Gambling You ever notice this?

[ezvpn]

Gambling You ever notice this?

I usually gamble with at least 5 lvl 80+ chars full of gold, currently using a lvl 93 paly. I clear my stash except for gheeds and start clicking. If I gamble straight amulets/rings, i might get one or 2 rare rings/amulets with usually a quality one every 3rd or 4th stash sell back. However if I gamble all rings/amulets buying each one that comes up, i'll have 10+ rares/set and the occasional unique, usually junk rares but still. I expect the answer to be just odds, but it doesn't really seem to be that way when there is such a noticable difference when gambling both ammy/rings than just a single item. Alot of times ill just be gambling ammys, magic magic magic than i'll gamble a ring and its rare. Anyone else gamble straight rings or ammys or notice a difference when gambling both?

ez

 

[Are]

Re: Gambling You ever notice this?

The stats of the items are created when you open the gambling screen, so any differences in stats based on what types of items you buy are entirely coincidental

 

[ezvpn]

Re: Gambling You ever notice this?

The stats of the items are created when you open the gambling screen, so any differences in stats based on what types of items you buy are entirely coincidental

I'm not denying any stats on items being created when the gambling screen is opened. Maybe I wasn't clear enough. Gamble 37 ammy's, post how many rare/unique/set are received. Gamble 37 rings and note the same. Than gamble rings and ammys until you have 37 total, count your unique/set/rares. Do this 5 times consecutively only stopping to get more gold from your other chars. I'll record my next gambling actions and post my results. Feel free to do the same to compare results. I should be full again by tomorrow.

 

[droid]

Re: Gambling You ever notice this?

Do this 5 times consecutively only stopping to get more gold from your other chars.

And then do it another 22 times....even at 37 gambles a round, thats not statistically significant. About 1000 total gambles should give you a decent idea, and I suspect you'll find you're right at the reported percentages for your overall magic:rare:set:unique ratios

 

[Dawnmaster]

Re: Gambling You ever notice this?

It doesn't only depend on said odds.

The odds are being fed by a random generator, which uses a weird seed.

Some games you can definetally see some stats pop up more than other times (it's most obvious if one game you don't get any skill amulets, and other games you get a random skill every 3 gambles)

Same goes for rerolling skillers, you can try out your luck per game to see which seed it uses, my recommendation would be to roll 7x per game and switch games until you get a skiller, then stay in that game.

 

[BKC]

Re: Gambling You ever notice this?

for something so incredibly random this is very very very farfetched if you ask me.... are you guys sure superstition isn't getting a hold of you?

 

[Hrus]

Re: Gambling You ever notice this?

I don't say that the random generator in D2 doesn't do weird things, but this...
Anyways I don't like to even think about gambling this way until I get some serious data sample.

BTW: I hope you don't gamble amulets with lvl80 characters, +2 charskill prefix is not possible for them. Also, use Edge for another gold discount.

 

[MuffinMan]

Re: Gambling You ever notice this?

It doesn't only depend on said odds.

The odds are being fed by a random generator, which uses a weird seed.

Some games you can definetally see some stats pop up more than other times (it's most obvious if one game you don't get any skill amulets, and other games you get a random skill every 3 gambles)

Same goes for rerolling skillers, you can try out your luck per game to see which seed it uses, my recommendation would be to roll 7x per game and switch games until you get a skiller, then stay in that game.

Interresting, I didn't know about the seed thing. Just started gambling bigtime with a lvl95 char, so this could be usefull information for me atm.

Any links regarding the "weird seed" and how it works would be appreciated.

 

[Dawnmaster]

Re: Gambling You ever notice this?

Interresting, I didn't know about the seed thing. Just started gambling bigtime with a lvl95 char, so this could be usefull information for me atm.

Any links regarding the "weird seed" and how it works would be appreciated.

Well I don't have any proof for that.

In my younger days I programmed a lot on a Commodore64 and 128 computer, and the random function used a seed to generate a sequenze of random numbers, but some seeds generate a recurring sequenze of random numbers.

x = 1 + int ( 6 * rnd ( 0 ) )
would be the standard line to generate a random number from 1-6
(dice simulator)
now, if you replace rnd ( 0 ) with -1 (iirc) you get a series of random numbers, but the next day, if you turn on your computer and start the program, it generates the exact same series of random numbers

I'm not saying that the stats are simply chosen by that seed, since that would give the same items over and over, but that every game uses a different recurring seed in conjunction with another random number, that would explain most results the best.

I'm guessing the most likely possible part of the seed is the gamenumber (used to hunt dclone)

However, the only way to "proof" this, would be to make 100 games in a row, and gamble 100 items per game, and write down all gambled stats and every gamenumber.
Unfortunatelly, I can only gamble up to 50 items in either 1 game or multiple games, so a testing project of this size would require a significant amount of gold.

 

[Qveasd1]

Re: Gambling You ever notice this?

I'm guessing the most likely possible part of the seed is the gamenumber (used to hunt dclone)

What is this gamenumber and how to find it? Looking through DClone hunting guide didn't help

Anyway, if the period of your random number generator was so small you could see it on a dice roll, then it is a very bad RNG It is understandable considering the resources of Commodore64, but using a bad RNG in an online rpg game is inexcusable. Using fewer different seeds than 2^32 is inexcusable too.

 

[wickedswami]

Re: Gambling You ever notice this?

I really think your observations have more to do with random chance then with seeds.
Though, to be sure, about 1 million games, rulling 1 million charms would be required, so I can't really tell...

 

[boogyman19946]

Re: Gambling You ever notice this?

Dawnmaster is right here =] Looks like he has done his homework in the programming field ;]

The numbers generated by the computer are never truly random. I'm not sure if true randomness can actually happen on a computer but even nowadays, it isn't that easy for a computer to generate a truly random number.

The randomness generators that I have used in programming, were actually producing something called pseudo-randomness. The way the computer determined a "random" number was by using a number called the seed. There is a formula for the computer to use or something among those things, that based the supposed randomness off the seed number, meaning that using the same seed number will always generate the same results. How to solve this? You need a random seed number.

Now, one might think "how can you generate a random seed number when you need a seed number to generate any randomness?" (hehe, i love how mangled up this is already ;] ), well, in my case, this is based off of time. The time it takes between seconds (milliseconds more likely) is very little and the seed number can be generated out of that, but since the processors nowadays run at almost 3 billion cycles per second, chances are that even lower terms of time measurement are used, such as microseconds or something of the like.

As an analogy imagine a magician flipping cards and telling you to stop whenever you feel like saying stop. This is the sort of way that's used to generate the seed number in the scenario i'm presenting.

Now we got our semi-random seed number, we plug it in to the functions and everything ticks like a Swiss watch (hehe.. watch=time... cheesy jokes anyone? ). As you can see, everything happens so fast, that it's impossible for a human, and probably impossible for a machine, to determine what the seed number was and perform the calculations fast enough to predict the next number, hence considering this type of generation as "random".

Those here are the very details of programming as I have learned them =] This is the way I learned to generate randomness and am speaking purely out of prior knowledge, I haven't done any particular research on this so I might be off

 

[Are]

Re: Gambling You ever notice this?

The fact that the random number is based on the computer clock still doesn't mean that 7 charm rerolls is anywhere near enough to determine if a particular seed is giving you a higher or lower chance for rolling a specific affix though. That would be superstition

 

[droid]

Re: Gambling You ever notice this?

Yes but even that kind of time-based pseudo-randomness isnt going to generate patterns coherent enough to notice based off a few dozen samplings, like he suggested happens with "hot games" to gamble or roll skillers.

Maybe over hundreds of thousands of samples, but for all intents and purposes here (unless it was just coded really poorly, but RNGs are a pretty basic coding exercise) Diablo's RNG-generated numbers should behave like true random numbers.

 

[boogyman19946]

Re: Gambling You ever notice this?

The fact that the random number is based on the computer clock still doesn't mean that 7 charm rerolls is anywhere near enough to determine if a particular seed is giving you a higher or lower chance for rolling a specific affix though. That would be superstition

I'm sorry I should've been more clear on this. I meant to say that dawnmaster was right about random numbers being determined by a seed, didn't say his calculations were correct. I should've specified that in my post, my bad.

Yes but even that kind of time-based pseudo-randomness isnt going to generate patterns coherent enough to notice based off a few dozen samplings, like he suggested happens with "hot games" to gamble or roll skillers.

Maybe over hundreds of thousands of samples, but for all intents and purposes here (unless it was just coded really poorly, but RNGs are a pretty basic coding exercise) Diablo's RNG-generated numbers should behave like true random numbers.

You are right. Firstly, I don't know where the seed number is being implement and if the game is using only one seed number. Let us agree that the general meaning of being random is that something can't be predicted. There is a million ways to achieve a pattern of numbers beyond the scope of being predictable. As I said, I don't know where the seed is implemented in the randomness generator, and I definitely don't know if a single game uses only one seed number. As far as we know, a seed number can be generated each time it is required (like everytime you open the gambling window since the stats are determined upon opening it), making it different throughout the game.

Experimental probability is always different from theoretical probability, and hitting the same microsecond or w/e Diablo is using, is based on chance (1/100 for anything below seconds.). The sole idea is to make the numbers unpredictable.

 

[BobCox2]

Re: Gambling You ever notice this?

You see results faster buying 1 space size items.

I think that's what is giving you the impression you get more of them.

 

[fledgeling]

Re: Gambling You ever notice this?

Maybe this is some form of blizzard's defense versus botting, not sure.

 

[ezvpn]

Re: Gambling You ever notice this?

That's one of my original thoughts as well, something kicks in when repeatedly gambling the same specific item over and over.

ez

 

[ezvpn]

Re: Gambling You ever notice this?

Well, thought i'd post my gambling results. Next time I will do all rings, than all amulets than all rings/amulets to get a longer trend. It is interesting to see a spike everytime that i switched from gambling one specific item to another. But according to the randomness probabilities this could be just random as well. Further testing should be done to compare results:

Rings 37 each time
5 rares, 1 set
7 rares
3 rares
1 rare
4 rares

amulets 37 each time
6 rares
3 rares
3 rares
3 rares
3 rares

rings/amulets 18/18
3/1
4/0 ran out of gold here, only gambled 16/16. I should done more of this to confirm/deny what i was talking about. I'll do this first next time.

 

[Dawnmaster]

Re: Gambling You ever notice this?

What is this gamenumber and how to find it? Looking through DClone hunting guide didn't help

Anyway, if the period of your random number generator was so small you could see it on a dice roll, then it is a very bad RNG It is understandable considering the resources of Commodore64, but using a bad RNG in an online rpg game is inexcusable. Using fewer different seeds than 2^32 is inexcusable too.

Gamenumber: (I might have goofed on using the correct term here, it's actually servernumber)
http://www.purediablo.com/forums/showthread.php?t=739712&highlight=dclone

About the dice roll: I said that particular commandline was to simulate a 1-6 chance, if you leave out the int, you'll get a billion different numbers rather than 6.

And to the rest, for a random seed, my calculations would be rubbish indeed, but for some reason everybody missed the part I said about some seeds giving the same series of random numbers, thus creating recurring randomness.

For all we know, Blizzard might have intented it that way, so it would give people a chance to find a game where a certain mod was more likely to get spawned, and then try your luck out there in a legal way.

And it's not just gambling, you see it in drops as well, some games drop zero runes while playing for hours, some drop one on average for every quest you pass, some games you get 3 times the same set item, while others you see none at all.
Some areas even have an unconfirmed increased chance of spawning certain items more than others (eg: runes in travincal, on average, you will get a random rune every 3 games, sometimes it skips one, but the general rule is pretty good there)

This might be all superstition on my behalf, but seeing how the situation only requires certain seeds to explain it, rather than making a raindance on your rooftop in a tutu while you drink a cold coffee, I'd say it's a possibility.