[Transparency request] Publish the RNG engine of the lottery system

[MiaShino]

I feel this is underappreciated

Would also be equally simple to start rng function at +1 but what can we expect from company who had auto fifteen minute disconnect code baked into their client during expansion launch?

[Shibi]

...The tl;dr form of this near-essay-length post: the failure could be in something entirely different than the actual RNG portion, and based on my own past game industry experience, there's a fairly reasonable chance it was. So even if they did publish the RNG code, that wouldn't do anything to show where the issue was.

Yes it could, it could even be something like SE not thinking what rolling a random on all these plots at the same time and then trying to write the number into the database.... causing a lock and failure to update.

The problem could surface from many faults.

Releasing for Good Friday was a bit silly though.

[MsQi]

INT(RND(1)*X)

[Iscah]

The lottery was not fair and was rigged.
Assuming there are 4 bidders, each one should have 25% chance of winning.

So when the lottery system rolls from 0 to 100 a 24, the first ticket should have won. However, because there was a 5th "ghost" person, there is now 20% chance of winning for each person.
That means now the winning range of the first ticket is reduced from being 0 to 25 to being 0 to 20.

Yes. The four people still have an equal chance of being the one whose number is called. That is still fair odds for each entrant, despite the accidental fail state added.

"Rigged" means that the results are being manipulated to guarantee a particular outcome, which is not what is happening in the example you are giving. It is still fully random and nobody has control over what number is produced to determine the winner.

Henceforth, with a roll at 24, where the first ticket should have won, now the 2nd ticket has won instead.

This is a scam and this is not fair, period.

This is randomness. It is simply one more factor of randomness as to which person won.

The roulette wheel got spun slightly harder or softer, and the difference in momentum changes the winner.

The person holding the box of paper raffle tickets shook it up-and-down instead of left-to-right. Maybe they spun the box around. Someone else's ticket could have been picked if it was facing the other way! Or if the person picking the winner moved their hand an inch to the right!

Random is not perfectly random, and all random chance draws are going to be shifted around by the factors that produce the numbers. But what matters is that all entires have an equal chance and nobody can predict (or alter) and take advantage of which random number the system is going to spit out.

A scam requires malice. It's not a scam unless they then shrug and say "too bad, you lost the raffle and the system is working as intended" – which is not the case here. They're looking to fix what went wrong.

Even if it's a more complicated programming error like Packetdancer is suggesting (thankyou for that explanation; it does sound plausible!) it's still not a scam. It is a glitch. An error. A mistake. They are not trying to rob you of your gil or manipulate the results. At most, a second lottery needs to be run to determine a winner for the plot.

The error only came to light because the lotteries stopped at zero instead of trying to reroll. If they had kept retrying until they produced a valid ticket, and every lottery now had a winner, would you need to see the lottery programming to confirm it was fair? If so, why was nobody asking a week ago to see the code? We just assumed it was going to be random (or pseudorandom) and if it hadn't generated these odd results, we never would have known a difference as long as there was a winner.

[Lieri]

Its probably just a code bug and the rng is your usual go to rng that you find in C, Java, etc. Either they mistakenly added 0 to the pool eg 5 lotto tickets numbers + 0 so that this zero can win. Or they mistakenly reduced 1 from the lotto winning ticket after rng, so that means every plot that only has 1 ticket automatically loses as 1-1=0, it could either be an UI bug where the winner cannot claim the plot or it actually happened server side so there are actualy 0 winners in the database.

Either way the effect of it is a convoluted mess that I doubt they can find a fair solution for everybody. Some people will be left disappointed.

[Oglaf]

No need, I can leak the code:

Random rnd = new Random();
int winnerTicket = rnd.Next(1, numOfParticipants);
str characterName = this.getCharName(winnerTicket);
housing.setProperty(characterName, server, houseId);
sysMailer.sendMail(characterName, "GG, u have won");

You welcome.

[Skivvy]

The lottery was not fair and was rigged.
Assuming there are 4 bidders, each one should have 25% chance of winning.

So when the lottery system rolls from 0 to 100 a 24, the first ticket should have won. However, because there was a 5th "ghost" person, there is now 20% chance of winning for each person.
That means now the winning range of the first ticket is reduced from being 0 to 25 to being 0 to 20.

Henceforth, with a roll at 24, where the first ticket should have won, now the 2nd ticket has won instead.

This is a scam and this is not fair, period.

Regardless of the number of bidders..why would you have the system picking a number between 0-100? Why not go with a more straight forward 'if there are 19 bidders, pick a number between 1 & 19?'.

We have no idea how they've programmed any of this, so to call it a scam is a stretch. A massive SNAFU? Sure! But a scam....?

[dutiona]

... snip ...

Excuse-me but I gave an example where the result was manipulated to make another person win instead of the real winner, demonstrating in particular why adding another ghost participant rig the lottery.

You still consider it fair and not rigged, fine, that's your opinion.

The scam part is the prejudice of all the entrant of the lottery having their wining range reduced, significantly, sometimes even up to 50% of their winning range disappearing.

I disagree when you say it is not a scam and is just an error, a mistake. They are advertising and selling a feature in the game for real money (subscription fees). And when you want to redeem this feature and actually try it out, you get kicked by the system because of a programming error? That is not even audible.

Yes I would not have asked to see their code if it worked properly first try but hey, that how life work. You give people their chances untill they screw up. At that point you hold them accountable for their screw up and ask to see what happened so that it never happens again.

Regardless of the number of bidders..why would you have the system picking a number between 0-100? Why not go with a more straight forward 'if there are 19 bidders, pick a number between 1 & 19?'.

We have no idea how they've programmed any of this, so to call it a scam is a stretch. A massive SNAFU? Sure! But a scam....?

That's because RNG engines basically all roll a floating point number between 0 and 1. Then the developers turns this number into something useful for them. Meaning they will do maths with the number. And if the maths is wrong, the fairness is affected. For instance let us draw a number between 0 and 1.
Then we multiply it by 100 (because maybe there cannot be more than 100 bidders on a plot? Just for the example)
Then you want to adjust it to the number of bidders : there are 45 bidders, you just mod the number by 45?
That means that people with ticket 1 to 10 are more likely to have their ticket drawn (3 chances/100) whereas all the other one only have 2 chances/100 to have their ticket drawn.
Correctly rolling rng numbers fairly is not easy and is an ongoing research area.

I don't know how it's coded maybe they did ceil(rng_number * nb_bidders), maybe something else entirely: I don't care. What I care is whether their system is fair.

[Iscah]

Excuse-me but I gave an example where the result was manipulated to make another person win instead of the real winner, demonstrating in particular why adding another ghost participant rig the lottery.

You gave an example where what you're calling "manipulation" isn't manipulation at all. It's a random shifting of the range prior to the drawing of results. The thing that matters is that each ticket's range is equal.

Manipulation means that someone is meddling with either the process or results to ensure that one particular person wins. Narrowing the range cannot affect that. It might affect which person randomly gets it, but that's no more rigged than the random number itself coming out slightly different.

Think of a raffle with paper tickets. Five tickets in the drawing box, each with a theoretically even chance of being picked out. A sixth blank ticket is added. This introduces the possibility of the blank ticket being picked, and slightly reduces the chance of any one of the other tickets being picked, but does not improve or reduce the chance of which of the five genuine tickets gets picked.

The scenario you are talking about is simply the digital equivalent of this raffle. Each ticket's number range got shifted along a bit. Maybe they got moved out of the winning position, maybe they got moved into it, but it's still a random lottery and everyone had an equal chance of that happening to them.

A rigged raffle, on the other hand, would mean someone conspiring to ensure a specific ticket is drawn out of the box somehow – or in the digital version, ensuring a specific output that corresponds to one particular person's ticket. That isn't the case in your example.

The scam part is the prejudice of all the entrant of the lottery having their wining range reduced, significantly, sometimes even up to 50% of their winning range disappearing.

I disagree when you say it is not a scam and is just an error, a mistake. They are advertising and selling a feature in the game for real money (subscription fees). And when you want to redeem this feature and actually try it out, you get kicked by the system because of a programming error? That is not even audible.

You're talking as if this failure is permanent – as if nobody will ever get that house now because the zero cane up. That is simply not the case.

All it takes to resolve this is to draw another number, this time with the fail chance removed. Someone will get the house.