Does the RNG really even work?

[Dunno]

Read damage calculations-even if you have a 70% chance of hitting there's always that 1% chance of missing all attacks. This DOESN'T mean that after 100 attacks 1 will miss all. It means that PROBABLY after 100 attacks 1 will miss all. If you're unlucky this may happen pretty often, live with it.

[Mmoor]

I personally don't understand why anyone doubts the RNG (whatever that stands for). This is not directed at anyone, I just don't seem to question wether or not probabilities will be consistent. It just seems natural to me that sometimes things will randomly put me out of favor, but I guess I got used to that from my days as a Warsong player.

[Hulavuta]

It stands for Random Number Generator

[Mithrus]

I've read the other post comparing the Wesnoth RNG with a hardware emulated version. I understand people have tested large samples and the end result is that the RNG performs accurately. My point still stands that "perception" can be more valid than reality. If my friends and I perceive a game has an unbalanced RNG, then reality takes a back seat. The distribution of the RNG is poor for this kind of game.

P.S. I can appreciate the frustration on the dev's part, as I am also a software dev (business, not gaming, so I don't use RNGs ever). I am sorry for hitting a nerve on this issue, but my friend and I (who play Wesnoth a LOT) keep feeling frustrated when a high hit chance misses every time (or every unit misses for that entire turn!). IMO, I care less about an RNG's long term ability to provide statistical accurate numbers than it's ability to allow me to enjoy the game.

How difficult would it be to make RNG PnPable like the AI engine? This way, should someone come up with different algorithms, it would be a bit easier to have the general public try them out and provide feedback.

[Sapient]

If the Problem is in your Perception, not in Reality, then by definition it is not a Real problem only a Perceived problem. The "fix" then, is an improved understanding of flaws in our own human perception of randomness.

As a software developer, you should understand that if Wesnoth displays "70%" and then changes the RNG so that it is not 70%, then that would introduce a Real bug, not just a Perceived bug.

Also, please provide your replay. We consider it important data for the "fix"

[Gambit]

IMO, I care less about an RNG's long term ability to provide statistical accurate numbers than it's ability to allow me to enjoy the game.

You might be interested in several projects being developed over at Wesnoth's experimental fork then.

[Dave]

My point still stands that "perception" can be more valid than reality.

This "point" cannot "still" stand since you didn't make it in your original post. You simply said there was a problem with the random number generator not generating numbers that closely match a random distribution and presumably would have been able to produce proof to back it up.

If you have any kind of evidence that the random numbers generated in Wesnoth do not closely match a random distribution, then please post a bug report at http://bugs.wesnoth.org with the evidence attached. I am locking this thread because we're not interested in a discussion about perceptions of random numbers. It's been discussed enough times already.

David