What is wrong with Worms Armageddon's random number generator??

[DragonQ]

I usually play a scheme that involves starting with infinite ropes and some blowtorches and parachutes etc., and weapon crates on 100% drop rate. But the thing is, there are always periods where we'll get ridiculously amazing weapons (like 10 in a row), or lots of awful weapons (Skunks, Blowtorches, Parachutes, Bungees etc.) in a row. It very rarely seems to give a balanced weapon selection.

Is the game's random number generator just rubbish?

[yakuza]

Okay, so you're complaining because you have all weapons on 100% and you're getting crap weapons at times, and awesome weapons at other times in a random crate weapon generator... Okay, so, what's the problem?

[Muzer]

He means he gets loads of crap ones in a row, then suddenly loads of awesome ones in a row.

[AndrewTaylor]

I usually play a scheme that involves starting with infinite ropes and some blowtorches and parachutes etc., and weapon crates on 100% drop rate. But the thing is, there are always periods where we'll get ridiculously amazing weapons (like 10 in a row), or lots of awful weapons (Skunks, Blowtorches, Parachutes, Bungees etc.) in a row. It very rarely seems to give a balanced weapon selection.

That's how random numbers and human intuition work. They don't play well together. I'd be sceptical that it was behaving wrongly unless someone did systematic well-constructed experiments. (I guess that'd be the QA department's job.)

[yakuza]

He means he gets loads of crap ones in a row, then suddenly loads of awesome ones in a row.



I know what he means, what part of random don't you understand? He gave us his scheme details, providing he's saying the truth he only needs to play more games he'll be able to figure statistics and randomness have their weird moments, and that this complaint is stupid, and that maybe, he doesn't get balanced weapons through the game because, frankly, if he thinks a skunk is a bad weapon then there might be a chance his judgement is pretty bad, and unless he acknowledges that there is an equal number of awesome+crap weapons, than there is of balanced weapons, chances are (and statistics back this up) he'll never get a balanced outcome.

[bonz]

Play a single game, grab every crate, write down its contents and repeat until you reach 1,000 crates.
Then see if the random generator is really faulty.

[yakuza]

Play a single game, grab every crate, write down its contents and repeat until you reach 1,000 crates.
Then see if the random generator is really faulty.

You can never actually prove a random generator is faulty, never in a lifetime, or infinite lifetimes, well, maybe in infinite lifetimes you can, but never in infinite lifetimes - 0.Ô1

[DragonQ]

Well knowing how many of each crate wouldn't help, you'd need to know the order, and then assign a colour for "goodness" of that weapon, then have a graph. If it kind of looked like red bit -> blue bit -> red bit etc. then it's not working properly.

He means he gets loads of crap ones in a row, then suddenly loads of awesome ones in a row.

Yes. I have been playing this scheme for ages and this is almost always the case. If it was random, there would be a fair lot of occassions where the weapons are average (i.e. good and bad weapons together), but that happens extremely rarely in my experience. That is why I am convinced it's not really random. If it were, flurries of amazing or crap weapons would only occur every now and then.

[yakuza]

Okay, you're right, it's not really random, in fact when the game starts, it reads your mind, and then makes the weapons appear accordingly to what you think is an "awesome weapon" and a "crap weapon" and drops series of these each time to annoy the hell out of you.

[Plasma]

You can never actually prove a random generator is faulty, never in a lifetime, or infinite lifetimes, well, maybe in infinite lifetimes you can, but never in infinite lifetimes - 0.Ô1
Look at the scource code.
:D

I usually play a scheme that involves starting with infinite ropes and some blowtorches and parachutes etc., and weapon crates on 100% drop rate. But the thing is, there are always periods where we'll get ridiculously amazing weapons (like 10 in a row), or lots of awful weapons (Skunks, Blowtorches, Parachutes, Bungees etc.) in a row. It very rarely seems to give a balanced weapon selection.

Is the game's random number generator just rubbish?
I'd say that's all down to the luck of the draw.

[yakuza]

Look at the scource code.
:D





Shh! Don't point the real faults of the random generator or you might make the OP here think he's got a real case :eek:

[DragonQ]

I'd say that's all down to the luck of the draw.

I wouldn't. Not since I've been playing the game for almost 2 years with that scheme.

[AndrewTaylor]

infinite lifetimes - 0.Ô1

That is the poorest attempt at conversational mathematics I have ever seen. There are at least five separate errors in that quote.

[yakuza]

That is the poorest attempt at conversational mathematics I have ever seen. There are at least five separate errors in that quote.

That is the poorest attempt at a conversational troll attempt I have ever seen from a mod around here, and I've seen quite a lot. You're in time to explain why it's wrong though, not that I care, but you might be able to change from a trolling attempt to a constructive criticism off topic post.

[Metal Alex]

constructive criticism

oh, the irony.

[AndrewTaylor]

That is the poorest attempt at a conversational troll attempt I have ever seen from a mod around here, and I've seen quite a lot. You're in time to explain why it's wrong though, not that I care, but you might be able to change from a trolling attempt to a constructive criticism off topic post.

If you don't care I'm not going to waste time listing the various differences between your own imagined number theory and that used by the rest of the world. If you change your mind and decide you want to know then email me about it.

Well knowing how many of each crate wouldn't help, you'd need to know the order, and then assign a colour for "goodness" of that weapon, then have a graph. If it kind of looked like red bit -> blue bit -> red bit etc. then it's not working properly.
It would look like that anyway. Runs happen and people spot them. It's all in that there Derren brown's book. And thanks to confirmation bias, the more you suspect it, the more obvious it will become, even if the game's behaviour never changes.

Like I say, without numbers it seems far more likely to be your imagination -- especially since the groups of "good" and "bad" weapons are so arbitrary and ill-defined.

[ShootMe]

Ok for anyone who wants to check the number generator, I took a shoppa scheme and set the probability of all droped weapons to the same value then let the game run for a while until I had 200 data points. As I was doing it you could tell some weapons poped up a lot more than others. I put the points in a text file with each weapon given a certain number to make it easyier writing.

# - Weapon Name - Times In Crate
05 - Bazooka - 4
06 - Homing Missile - 8
07 - Mortar - 11
08 - Homing Piegon - 11
09 - Sheep Launcher - 7
10 - Grenade - 9
11 - Cluster Grenade - 8
12 - Banana Bomb - 10
15 - Shotgun - 7
16 - Handgun - 6
17 - Uzi - 15
18 - Minigun - 6
19 - Longbow - 8
25 - Dynamite - 8
26 - Mine - 8
27 - Sheep - 10
28 - Aqua Sheep - 6
29 - Mole Bomb - 4
38 - Baseball Bat - 8
46 - Holy Hand Grenade - 5
47 - Flame Thrower - 7
50 - Petrol - 12
51 - Skunk - 11
55 - Cow - 6
56 - Old Woman - 5

Uzi and petrol poped up quite a lot, but I would say the number generator works well since all the numbers are fairly close to the expected average of 8.

[DragonQ]

Stupid Petrol Bombs <_<

Always seem to get either none between us on a level, or ~4. I still would like to know how the number generator works, but that probably won't ever happen. Oh well.

[AndrewTaylor]

Stupid Petrol Bombs <_<

Always seem to get either none between us on a level, or ~4. I still would like to know how the number generator works, but that probably won't ever happen. Oh well.

Computerised pseudo-random number generators almost all work by manipulating massive numbers in obtuse ways, and returning the last few digits, which are so unpredictable that they act as if they were random.