Curse

BlueNovember · May 05, 2007, 02:43 PM // 14:43

Quote:

Originally Posted by Caoimhe

...Each coin toss represents an individual event, not a progression of events, so each coin toss is determined only by its toss, not by ones that precede it.

Would 100 straight heads-tosses be extraordinary? Absolutely. Likely? Not very. Possible? Yup. There are more factors that should be considered before considering the exercise biased, though.

I agree that regardless of previous results, the probability of a coin being a head is 0.5 (treating it as a Bernoulli trial and discounting silly things like "edge"). However there's a difference between "100th throw being a head, given 99 heads" and "100 heads"; the difference beeing 0.5 vs ... some really small number.

Mmm. Diverging from the topic a _little_ bit, but for the sake of completeness,
Probability of h heads in n tosses of a coin with a probability of heads equal to p:

(h + t + 1!) / h!t! * r^h * (1-r)^t

h=100, t=0, r=0.5, (presuming fair coin)

(100 + 1)! / 100! * 0.5^100

101 * 0.5^100,

Which, according to google, is
7.96749514 × 10-29

...which is a number so small I don't even know the si prefix for. The coin is biased. Certainty is well over 99.99%.

Edit: I've never really done probability, just have an interest in it, so any and all of these values could be wrong :P

phil_carter · May 05, 2007, 03:40 PM // 15:40

i would be inclined to think that that was just a result of probability of independent events, yes its surprising that you got the same pet on 3 characters twice, but its possible and given that each one is independent not really that unlikely to happen, to see the real distribution youd need to carry out far more tests, if this happened to more than 10 of your 100 characters then yes i'd be concerned, but its still alot more likely than my hand passing through my desk, or an ice cube jumping out of my glass.

Someone else could equally claim that it is odd that they recieved different pets on every character, its just how the world works, and probably indicates a good random number generator rather than a poor one in which extremes never occur.

And no ive never understood conditional probability, but does it even apply to independent events?

TRUEgamer · May 05, 2007, 03:55 PM // 15:55

I have received one jungle troll and one siege turtle. No complaints here.

Refer to thisfor statistical data showing how the number system works.

netniwk · May 05, 2007, 04:02 PM // 16:02

3 char teddy bears here

LifeInfusion · May 05, 2007, 04:04 PM // 16:04

It's an algorithm. Algorithm = not entirely random.

The OP is correct that it is indeed NOT random. They may use random factors to GET random-ness, but the result in itself is not completely random.

For example, if it was based on any factor, say time, then there is a independent variable the result is based on.

Given that C++ and generally all computers operate on logic statements , we must assume it is "close to random" but not random. Moreover, if the independent variable itself cannot be controlled, it yields a better
"randomness".

If you use a random generator, even that is not random. Any result that can be entirely replicated by manipulating the independent variable is NOT random. It is dependent on an unknown variable.

That's probably why computers aren't used to do lotteries and instead ping pong ball things rolling in a sphere shown on television are used. At least it cannot be rigged. Or can it?

anyway, my suggestion is to try opening it at different times or randomize YOURSELF since it is highly likely it involves timing or what time you logged on, how much you played, or the like.

A good example is the chest drop rate. If you do the same chest over and over you will probably trigger whatever codes they have, and therefore it is a bad idea. Randomizing yourself is the way to go.

bamm bamm bamm · May 05, 2007, 04:18 PM // 16:18

Quote:

Originally Posted by Myrkwid

The probably standard one used by ANet, but you don't know, so this comment is worthless.

Perhaps I shouldn't have even mentioned different algorithms, since it matters less than the sheer volume of people playing and has just allowed you to dismiss my comment out of hand without really saying anything of merit. It doesn't matter how unlikely something is or how random their algorithm is, GW has millions playing. Lets estimate the chance of receiving a common minipet three times in a row is ~0.0007 ((0.7*0.125)^3). Multiplied by 2 million is 1400 occurrences. That's an estimation (and probably wrong) but it helps to illustrate my point. According to Guru there are 2368 people currently online. ((0.7*0.125)^3)*2368 = 1.586375, or 1 person. I guess that would be you.

Quote:

Originally Posted by Myrkwid

Again, even the weirdest things can happen, the lottery tossing the same numbers twice in a row. But if such things happen on a frequent basis, then there is no real randomness.

But this isn't happening on a frequent basis, so what's your point? A pseudo-random generator is broken because it has produced something unlikely at least once? That's kind of what they do. I would have to see this happen many hundreds of times before you convince me there is anything wrong going on here.

Quote:

Originally Posted by Myrkwid

If a small percentage experiences a bug, do you want to tell them "there is no bug, look at the others who don't experience it! COme back with 20.000 others and we might have a look at it." And if you read it, then you might have noticed, that I'm asking if others experiencing the same. But obviously posts in a forum are subjected to fanboism.

They haven't experienced a bug at all. People are randomly given minipets. You have received three the same. You might even get a fourth, but the chance is drastically diminished. Where's the bug? Not to say that there isn't one, but your evidence proves nothing, and certainly provides no grounds for a firing.

Thunder79 · May 05, 2007, 04:26 PM // 16:26

Anyone who has studied statistics knows that the statement made by the OP is a bad statement because the sample size is too small to make a good statement about the randomness of the sample.

The population is all the minis that people got as birthday presents.

Your sample is all the minis that you received on your characters.

the population is made up of hundreds of thousands of birthdays (maybe approaching or more than a million, I don't know)...

your sample is talking about a very small number...so small that you can't reach a good conclusion about the randomness of the random number generator because your sample is most likely not a representative sample for the population.

I've never gotten anything better than a purple for birthday presents...most have been white...do I care? No...it's something that's given out free...as a birthday gift. The fact that I get a gift makes me happy...because most games don't do stuff like that.

XvArchonvX · May 05, 2007, 06:44 PM // 18:44

I swear, every day there has to be at least 1 new thread from some Joe Blow who has a streak of bad luck and has to come here to cry about it and blame the coding of the game.

HawkofStorms · May 05, 2007, 07:30 PM // 19:30

Agreed Archon. I LOVE the conspiracy theorists that come on these forums. There is nothing broken, get over it. The only reason you think it is is because only the people who think that are are posting "evidence" on this thread to support that.

VitisVinifera · May 05, 2007, 07:52 PM // 19:52

wow it's not just me

both canthans on their 1st a week ago or so got jade armors

vixro · May 05, 2007, 08:08 PM // 20:08

While I agree somewhat that minipet randomness is far from good random code, retaining lock picks, salvaging, gold/purple chest drop rates, drop rate findings while farming and even coffers are all random.

I can do a chest run and get 15-20 golds in a row, while retaining a large amount of lock picks(maybe 8-10 of them with only ~ 25% chance of saving) and then 5 hours later do the same chest run and lose every lock pick and get all purples. Sometimes when I used to farm tengu in twin serpent lakes I would get as many as 6 golds within the first 2 groups I killed, sometimes I would get absolutely none in the entire run.

When I bought around 30 coffer of whispers I didn't get anything good, but everything I got was different almost every single time. I never got the same thing 3 times in a row. It may have been the same 5-6 items, but that's because I was getting the shitty stuff instead of the good stuff. That happens because of the drop rate of the better items.

Now, the problem the OP is seeing and that I see too is that this isn't the case with minipets. Sure it would be frustrating to have all white minipets, but that can happen just because the drop rate is going to be higher on the common pets over the uncommon, rare, and unique drops. What ISNT random is when you keep getting the same minipet over and over again. If you have 10 presents open side by side and you get 5 minipets of one type and then 5 of another out of a possible 15 minipets then something is seriously wrong. That is a 1 in 759,375 chance and you hit it twice. That's far from unlucky, that's just plain retarded. Although that number would mean that each minipet is considered equal, the number is still well in the hundreds of thousands in odds.

Dragonious · May 05, 2007, 08:24 PM // 20:24

Quote:

Originally Posted by makosi

I recieved 3 Fungal Wallows on my main account and a Charr Shaman on my abandonded secondary account.

I know it's a conspiracy theory but I believe its related to farming flags in some way.

LOL! Thats basically what I got! I had a whiptail and 2 fungals and on my second I had the same charr shaman.. :-P

Marth Reynolds · May 05, 2007, 08:31 PM // 20:31

Quote:

Originally Posted by XvArchonvX

I swear, every day there has to be at least 1 new thread from some Joe Blow who has a streak of bad luck and has to come here to cry about it and blame the coding of the game.

But what if the coded chance generator really has bugs?

So far speaking from my own experience i have lucky streaks or unlukcy streaks with so now and then single timed luck or bad luck.
For example using 42 locpicks in the last 18 12 didn't break, Doing chest runs with 12 golds in a row and then 24 purples in a row.
I know for sure this generator spits out more streaks then normal "(bad) luck spikes". for me at least, i can't say it's the same with other people simply because i haven't moniterd their drops .

neshim · May 05, 2007, 09:33 PM // 21:33

You can always count on A.Net to create bad coding!

Neo Nugget · May 06, 2007, 02:45 AM // 02:45

same account.2 horsemans.
only double of the mini's

HawkofStorms · May 06, 2007, 05:27 AM // 05:27

Any person with any mathematical training will tell you...
statisitical anylsis with small, biased sample sizes can not get statisitically significant results.

Kool Pajamas · May 06, 2007, 07:06 AM // 07:06

Ive gotten about 6 1st year birthday presents and only 1 have been double. Yay me i guess.

Midnight Harmony · May 06, 2007, 01:42 PM // 13:42

I've had two first year presents so far, one being a Temple Guardian and the other a Kirin.

teenchi · May 06, 2007, 04:09 PM // 16:09

An alliance member of ours just got 3 rot wallows on his 3 in the same day.

Echuu Ishtar · May 06, 2007, 04:20 PM // 16:20

Something really needs to be done about the randomness of the birthday gifts. The probability of getting a rare one should increase depending on the /age of that character. It isn't really fair if somebody puts their heart and soul into a character for a year and opens up a Fungal Wallow as a gift, yet somebody with a second account that just collects bday gifts could possibly get a Bone Dragon. If they could somehow get around people that would afk to rack up hours, this method would be perfect I think.

Or better yet, take away the rarity of the pets.