• We are looking for you!
    Always wanted to join our Supporting Team? We are looking for enthusiastic moderators!
    Take a look at our recruitement page for more information and how you can apply:
    Apply

Developers, please fix a randomizer for Carnival!

  • Thread starter DeletedUser34480
  • Start date

DeletedUser34480

This random doesn't work remotely close to what is declared.

I have tried more then 10 times for Upgrade from chest. Declared chances: 20%, 15%. Actual result: 0.
OK, this stinks but somehow is within reasonable deviation.

But, games are just a lost case!
Double chance on a game #2 -- time after time it takes me the whole ticket scale, on double chance. Yet, it supposed to be 30%. Is it just anywhere close?!
Just now, I've spent 25 tickets on game #3. All on the double chance. So, it supposed to be 10%. Once again, care to compare what it supposed to be and what it actually was?!

And please, don't tell me it is random, some crazy things can happen. Yes, they can. But not time after time!
Because if it does, it is simply broken!
 

DeletedUser31308

I'll give you the benefit of the doubt, since probability does seem to be a topic that thoroughly escapes most people. Declared chances can be perfectly accurate, while not giving you a guarantee that at any given moment the distribution will perfectly match those chances. If you add up the experiences of the guy who got 3 SoKs in 3 attempts with the people that tried 20 times at 10% and got 0, the distributions should be very close to the expected probabilities. The more samples you have, the more true that is. The 10 tries you've based your conclusion on is a tiny tiny sample size that is inherently vulnerable to skewed results.
 

DeletedUser34480

I'll give you the benefit of the doubt...
f you add up the experiences of the guy who got 3 SoKs in 3 attempts with the people that tried 20 times at 10% and got 0.
The 10 tries you've based your conclusion on is a tiny tiny sample size that is inherently vulnerable to skewed results.
Thank you for benefit of doubt because you seemed to miss key points.
10 tries were for chest. If you check, I said that it is bad but within the reasonable range.
Game #2, I've played multiple times. So the sample is significant.
Now, if randomizer casually swings times out average, it IS BROKEN.
 

DeletedUser31308

Thank you for benefit of doubt because you seemed to miss key points.
10 tries were for chest. If you check, I said that it is bad but within the reasonable range.
Game #2, I've played multiple times. So the sample is significant.
Now, if randomizer casually swings times out average, it IS BROKEN.
The chance to fail a 30% chance 11 times in a row (your whole bar) is 2%. Meaning 1 out of every 50 people that dump their bar on the double chance cups game will not win anything. If this has happened to you twice, you are exceptionally unlucky, but nothing is broken. There are NO guarantees with random.
 

DeletedUser34480

The thing is, you are looking at it as some "Win a car" in Vegas.
This one, is an ordinary game (at least, #2).
Treat it as something else that is ordinary. For example, a bus schedule. If it says it runs 15 minutes on average, you might be surprised by two buses within 5 minutes. Then, waiting for 2 hours, would be plain wrong and broken.
Anything beyond 100% is a failure.
 
Last edited by a moderator:

DeletedUser

The thing is, you are looking at it as some "Win a car" in Vegas.
This one, is an ordinary game (at least, #2).
Treat it as something else that is ordinary. For example, a bus schedule. If it says it runs 15 minutes on average, you might be surprised by two buses within 5 minutes. Then, waiting for 2 hours, would be plain wrong and broken.
Anything beyond 100% is a failure.
This is hilarious, and shows that you understand exactly nothing about random chance and the chests.
 

Algona

Well-Known Member
I agree the whatchamavallit thingamajig is broken. I keep winning the 5% game on 11 or less tries.

Two questions for you, Ace.

How many people play this game?

Take a few minutes and look around the forums before answering this one: Hoe many complaints do you see about the RNG?

a bus schedule. If it says it runs 15 minutes on average, you might,,,

,,,want to find another bus company.

And you know better because... ?

He's not comparing an RNG to a bus schedule.
 
Last edited:

Falconwing

Well-Known Member
Look, you are up against the most evil creation man has ever known, the RNG monster.

The RNG (Random Number Generator) can, on a good day, make you feel like a God among Gods! You're every touch turns to gold. You can do no wrong. And, on other days, it will make you fell like something you stepped in that you have to scrape off the bottom of your shoe.

Personally, the RNG has been my friend this Event, Many times I come on to spend my 11th ticket, see 400+ on the 5% chest, use one ticket and BOOM! I'm a winner. When the Architect chest was here, I bought 9 chests and won 3 upgrades and 4 SoKs. I think I've collected a total of a little over 10k florins this Event.

Sadly, in order for the RNG to treat me so well, it has to treat you, and many others, like a steaming pile of poo. Sorry.
 

DeletedUser34480

Gentlemen, the fact that you're playing this game longer than I do, somehow led you to an idea that you understand what random is (and can be made) -- way better than I understand.
If you think, those are not related. I do know a lot about random and how it is (and can be) programmed.

Example of "bus runs every X minutes on average" is clear enough.

As for chest, let me beef up stats.
I now have 1 Upgrade and 8 Canals -- while declared stat is 1 : 2.
 

DeletedUser31308

...and I got 3 SoKs in 10 tries. So "obviously" the SoK drop chance isn't 10%, it has to be 30%.

I'm not making any assumptions based on longevity playing this game. I am an engineer by education and profession, I code for a living, and have taken multiple classes on probability theory. Bus schedules are nothing like random number generators, in fact a "bus runs every X minutes on average" is not even a promise of future behavior, just an observation of past behaviors. Let me show you how something like these chests would be programmed with a simple example:

int foo = rand () % 20; // We divide a random number by 20 and store the remainder, giving us a random number between 0-19.

if(foo >= 0 && foo < 3) //the range [0,1,2] will give a bridge aka 3/20 numbers will give bridge aka 15% chance
{
//drop a bridge
}
else if (foo >= 3 && foo < 5) // the range [3,4] gives a SoK, aka 2/20 numbers gives SoK, aka 10% chance
{
//drop a SoK
}
.....etcetera

Certainly there are cleaner ways of doing this, but this will give the expected distribution given a large enough sample size. We're talking on the order of thousands of samples to be sufficient. If you're familiar with randomness in code, write your own program like the above that runs the above snippet a configurable number of times. Run it 10 times, note the distribution (hint: it will be unlikely it matches the actual probabilities). Run it 100 times, note the distribution. Then run it 100,000 times, and not that the distribution will be very close to what was expected.
 

DeletedUser34480

Bus intervals are very much a promise of expected range. If bus shows up in 20-25 minutes instead of 15, we say "unfortunate".
When we wait 30-40 minutes, twice, occurring not far apart -- then we say that declared average is not true, don't we?
Now, waiting for an hour or longer is just plain wrong?
Won't you agree.

For your sample, when you run your 1000 times, do you check of deviation?
If it produces wild swings, it should not be used.
As I've said, there hundreds of fine-tuned randomizers.
Finally, if nothing better at hand, use the last millisecond of the input as your 0-9 value.

ps. It doesn't really matter, but I've got just 2 SoK.
 

wolfhoundtoo

Well-Known Member
Bus intervals are very much a promise of expected range. If bus shows up in 20-25 minutes instead of 15, we say "unfortunate".
When we wait 30-40 minutes, twice, occurring not far apart -- then we say that declared average is not true, don't we?
Now, waiting for an hour or longer is just plain wrong?
Won't you agree.

For your sample, when you run your 1000 times, do you check of deviation?
If it produces wild swings, it should not be used.
As I've said, there hundreds of fine-tuned randomizers.
Finally, if nothing better at hand, use the last millisecond of the input as your 0-9 value.

ps. It doesn't really matter, but I've got just 2 SoK.


The fact that you are no longer complaining because you got what you wanted indicates a basic flaw in this situation but it's not with the RNG.
 

DeletedUser31308

When we wait 30-40 minutes, twice, occurring not far apart -- then we say that declared average is not true, don't we?

No, we don't, unless we have no idea what we're talking about. For every odd occurrence of 1 hr late buses is a couple 10 minute waits that were lower than average that balances the average out. There's no such thing as a "false" average with probability, just the average, which happens to be displayed on these chests.

Even if you did comprehend probability, why are you the one who gets to declare "if it produces wild swings, it should not be used"? Should not? Why not? Because some people will complain about it? Because you don't like it? I think the swings make Inno a lot of money, because all the gambling addicts in this game live for those big swings in their favor. And in order to have those swings, there have to be swings against peoples favor in order for the average to be where it is.
 

DeletedUser34480

No, we don't, unless we have no idea what we're talking about. For every odd occurrence of 1 hr late buses is a couple 10 minute waits that were lower than average that balances the average out. There's no such thing as a "false" average with probability, just the average, which happens to be displayed on these chests.
I've got annoyed by your lousy tutoring pose.
Here, you are the one who have no idea what you are talking about. Moreover, you are unable to comprehend a really easy sample.
No, if three buses come 5 minutes apart and then 3 buses come an hour apart -- you CANNOT average it out.
Just because math lets you can add and divide them, it doesn't mean at all that you can apply it in this situation.

Even if you did comprehend probability, why are you the one who gets to declare "if it produces wild swings, it should not be used"? Should not? Why not? Because some people will complain about it? Because you don't like it? I think the swings make Inno a lot of money, because all the gambling addicts in this game live for those big swings in their favor. And in order to have those swings, there have to be swings against peoples favor in order for the average to be where it is.
Oh, so now you think game is rigged on purpose to produce wild swings? This is possible but I hope it is not true.
As for "why" it should not be that way, there is a sample above. When you declare 30% chance, it should not turn time after time into 10%.

PS. Feel free to reply, I'm done with this discussion with you.
 

DeletedUser34480

The fact that you are no longer complaining because you got what you wanted indicates a basic flaw in this situation but it's not with the RNG.
This fact didn't happen.
Should have it happened, what would it really prove?
It was a rhetorical question, in case you haven't notice.
 

DeletedUser

I've got annoyed by your lousy tutoring pose.
Here, you are the one who have no idea what you are talking about. Moreover, you are unable to comprehend a really easy sample.
No, if three buses come 5 minutes apart and then 3 buses come an hour apart -- you CANNOT average it out.
Just because math lets you can add and divide them, it doesn't mean at all that you can apply it in this situation.


Oh, so now you think game is rigged on purpose to produce wild swings? This is possible but I hope it is not true.
As for "why" it should not be that way, there is a sample above. When you declare 30% chance, it should not turn time after time into 10%.

PS. Feel free to reply, I'm done with this discussion with you.
You know, with each post you prove more emphatically that you have no idea what you're talking about. Besides not making much sense, you are deliberately misinterpreting what @ThruTHEhead said about wild swings. Nowhere did he say that the game was "rigged on purpose", or even imply that. He simply said that Inno probably made a lot of money off of the fact that there are "wild swings". The fact is that any RNG will occasionally produce "wild swings" just because it is random, not because it is "rigged".
 

DeletedUser34480

I agree the whatchamavallit thingamajig is broken. I keep winning the 5% game on 11 or less tries.
So, you feel the game is rigged?
It increases chances for a "well-known" members?

Two questions for you, Ace.

How many people play this game?

Take a few minutes and look around the forums before answering this one: Hoe many complaints do you see about the RNG?
Why don't you tell me, how many active, real players?
How many of them are even registered for this Forum?
How many of those would come to complain.

Or, you can take it from the other side: how many supporters are here, compare to the overall number of players?

Face it, you don't know, if there was a true vote on this, which option would take majority of votes.
 
Top