I take issue with randomness of pulls, please prove me wrong

dr tinykittylove

simonsez wrote:

Also, your comments make me think that some people are misunderstanding the "token pulls aren't random" argument. No one is saying they've constructed a fake-RNG that spits out biased outcomes. What they're doing is using a RNG against a table of probabilities that are skewed towards a rotating set of characters.

Someone please let me know when iceman enters this skewed rotation. I could use some covers

Bowgentle

dr tinykittylove wrote:

simonsez wrote:

Also, your comments make me think that some people are misunderstanding the "token pulls aren't random" argument. No one is saying they've constructed a fake-RNG that spits out biased outcomes. What they're doing is using a RNG against a table of probabilities that are skewed towards a rotating set of characters.

Someone please let me know when iceman enters this skewed rotation. I could use some covers

Currently it's Red Hulk week. I pulled one of each of his colours during the last 5 days or so.
That's 3 RHulks from 5 or so tokens.

NighteyesGrisu

simonsez wrote:

I agree, I don't care how it feels to people. That's why every time someone happens to post the results of a bucket of token pulls, I'm posting the probability of getting whatever sort of clustering their results are showing. And despite consistently unlikely outcomes, most part people still prefer to believe in the RNG fairy.

That's the point there are not conistently unlikely outcomes....there are inconsitently unlikely outcomes...

NighteyesGrisu wrote:

That's not what they're doing, no one said they are, so keep your strawman out of this.

Then what are the guys claiming the token pulls must be not random bacause they got 10 moonstones in a run saying? Either it's random or if it isn't they must have made some change that favours getting moonstones?

And it hardly has to be an elaborate algorithm. I would have to assume that all the probabilities are stored in a table that the RNG hits up against. They just need to change the values in that table every few days.

And why would they do this when they explicitly list the probabilities in the token sets? Why not just list it accurately? What you're suggestion they're doing here is nothing short of fraud.

Also, your comments make me think that some people are misunderstanding the "token pulls aren't random" argument. No one is saying they've constructed a fake-RNG that spits out biased outcomes. What they're doing is using a RNG against a table of probabilities that are skewed towards a rotating set of characters.

Maybe you personally aren't saying that, but claiming that "no one is saying they've contructed a fake-RNG that spits out biased outcomes" is clearly wrong, even if we're just limiting it to statements in this very thread.

simonsez

NighteyesGrisu wrote:

And why would they do this when they explicitly list the probabilities in the token sets? Why not just list it accurately? What you're suggestion they're doing here is nothing short of fraud.

They can just claim those are long-term averages, and I'm sure their long-term numbers would probably support that.

Dayv

simonsez wrote:

Bishop wrote:

I know **** is in the title but **** needs to stay thier.
I have all these colors already and this isn't the first time this has happened. Could use IW green or blue,haven't seen them. Could use PX purple haven't seen it,could use many more but I'm confidant I'll get these. I think there's better odds in Vegas. My last pulls..only the 5* was useful.

I could point out that there's less than an 8% chance of getting 2 or more pairs out of 6 pulls, but this is getting repetitive, and I'm getting the feeling that no one who doesn't want to be convinced that pulls are clumped instead of random, is going to be convinced, no matter how many of these there are.

You know who doesn't post their token pull history to the forum? The vast majority of players with no particularly interesting patterns, good or bad, showing up in their draws. You can't draw any conclusions on the randomness of a system from self-selected reports. You simply can't.

Dayv

simonsez wrote:

NighteyesGrisu wrote:

And why would they do this when they explicitly list the probabilities in the token sets? Why not just list it accurately? What you're suggestion they're doing here is nothing short of fraud.

They can just claim those are long-term averages, and I'm sure their long-term numbers would probably support that.

There's a fairly easy way to test this. Get an account sandboxed, generate unlimited funds, and open enough tokens to have a viable statistical sample.

Of course, if you do that, and it shows a reasonable random distribution, you could always say that there are different odds for sandboxed accounts. At a certain point, you simply can't argue with religious belief.

simonsez

DayvBang wrote:

At a certain point, you simply can't argue with religious belief.

It's **** hilarious that you're referring to probability calculations as "religion".

And I'll say it again, since this is another familiar complaint from believers in the RNG fairy: most people aren't posting because they see "interesting patterns" in their pulls. Most are just complaining they're not getting anything useful. Getting/not getting the iceman or cykes one is looking for is unrelated to the distribution of the pulls. So no, these aren't cherry-picked observations of people specifically posting pulls with patterns in it.

simonsez

DayvBang wrote:

There's a fairly easy way to test this. Get an account sandboxed, generate unlimited funds, and open enough tokens to have a viable statistical sample.

That'd be a waste of my time. We already have a dev telling us that someone got four 5*s via 80 taco tokens. No study anyone does on a sandboxed account will ever generate an outcome as absurdly statistically impossible as this. And yet I'm still having this **** argument.

Malcrof

simonsez wrote:

DayvBang wrote:

At a certain point, you simply can't argue with religious belief.

It's tinykitty hilarious that you're referring to probability calculations as "religion".

Nothing to do with anything, but ever met a mathematician? I think the analogy is fairly accurate.....

What we would need, is for one of the devs to provide a nice raw data sheet.. no names, just legendary token pulls and what was pulled. Of course conspiracy theorists would say the data was altered, but i doubt they would do that.. game data is vital to them, to understand the goings on, and this is the only way to get a true look at it. The thousands of players who don't use the forum would complete the data we need..

Pylgrim

simonsez wrote:

DayvBang wrote:

At a certain point, you simply can't argue with religious belief.

It's tinykitty hilarious that you're referring to probability calculations as "religion".

Especially, taking in account that we have all these people trying to dismiss empirical data with entirely theoretical tenets. Who is religious, again?

Anyway, I'll keep insisting that this discussion goes nowhere. The people who present evidence to raise the concern of unlikelihood of true randomness will always be dismissed by the believers of said true randomness because of the comparatively minuscule sample size that the concerned are able to present, all the while their own statements to support true randomness are by nature, an unprovable hypothesis based on the starting assumption that there is true randomness.

Again, we could spare ourselves this tedious argument by simply making it so luck cannot screw players at this level of playing, truly randomly or otherwise. Since staunch supporters of true randomness have been so far, without a fault, people who have had enjoyed at least a moderately good luck, it's hard for them to understand that even if they are right, and we are simply being screwed by cold, apathetic chance, that fact doesn't provide any consolation, nor a way to remediate the unfair and undeniable disadvantage that it arbitrarily inflicts to a group of players.

simonsez

Malcrof wrote:

Nothing to do with anything, but ever met a mathematician? I think the analogy is fairly accurate......

No it's not. One bases their beliefs on facts and data, and the other bases them on faith in fairy tales that are thousands of years old.

But nice job in trying to shift the debate into something way, way more volatile.

The Bob The

OneLastGambit wrote:

Oh my Lord... This thread is still going?

Are you kidding? This thread sums up like 75% of the forums' content. It's nice to have it contained in one place.

Zen808

I love how this forum celebrates Thanksgiving in the most American way possible - tedious idealogical squabbles! All we need is for IceIX to make some virtual booze icons, and we'll be celebrating in style!

Punatulkku

past 2 months i have gotten 2 4* guys and 3 3* guys. Rest is 2* **** and 1* basic token stuff... Feeling really happy that game is limiting my progress...

OneLastGambit

The Bob The wrote:

OneLastGambit wrote:

Oh my Lord... This thread is still going?

Are you kidding? This thread sums up like 75% of the forums' content. It's nice to have it contained in one place.

Yes indeed, I'm surprised anyone is debating this anymore. You have one set of people tenuously grasping the one anamolous result (the quickly whaled 5*) as their smoking gun, and the other set of people using every single other result in the data set (which is quite clearly the statistical trend) refuting it. The religious analogy was not drawn to compare the discussion to religion but rather the self-supporting belief system the participant A (the people clinging to one tenuous anamolous result) has in place and thereby are only looking at the one result that supports their argument in the same way that religious people often only look at evidence which supports their belief and ignore anything to the contrary.

Nobody will ever win this argument, I've given up on it. Though I'll admit it's far more fun to read the discussion when you're not the one frustrated by it.

simonsez

OneLastGambit wrote:

You have one set of people tenuously grasping the one anamolous result

This is why I'm out. This is like talking to a bunch of 6 year olds with ADHD.

It's not a tenuous grasp. I've got a reaaal firm grip on it. It's not one smoking gun, it's several smoking guns, all related to the different ways the initial 5*s were whaled up. It's all on the forum in those older threads for anyone to back and revisit for anyone who actually gives a ****.

And then this was followed by dozens of other sets of token pulls that have done nothing to refute the initial reaction MOST people had back then that, "There's no way those pulls could've been random".

But sure, go on believing that all of this is just about "one" event that "could" have happened. I'm done punching down. I've got other things to waste my time on.

OneLastGambit

simonsez wrote:

OneLastGambit wrote:

You have one set of people tenuously grasping the one anamolous result

This is why I'm out. This is like talking to a bunch of 6 year olds with ADHD.

It's not a tenuous grasp. I've got a reaaal firm grip on it. It's not one smoking gun, it's several smoking guns, all related to the different ways the initial 5*s were whaled up. It's all on the forum in those older threads for anyone to back and revisit for anyone who actually gives a tinykitty.

And then this was followed by dozens of other sets of token pulls that have done nothing to refute the initial reaction MOST people had back then that, "There's no way those pulls could've been random".

But sure, go on believing that all of this is just about "one" event that "could" have happened. I'm done punching down. I've got other things to waste my time on.

Please don't take it as a personal insult to yourself, it wasn't meant that way at all.

All I'm saying is until there is empirical evidence (anecdotal evidence is not sufficient - no matter the source) I cannot, as a scientist, accept anything other than what is presented.

XandorXerxes

OneLastGambit wrote:

simonsez wrote:

OneLastGambit wrote:

You have one set of people tenuously grasping the one anamolous result

This is why I'm out. This is like talking to a bunch of 6 year olds with ADHD.

It's not a tenuous grasp. I've got a reaaal firm grip on it. It's not one smoking gun, it's several smoking guns, all related to the different ways the initial 5*s were whaled up. It's all on the forum in those older threads for anyone to back and revisit for anyone who actually gives a tinykitty.

And then this was followed by dozens of other sets of token pulls that have done nothing to refute the initial reaction MOST people had back then that, "There's no way those pulls could've been random".

But sure, go on believing that all of this is just about "one" event that "could" have happened. I'm done punching down. I've got other things to waste my time on.

Please don't take it as a personal insult to yourself, it wasn't meant that way at all.

All I'm saying is until there is empirical evidence (anecdotal evidence is not sufficient - no matter the source) I cannot, as a scientist, accept anything other than what is presented.

He's showing you empirical evidence, you're taking a redname's word for it - that's the opposite of what you're going for.

If I roll a die 10,000 times and it comes up with a 1 around 80% of the time, do you shrug and say that's bad luck or do you look at the die to make sure it's not weighted or has a manufacturing defect? I entirely believe the developers when they say it's coded to be random, just as I believe dice are created to be random. That doesn't mean that there aren't defects that inhibit that randomness.

OneLastGambit

XandorXerxes wrote:

OneLastGambit wrote:

simonsez wrote:

OneLastGambit wrote:

You have one set of people tenuously grasping the one anamolous result

This is why I'm out. This is like talking to a bunch of 6 year olds with ADHD.

It's not a tenuous grasp. I've got a reaaal firm grip on it. It's not one smoking gun, it's several smoking guns, all related to the different ways the initial 5*s were whaled up. It's all on the forum in those older threads for anyone to back and revisit for anyone who actually gives a tinykitty.

And then this was followed by dozens of other sets of token pulls that have done nothing to refute the initial reaction MOST people had back then that, "There's no way those pulls could've been random".

But sure, go on believing that all of this is just about "one" event that "could" have happened. I'm done punching down. I've got other things to waste my time on.

Please don't take it as a personal insult to yourself, it wasn't meant that way at all.

All I'm saying is until there is empirical evidence (anecdotal evidence is not sufficient - no matter the source) I cannot, as a scientist, accept anything other than what is presented.

He's showing you empirical evidence, you're taking a redname's word for it - that's the opposite of what you're going for.

If I roll a die 10,000 times and it comes up with a 1 around 80% of the time, do you shrug and say that's bad luck or do you look at the die to make sure it's not weighted or has a manufacturing defect? I entirely believe the developers when they say it's coded to be random, just as I believe dice are created to be random. That doesn't mean that there aren't defects that inhibit that randomness.

Umm I'm not taking ICE's word, I'm doing the opposite. I'm saying if someone wants me to believe that a random system is not random they need to provide empirical evidence to back up their point, I would also like for that evidence to be p<0.05 too if it all possible. One (self reported and unverified) result would not be within this normally accepted and very reasonable probability margin.
I know that people saying that ICE has confirmed this, but he did not. He only confirmed that the odds of this happening by chance were very slim. He did not verify what the poster had claimed, that part of the data was self reported and unverified so it can't really be taken as accurate. How do we know that person was telling the truth? How do we know it was not a sandboxed account?
How do we know it was not hyperbole? the internet (forums in particular) are very prone to this last one.

Punatulkku

Could anyone make somekind of webpage where we can record our token drops? So that we could get some raw data. Cause for me, the token drops feel gimbed and arranged...