Some adjustments to the loot system?

[Airget]

If my math isn't wrong even with all your VODS the odds of rolling the same number twice is that low since you have 1/100, so if each person can equally roll those 100 numbers that means there is a 1% chance to roll any number. In order to have a substantial test size you have to first understand the odds of the same number being rolled twice, aka the odds of the same 1% being hit twice.

The basic equation is 1/100 * 1/100 which equals to 1/10,000 that means there is a 1 in 10,000 chance to roll the same number or a .01% chance that when you roll need on an item and get a 74, another person getting that same number is .01% chance.

And here's the thing to consider. If rolling the same number wasn't possible, then logically speaking people who roll later have a higher percentage of getting a better number.

[Bourne_Endeavor]

In the past, I've had to do normal so many times

This is precise why they use a RNG based system. You've added to the queue for other players, thus keeping Omega Normal active. A token system means people will immediately stop queuing once they have what they want. I mean, think about it. Whenever you've lucky enough to get your pedal or bolt from V1N after only 1-3 tries. Do you go back or wait until the following week? If you want guaranteed gear, you have Lakshmi and crafted stuff as an alternative.

[Dzian]

If my math isn't wrong even with all your VODS the odds of rolling the same number twice is that low since you have 1/100, so if each person can equally roll those 100 numbers that means there is a 1% chance to roll any number. In order to have a substantial test size you have to first understand the odds of the same number being rolled twice, aka the odds of the same 1% being hit twice.

The basic equation is 1/100 * 1/100 which equals to 1/10,000 that means there is a 1 in 10,000 chance to roll the same number or a .01% chance that when you roll need on an item and get a 74, another person getting that same number is .01% chance.

You seem pretty bad at math
How can you have a 1/10,000 chance when there are only 99 possible numbers or outcomes?

It really doesn't matter what the first guy rolls. The second guy has a 1/99 chance of rolling the same number.

If the first guy rolled a 74. The second guy has a 1/99chance of rolling a 74.

If the first guy rolled a 63 the second guy has a 1/99 chance of rolling a 63.

What the first person rolls doesn't matter. There's always a 1/99chance the second roll will be the same.

Or a 2/99chance thenthird roll will match any of the previous 2 rolls regardless of what numbers those rolls were.

%chance = ((N-1)/99)×100. Where n is the number of people rolling.

What the rolls actually are doesn't matter the chance of any 2 being the same is whatever you get from that equation depending on the total number of rolls

2 people = 1.011% of both being the same
4 people = 3.033% of any 2 being the same
8 people = 7.077% any 2 will roll the same number

[Phelios]

If you roll on the loot and then leave the dungeon afterward, does that mean you forfeit on the roll or will it still count after you leave?

[Pells]

A) Horribly complicated algorithm to create an entirely new version of how rolls take place after each player (who may be trying to roll at the same time, so even defining "after each" would eliminate normal threading).

B) Trivially simple random number generator roll, then an extra tie-breaker decision if necessary. And which apparently the devs themselves have said they use.

I think you're wrong on where the burden of proof would lie. Assume the one that actually makes sense over the one that doesn't unless there's clear proof to the contrary. And what proof does anyone have? That only a few people have noticed it happening? With 100 possibilities, there usually wouldn't be a tie anyway. (Ok, maybe a bit more often for 24 man content, but there how often do people actually check all 24 rolls for duplicates?)

It's B, but we'll never see it. The first person to roll keeps their number. The tie breaker roll is made when the second (tying) roll is made, and it will keep rerolling until it gets a different number. We aren't given a number until it's different. This isn't a horribly complex algorithm.

Further, tied rolls would be incredibly common if they were actually possible. Far more common than getting NQ on a 99% quality craft, and people see those often enough it's become a trope.

If you roll on the loot and then leave the dungeon afterward, does that mean you forfeit on the roll or will it still count after you leave?

If you leave while a roll is in progress, you automatically pass on everything. You would forfeit any loot.

[ChandraNoctis]

Here's why I think there is an apparently reasonable chance of a double roll (about 6% in a four man) but you rarely see it:

Since you don't see the numbers rolled by others until they have been resolved (and if a duplicate the re-roll has already been done) you will know there match if you are one of the players that rolled a duplicate number.

Even then you will only detect a match has occurred if you roll a number, remember what you rolled (or check the log) and notice it is different than your final loot roll result.

I think have seen this - where I have rolled a 99 and then my final roll was shown as 30 something and someone else had a 99 but I'm not 100% sure. I will have to watch for it in the next few instances I run.


If this is the case there is also no benefit to going early or late in the roll - due to latency, hacking concerns etc the rolls appear not to be resolved until all are registered. That's why you don't see anyone else's rolls until they are all in. So there is no real order in the rolls, they are all independent and resolved as a group (with a probably re-roll and assigning a new number to the loser on the event of a match - I'm not sure if it changes both of the matched rolls or not)

[DeaconMoore]

Nah, they don't need to change the timer. They need to upgrade their servers so you can roll and leave and if your roll wins it gets mailed to you.

It's the leave and forfeit that's the problem.

[NolLacnala]

snip

You can already tell who tried to wait out the timer because they show up as the first name in the rolls.

[Aylis]

I'd be completely okay with a timer reduction to 1 minute. That's plenty of time for anyone to decide if they want something or not. It shouldn't take anyone 5 minutes to decide. I wouldn't mind either if it forced a pop up either instead of having to click on the tab to bring it up.