RNG superstitions

[AI]

Elves' faerie fire is magical, this ensures it misses 70% of the time.

[DDR]

The only way to kill a full health shadow (on hills) is with a half-health ulfserker (on plains). This ties in perfectly with the first point made in this thread. And it is actually true.

[JW]

Nagas lose to mermen at any time of day because they are blind.

[DDR]

A high-level unit surrounded by low-level units will never hit anything: all the luck will go to the surrounding units, especially those with 2-2 attacks and the like.

[A Guy]

-The closer a unit is to leveling, the easier it is to hit.

[kodama]

This thread gave me a good laugh

Actually, I wonder how many RNG instances are used to control the units' fight. The RNG algorithm used, while not a bad thing, is simple; it might not be multi-dimensionally equidistributed, e.g. if two units draw a sequence of numbers from the same instance of the RNG, the sequences they get might not be statistically random enough, although if one single unit had drawn all the same numbers (the two above sequences mixed into one), the statistical randomness of this sequence might have been higher.

After all, there could be something behind all this RNG superstition

[grrr]

I tell you a secret now, kodama: numbers come from a single source (actually, if I understood it right there is even a cache, which is easily exploitable then). I was advising to have a single source for each player a while ago, but the current model is simply easier. And random numbers are random, so it should not matter anyways, right?

Also, good players have a close look at the generated numbers. If all goes too well and they got their lucky hits they won't push further but rather set up for defense, simply because *they know* the counter attack has a high chance to fail now.

And you can test the RNG (that is, trying to miss twice or thrice in a row) too before you go for a critical hit (say, horsie on leader) =)

So, the question is, do you still laugh kodama?

[Jozrael]

The above post has me baffled @_@

[TobiasParker]

Grr if that is true (and i will take your word for it as a 'vet') than everyone complaining about the RNG is absolutely correct. If you can predict the outcome than the numbers are not truly random. I mean, we may not know the exact numbers but if you can use your knowledge of the RNG to plan strategy...then...i just don't know.

[zookeeper]

No, it's not true.

[Jozrael]

Ok, now Tobias remark at least has illuminated the POINT of this post. Honestly I guess I was just tired, I didn't understand a word it was saying.

I tell you a secret now, kodama: numbers come from a single source (actually, if I understood it right there is even a cache, which is easily exploitable then). I was advising to have a single source for each player a while ago, but the current model is simply easier. And random numbers are random, so it should not matter anyways, right?

Yes, the numbers come from a single 'source', the RNG. The cache, IIRC, is entirely unrelated to the RNG, that just loads Wesnoth data for easy use. All modern processor intensive games have a cache, its what reduces the need for those lame 'loading' screens. The cache is not exploitable in the least in terms of MP cheating.

This is not to say that RNG cheating is impossible in MP, but its not done by most players. You'd have to ask a coder dev as to why they haven't fixed this: its probably very difficult to prevent and its happening in so few games they don't deem it necessary.

Also, good players have a close look at the generated numbers. If all goes too well and they got their lucky hits they won't push further but rather set up for defense, simply because *they know* the counter attack has a high chance to fail now.

Entirely false. This is a logical fallacy that you'd do best not to fall into. Just because I just rolled 6 hits on a row against someone with 70% defense does NOT mean my next hit has ANY chance to hit besides exactly 30%. Its what a RNG does. Generates RANDOM numbers, that have no regard whatsoever for previous rolls. Human thinking likes to make patterns out of it, so they'll either think that the unlikely has already happened, so it won't happen again so soon, or that its 'hot' and it'll continue giving unlikely results.

Neither of these is true. The Wesnoth RNG is a perfectly random RNG. Also, using the SAME RNG for two or more players does NOT decrease the randomness for either player. It changes the specific results, but each random roll i is independent of the next roll, i+1, and of the next roll, i+2. So it doesn't matter whether the numbers come from a single source or not.

And you can test the RNG (that is, trying to miss twice or thrice in a row) too before you go for a critical hit (say, horsie on leader) =)

So, the question is, do you still laugh kodama?

Also a logical fallacy in the same vein. Previous results guarantee nothing with the RNG, nor even hint at later results.

[turin]

Hm... I can't tell if Jozrael is brilliantly extending the joke even further, or if he missed the fact that this entire thread is a joke.

[Wintermute]

Hm... I can't tell if Jozrael is brilliantly extending the joke even further, or if he missed the fact that this entire thread is a joke.

It's like april fool's day all over again!

[Jozrael]

I guess I should have included some logical fallacies of my own for a more accurate reply

[DDR]

There is a small part of all of our brains that does nothing but look for patterns. True randomness mucks it up, and we 'see' patterns, like being overdue for a hit after a few rounds of misses.

I love psychology.