• 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

DeletedUser31308

The conditional probabilities in the example are changed by design but the unconditional probabilities don't change. I don't know what you're argument is.
Using your example, if I knew that I'd have a 75% chance of winning again if I won the prize I wanted and a 75% chance to lose again if I failed to get the prize I wanted, I would only open the chest more than once if I manage to win the first time and I'd stop at my first loss. This is because I'm paying the same amount for each attempt, and some attempts will give me the EV of 0.75 of what I want while the other attempts only give me an EV of 0.25 of what I want. I know your example isn't one you'd actually suggest to be implemented, but my point is as I said: as soon as you make successive trials dependent on one another, it opens the door to abuse.

Phrased another way: As soon as the absolute probabilities of individual trials differs, as in your example of sometimes having a 75% to win and other times having a 25% chance to win, there is room for abuse. If a pattern exists, and players figure out what it is based on data analysis, the players will abuse the pattern to avoid the trials with the worse probabilities.
 

DeletedUser13838

Using your example, if I knew that I'd have a 75% chance of winning again if I won the prize I wanted and a 75% chance to lose again if I failed to get the prize I wanted, I would only open the chest more than once if I manage to win the first time and I'd stop at my first loss. This is because I'm paying the same amount for each attempt, and some attempts will give me the EV of 0.75 of what I want while the other attempts only give me an EV of 0.25 of what I want. I know your example isn't one you'd actually suggest to be implemented, but my point is as I said: as soon as you make successive trials dependent on one another, it opens the door to abuse.

Phrased another way: As soon as the absolute probabilities of individual trials differs, as in your example of sometimes having a 75% to win and other times having a 25% chance to win, there is room for abuse. If a pattern exists, and players figure out what it is based on data analysis, the players will abuse the pattern to avoid the trials with the worse probabilities.
I disagree with your concern about abuse. If you have 1000 florins, tickets, or whatever, you're going to use them. I don't think you're going to hold on to them because the next chest or whatever has a smaller percentage. And there is no pattern. The results are still random.
 

DeletedUser3882

Congratulations! @Algona The argument over squiggly line placement in a line of code ensues to page 5!

I’m focused on creating an actual predictable pattern ( I skipped p.4 because “are you serious?!?” and did read 5 because “only two!” :rolleyes: ) soooooo...

See you again on 7, I reckon... If mathematical laws of predictability and rational thought hold serve..
 

DeletedUser31308

I disagree with your concern about abuse. If you have 1000 florins, tickets, or whatever, you're going to use them. I don't think you're going to hold on to them because the next chest or whatever has a smaller percentage. And there is no pattern. The results are still random.
I agree that most people would "still use them", because most people are either idiots or don't care enough to optimize their play. Those, like me, that are practically addicted to optimizing would not "still use them", and would do better on average because of it. That is exactly why I made the point that a system like this could cause more players to think the system is rigged. If they realize that a small handful of players have figured out how to effectively game the system, they will be angry that the system is game-able in the first place. Could lead to a lot of people quitting out of frustration.
 

DeletedUser13838

How is wasting tickets by not using them considered sub-optimal? Your probability of success drops from something positive to 0 which is as negative EV you can get. I don't think you understand what I'm saying if you think you are gaming the system.
 

DeletedUser31308

How is wasting tickets by not using them considered sub-optimal? Your probability of success drops from something positive to 0 which is as negative EV you can get. I don't think you understand what I'm saying if you think you are gaming the system.
I am more focused on the chest drop randomness, so any unspent curreny would be florins which do not have a cap like tickets. In the tickets case, you may be forced to take suboptimal individual probabilties but you'd still be able to game the system in terms of choosing not to pay to double your chances when your chances are worse in the first place, or having better information to choose the highest EV game at a given moment.
 
random ? how about more upgrades in the rewards and what player needs 3 level 1. Upgrades should be offered at least every 2nd or 3rd change of rewards and not at all for i think maybe 6 changes.
 

DeletedUser31882

Personal Anecdotal evidence/Data: I've had a few crazy streaks on the 15% carnival game. 3 of my secondary cities won 3-4 of the 15% game in a row. This also allowed many of my cities to purchase a chest with a chance of a WW. 2 of my cities ended up snagging a new WW.

Based on this data, I've come to the conclusion that we would need to start data collection threads if we really wanted to have a chance to catch Inno using a broken randomizer. When I played FFBE (A Gocha type game), some ambitious people on reddit did this for events & banners. I wonder if people on beta do something similar. *shrugs*
 

DeletedUser13838

I am more focused on the chest drop randomness, so any unspent curreny would be florins which do not have a cap like tickets. In the tickets case, you may be forced to take suboptimal individual probabilties but you'd still be able to game the system in terms of choosing not to pay to double your chances when your chances are worse in the first place, or having better information to choose the highest EV game at a given moment.
I'm not sure I was suggesting a change to chest drops (sounds weird I know). The problem with them is that sometimes there is a uniquely good prize with a bunch of goats as consolation (eg SoKs) and sometimes there isn't. You already have the ability to hold on to your resources and choose which chest to go after. Kind of like waiting until powerball hits $500 million before buying lottery tickets.

Speaking of Monte Hall, how about a Monte Hall mechanic where you make a selection (say 1 out of 5) , Monte opens a couple of chests (say 1-3 at random) and you can choose (pay extra?) to switch. OK I'm not sure I'm being serious here.
 

DeletedUser30900

Why would you blame your bad luck on Inno’s “broken randomizer” you didn’t even do it enough time to prove your theory (yes, I’m talking about over 1000times) Stop thinking like you are one of the main character of this world so the game should go wherever you want that :p
 

DeletedUser31308

I'm not sure I was suggesting a change to chest drops (sounds weird I know). The problem with them is that sometimes there is a uniquely good prize with a bunch of goats as consolation (eg SoKs) and sometimes there isn't. You already have the ability to hold on to your resources and choose which chest to go after. Kind of like waiting until powerball hits $500 million before buying lottery tickets.

Speaking of Monte Hall, how about a Monte Hall mechanic where you make a selection (say 1 out of 5) , Monte opens a couple of chests (say 1-3 at random) and you can choose (pay extra?) to switch. OK I'm not sure I'm being serious here.
Another classic case of two people talking, thinking they're talking about the same thing, and happen to be talking about something completely different. :p

Nothing is as impressive as the human ability to misunderstand one another. For what it's worth, Happy Friday!

Edit: I would love to see more "game theory" related mechanics in this game. Being able to pick a door, having another revealed, being able to switch, etc. Would be a lot more fun than just clicking "Open" on a chest and seeing what RNG dictates.
 

DeletedUser30312

Speaking of Monte Hall, how about a Monte Hall mechanic where you make a selection (say 1 out of 5) , Monte opens a couple of chests (say 1-3 at random) and you can choose (pay extra?) to switch. OK I'm not sure I'm being serious here.

As long as there's always a chance to win a Goat Farm. Zonk!
 

DeletedUser10720

A basic explanation of probabilities.

Equal probability: If you have a 20 sided die, and are betting that you will roll a 1, you have a 5% chance to be correct. Each time. Every time. The odds never change. Just like the RNG used for this event. You will see runs of 'hot streaks' and 'cold streaks' and at no time will your 5% chance to roll a 1 change.

Compounding probability: a game of Russian roulette. You start with a roughly 16.5% chance at finding the bullet, with the first click, it becomes a 20% chance, then 25%, 33%, 50%, 100%.

People often confuse equal probability scenarios with compounding probability scenarios. Each time you lose the game you add that previous 5% loss to the current 5% chance and believe you've secretly for a 10% chance. So by the time your 11th ticket comes up, those are 55% odds in your favor! When in reality the die has not changed.

This mentality is called confirmation bias. It is what drives many things such as the gambling industry, arcades, and yes this game.
 

DeletedUser34480

A basic explanation of probabilities.
...
... and you noticed that about 5-10 people have already laid it out here, in this thread.

Out of curiosity, what do you think of my examples?
They serve as demonstration of my position that if a high chance of event declared, any deviation of 100%+ are unacceptable.
 

DeletedUser10720

... and you noticed that about 5-10 people have already laid it out here, in this thread.

Out of curiosity, what do you think of my examples?
They serve as demonstration of my position that if a high chance of event declared, any deviation of 100%+ are unacceptable.

Your sample size is too small. I run 7 world's total. In many I have seen my 11 tickets yield nothing on the 15% chance game.

But I have also seen, several times over, 8 or more tickets in a row win that game immidiately.
 

DeletedUser34480

Your sample size is too small. I run 7 world's total. In many I have seen my 11 tickets yield nothing on the 15% chance game.

But I have also seen, several times over, 8 or more tickets in a row win that game immidiately.

Not arguing that.
My point is, 15to30% chance is too high to let deviation swing 100%+.
High probability comes with an expectation range. It is not a roulette.
 

DeletedUser10720

It is exactly that. The roulette wheel at a casino. A random number picked by a mechanism that has no increasing ratio to odds.

The deviation is not at 100% as you claimed. If you have spent every ticket and NEVER won once, then ir would be. But you're looking at runs of 10-11 and internally compounding your expected payout.

When the casino has the display above the roulette wheel, telling you the last 20 or so numbers drawn. Those numbers mean nothing. They are there to play on this exact logical error your brain is hardwired to see. Just because 25 came up 3x in the last 20 spins, does not mean that it has an increased chance of coming up again, nor does the lack of a 17 mean that 17 is "due to come up soon" each of the numbers always has the exact same odds as it had before.
 

DeletedUser34480

I know all that.
My point is, if something includes probability, it doesn't always mean it is free to deviate at will.

Anyway, I rest this case.
I made my point clear, it is up to anyone to agree or disagree.
To math fans, sorry, but your disagreement doesn't mean your are right. Amount of pictures and offensive jokes doesn't improve your position a bit.
 

DeletedUser

My point is, if something includes probability, it doesn't always mean it is free to deviate at will.
Ummm, a Random Number Generator doesn't have a will, any more than a pair of dice or a roulette wheel.
 
Top