionally favoured rolling specific ranges. But if you ever say that, you get dismissed with "lol no its confirmation bias," as you can see with the Ravenloft thread. That thread still makes me salty.

[Tryn Dralar]

I never use a wand for the same stat back to back. It's a head game. For me strength is most important. I use that first. If I roll a 2 but in need of a 5, I will use a different buff first before trying straight again. At least it seems to work better that way. *Shrugs*

[Marsi]

I believe that it was mentioned some time ago somewhere that the engine doesn't actually generate random numbers, but a stream of numbers that approximates randomness. As a result the variance allows more often for instances of incredible luck streaks and vice versa. [quotation needed]

That's what I've noticed, sometimes the number generator gets caught within ranges. It often doesn't matter, but when it starts to be in a low range (or even high range) it's very observable. My guess was it was something with how it drew the randomness seed, that the seed occasionally favoured rolling specific ranges. But if you ever say that, you get dismissed with "lol no its confirmation bias," as you can see with the Ravenloft thread. That thread still makes me salty.

Here's an interesting tidbit from a different thread:
https://forums.beamdog.com/discussion/6 ... lls-rigged

Random() is random in NWN. It just uses the underlying libc rand() function.

There is a tiny bias towards lower numbers, but you'll notice it once in 4000 rolls at worst, and likely once in 200 million rolls. Not something you should be concerned about.

There were indeed problems with random seed. The client re-seeds the RNG based on static objects and tiles in the area. This is done so that the randomly placed grass is placed in the same spot for all clients. In versions prior to 1.65, the game would then forget to reset this seed, resulting in repeatable random sequences. That has since been fixed.

[susitsu]

I came here for the sweet nectar of rage.

I left here feeling it.

[MoreThanThree]

Here's an interesting tidbit from a different thread:
https://forums.beamdog.com/discussion/6 ... lls-rigged

The client re-seeds the RNG based on static objects and tiles in the area. This is done so that the randomly placed grass is placed in the same spot for all clients.

god save bioware, lord of the game

[dominantdrowess]

I'm at 33 Masterly Damask Rapiers sacrificed for a chance at keen (and counting), if it makes you feel better.

When rolling a d20 IRL, to nervously get a feel of how many adamantine greatswords I'd need to sacrifice I rolled 192 times on my table, next to my keyboard without rolling a natural 20, counting every roll. =/

Then? I got the Keen Greatsword the next day, on my 13th Greatsword.

There's nothing wrong with Neverwinter's Random Number Generator. You guys just have an inaccurate idea of how probability works. >D

Knowing how the number generator works does not allow you to predict it's numbers in game-play, which is what makes it effectively random. I'd say I have more (personal) control over my own actual dice rolls on a d20 than over the number generator ... which effectively makes the natural dice rolls IRL less random than our game's RNG: Which is good for the game.

[Red Ropes]

I'm at 33 Masterly Damask Rapiers sacrificed for a chance at keen (and counting), if it makes you feel better.

So many of Qaelan's children are drowned.

[Alox]

(guess this post turned out be some random thoughts on RNGs more than an actual response....)

Random number generators are not true random (as already mentioned). Though most of the times RNGs are tested on the final distribution, and not much care is given to how streaks distribute (correlation of successive values). So they are likely to roll equal amounts of 1s and 2s etc., but maybe streaks appear more often (or less often) than they should.

I would assume that IF the RNG used by NWN is streaky, it is likely that streaks in the boundary numbers like 1 and 20 is something you react on, while a streak of 11s might go unnoticed, even though they appear equally often.

This thing with "player perception" is a phenomenon that is big enough to cause a lot of game developers to work on this, something that is also discussed by Sid Meier in his known talk from GDC 2010. To prevent streakiness one can make small pools of possible outcomes and then use the RNG to randomize the order the outcomes are given in, like drawing from one deck of cards, which at the most extreme gives you four aces in a row, never more. This obviously will prevent a streak given small enough pools (but if you know the pool size you can count the outcomes and predict future rolls). In other games, the player simply gets a bias (cheat die), so they don't experience the bad streak so much. Or even a diminishing return with a forgetting factor on outcomes, to prevent the same outcome from appearing too much. The player psychology is as you can see also an important factor.

To return to the streakiness, the RNGs used by games (often the RNG called LCG) is known to exhibit "streakiness" (more precisely it follows Marsaglia's Theorem). Modern operating systems offers special functions to avoid these old RNGs, like CryptGenRandom in windows. So it is possible some players who play a lot may eventually pick up on this... But maybe the RNG in C compiler used by beamdog is a new RNG with have improved upon these old LCG problems?

[Astegard]

I swear you have a higher chance to crit when an enemy is already near death.
I'm always like " ugh cant you do that when they are full health?"

RNG needs to ensure they are absolutely dead.

[Diegovog]

I swear you have a higher chance to crit when an enemy is already near death.
I'm always like " ugh cant you do that when they are full health?"

RNG needs to ensure they are absolutely dead.

I can confirm this.

[Rabbid]

I swear you have a higher chance to crit when an enemy is already near death.
I'm always like " ugh cant you do that when they are full health?"

RNG needs to ensure they are absolutely dead.

I can confirm this.

Can confirm it mostly happens like this on a scythe wm