Just to add some fuel...

On top of losing auctions as the only bidder, not being able to get your refund, and the completely unfair lack of moderation on bidding (You can bid until you run out of money, although you can only own one house, effectively having the ability to win multiple houses in one week and cut however many other people out of a week's worth of waiting. There is no 'runner up' so those lots sit empty until the next bidding cycle, a freaking week later.), there's something I haven't seen anyone mention yet.

Every single person who won this week's lottery was NOT the actual winner. How to explain in layman's terms...

When you generate a random number in code, it looks something like random(100), meaning a random number between 0 - 99, NOT 1 - 100. Computers start at zero by default in almost every language. The problem we ran into is a rookie mistake that a first year student wouldn't even make after week 3. To compensate, you return the number generated PLUS 1. Somebody forgot the plus 1. It's really that simple. Yes, maybe other mistakes were made in other parts of the system as well, but there wouldn't be so many number 0 winners had that very simple part of the code been correct. Even if the number was written to a database before the award code read it, I highly doubt anyone would've accidentally subtracted a 1 before writing or after reading the database. The fact remains we got zeros and the numbers started at 1. Anyway...

The meaning behind this is that every single person who won this week was not the actual winner, it was the next person in line.

Let that sink in, and know they will do NOTHING about it. Every single one of those houses should be refunded and put back up for sale.

Have a nice day.

You do realize that technical error or not, taking something away from someone due to a technical error that they could not influence will just substitute one PR nightmare for another, of equal if not worse magnitude.

You assume it's the random selection that's at fault and not something like the database writes failing due to some lock or race condition, or any number of other myriad edge cases to stop the correctly chosen number being put into the database.

Quote:

---

Originally Posted by GomezAddams

When you generate a random number in code, it looks something like random(100), meaning a random number between 0 - 99, NOT 1 - 100. Computers start at zero by default in almost every language. The problem we ran into is a rookie mistake that a first year student wouldn't even make after week 3. To compensate, you return the number generated PLUS 1. Somebody forgot the plus 1. It's really that simple.

---

Not really. You could get also overflows if you add 1 to the generated number. And there are programming languages like C++ where overflows have disastrous effects.


Cheers

Quote:

---

Originally Posted by Kaurhz

You do realize that technical error or not, taking something away from someone due to a technical error that they could not influence will just substitute one PR nightmare for another, of equal if not worse magnitude.

---

Nope, people don't understand this. It is what it is. Even if they take away houses and give them to the correct winners, it is far to late for that as people have already relocated houses and bought furniture. You will get what you get when they announce a fix which will likely be your gil back at minimal.

This is very likely not how the bug actually occurred, and even if it was "redistributing" the land to the "real" winner is not the solution. I broke it down in this post: https://forum.square-enix.com/ffxiv/...=1#post5906614

Quote:

---

Originally Posted by IdowhatIwant

Nope, people don't understand this. It is what it is. Even if they take away houses and give them to the correct winners, it is far to late for that as people have already relocated houses and bought furniture. You will get what you get when they announce a fix which will likely be your gil back at minimal.

---

Anyone who has dealt with a bad storm will know it. At the very least, the way I view it. The moment the results were released it was too late. People spent the Gil, they purchased the housing permit, and have on top of that already got around to furnishing as per your post, so it isn't just a simple nature of "Just take it away". Personally, I got the house I wanted and already spent upwards of 12M decorating it. Good luck with rolling that back, especially when many people are in the same boat.

At the very worst I estimate those who were subjected to this error will retain the Gil, and be allowed for a reroll, albeit with more competition. Possibly with an increased chance of winning if it is on the same plot.

We all had equal chance for the number to roll 0 or our ticket. Having one extra 'ticket' makes no difference. It rolled what it rolled, and there is no guarantee what COULD have happened, as if it were fate that it roll one way and not the other, get over it. You lost. We were all in this lottery under the SAME circumstances and the people who won, won.

It's so beyond petty looking to take houses away from players that managed to avoid a bug and get the joy of winning. You deserve the loss with such a poopy attitude, you and every other saltminer that demands to take away things from others. There are hundreds of houses left open on my server, 7 plots were taken on my ward.. 7. Hardly upping competition for remaining spots to an astronomical degree now is it?

Quote:

---

Originally Posted by GomezAddams

When you generate a random number in code, it looks something like random(100), meaning a random number between 0 - 99, NOT 1 - 100. Computers start at zero by default in almost every language. The problem we ran into is a rookie mistake that a first year student wouldn't even make after week 3. To compensate, you return the number generated PLUS 1. Somebody forgot the plus 1. It's really that simple.

---

This has been pointed out a few times. It is only one possibility. They are unlikely to reveal the actual technical details. They will fix the issue and say "good luck next time" and carry on as if the housing system isn't utter crap.

OPs explanation cant account for the situation where there was only 1 bidder AND that person won.
as by his example that wouldn't be possible.

no one knows the true nature of what caused this, its all speculation at this point.

but as mentioned above its too late to take back from those who won.

its a horrible mess to be in and i don't envy them the difficulties they'll face trying to come to a resolution to this that feels "fair and just" to all parties.

Recent SE Japanese post has a lot of details. They are clearly stating it's not a fault in the random function.

They also suggest that for the Zero-Win bug (there are a few bugs) the winning number is in the database and the plot is actually allocated to the winner, it's just not displayed on the client. So, those winners are locked in, and existing winners were chosen correctly. Zero is client side only and doesn't reflect the lottery results or the state of the RNG.

Quote:

---

the winning number is displayed as 0 even though the internal lottery is correctly drawn and a winner exists.

---

https://pastebin.com/jE8icAJZ

Let's put it's a fault in RNG to rest now.

It amazes me to see how everyone suddenly became a programmer in light of this bug.

As other people have said, it's not just a matter of counting from 0 or 1, because there were occurrences of the last ticket winning. This wouldn't be possible if the issue was counting from array[0] instead of array[0+1]. Assuming that was even the way it was coded.

Even so, this would have easily been pinpointed while testing, which I'm sure they did because they had to make sure money was given back to losers.

No one knows for sure what happened, but it has been said that some of these losing bets were placed before the maintenance. Were somehow these tickets lost? Maybe the number of participants was stored separately as an incremental value? All of this is speculation until a response is provided. But you can't say winning tickets were not all valid given the information we have either.

Quote:

---

Originally Posted by Shibi

Recent SE Japanese post has a lot of details. They are clearly stating it's not a fault in the random function.

They also suggest that for the Zero-Win bug (there are a few bugs) the winning number is in the database and the plot is actually allocated to the winner, it's just not displayed on the client. So, those winners are locked in, and existing winners were chosen correctly. Zero is client side only and doesn't reflect the lottery results or the state of the RNG.


https://pastebin.com/jE8icAJZ

Let's put it's a fault in RNG to rest now.

---

Yeah it was definitely not a random function bug or array count, this would have been caught easily. But it's nice to finally see some light. Thanks.

Quote:

---

Originally Posted by GomezAddams

On top of losing auctions as the only bidder, not being able to get your refund, and the completely unfair lack of moderation on bidding (You can bid until you run out of money, although you can only own one house, effectively having the ability to win multiple houses in one week and cut however many other people out of a week's worth of waiting. There is no 'runner up' so those lots sit empty until the next bidding cycle, a freaking week later.), there's something I haven't seen anyone mention yet.

---

Wait what? I was only able to bid on one house during the lotto period. Am I missing something?

Quote:

---

Originally Posted by GomezAddams

On top of losing auctions as the only bidder, not being able to get your refund, and the completely unfair lack of moderation on bidding (You can bid until you run out of money, although you can only own one house, effectively having the ability to win multiple houses in one week and cut however many other people out of a week's worth of waiting. There is no 'runner up' so those lots sit empty until the next bidding cycle, a freaking week later.), there's something I haven't seen anyone mention yet.

Every single person who won this week's lottery was NOT the actual winner. How to explain in layman's terms...

When you generate a random number in code, it looks something like random(100), meaning a random number between 0 - 99, NOT 1 - 100. Computers start at zero by default in almost every language. The problem we ran into is a rookie mistake that a first year student wouldn't even make after week 3. To compensate, you return the number generated PLUS 1. Somebody forgot the plus 1. It's really that simple. Yes, maybe other mistakes were made in other parts of the system as well, but there wouldn't be so many number 0 winners had that very simple part of the code been correct. Even if the number was written to a database before the award code read it, I highly doubt anyone would've accidentally subtracted a 1 before writing or after reading the database. The fact remains we got zeros and the numbers started at 1. Anyway...

The meaning behind this is that every single person who won this week was not the actual winner, it was the next person in line.

Let that sink in, and know they will do NOTHING about it. Every single one of those houses should be refunded and put back up for sale.

Have a nice day.

---

https://i.kym-cdn.com/photos/images/...91/035/135.png

You could code an array that includes numbers from 1 to 100, then use a random number picker that grabs one from said array without having to pick one and summing plus one.
This is Japanese programming we're talking about, not Western pseudo-coders.

Quote:

---

Originally Posted by AluminumSiren

Wait what? I was only able to bid on one house during the lotto period. Am I missing something?

---

FCs could have multiple members bid on one, or many plots. Or you could bid with alts.

Quote:

---

Originally Posted by rance-sama

https://i.kym-cdn.com/photos/images/...91/035/135.png

You could code an array that includes numbers from 1 to 100, then use a random number picker that grabs one from said array without having to pick one and summing plus one.
This is Japanese programming we're talking about, not Western pseudo-coders.

---

Was that png necessary to make your point?