Feedback The tale of randomness of random

Random is DSO's catch-all concept, used as a simple answer for drop complaints. The underlying mechanics are much more complex. More accurately, DSO uses a series of programmed schedules to determine what falls and when.
One of the schedules was even posted in the forums. There were 4 pieces required to workbench an item. One fell about 52.34% of the time, one fell only about 5.86% of the time, and with built-in priorities and exclusions. (Numbers subject to memory distortion, but decimals were used.)

So, "luck" may be a factor of right place/right time, but only to the extent that over-all results are determined by BigPoint. They choose the algorithms, knowing full well some people will be left swinging in the wind. Maybe they think it encourages revenue.

They can also change the math and results on a daily basis, when they want.

There probably is a model that rewards high dollar or high level players with higher tier drops. That evidence is anecdotal.

Some of the early posts hinted at a group dynamics matrix, which also influenced boss and minion parameters, which in turn effects the drops. Complex stuff for most.

We don't really need to know the details, we can see the results all too well.
I would rather have the system working for us, instead of us working against the system.

 

Yeah, I am with you and Greg speaking of probabilities like we are a bunch of kids kind of confirms that. They surely use some kind of bastardized function to work in their best interest. But with all the defects they have, especially some crazy ones like block working the opposite way, they must have problems with QA processes, because the block should have been tested on the unit/integration levels, and there should have been a manual test case for this at least, so QA/QC should have ran this test case and mark Passed in their QA tool during regression and work on the feature/user story. So with all that, there might an issue with random they are unaware of, even assuming they do some financials maximization with it.

 

Do they even have Quality Assurance protocols? I thought the department heads just pitched their stuff upstairs!

 

Small startup with a hand full of actual coders, I doubt it. But they said something about quality controls that were established in a prior twitch stream. Its hard to maintain in a small team where everyone know everyone. Not defending them, just trying to explain.

 

In business since 2002. 500M USD or so in funding in a couple of series. They are profitable. 500 or so employees. They hire ppl like Haruki. They are targeting Asia. I doubt you can call them small startup. This is mature organization, or supposed to be a mature. And they have surely passed the quick&dirty/MVP phase.

Plus they recently switched to 3 week sprints. That is usually driven by two things - 1. QA saythey can't test things in 2 week iterations as Dev keep checking in the code up until the last day, no time for regression.
2. Devs say sprint is too small to incorporate 'large' user stories. Both are illusions and will not fix the root cause, but anyway it means they do have QA and some sort of processes they call agile.

Upd. They must have above a thousand employees, offices in San Fran, which is the A-series kind of an office, after the seed period of start-up, office in Malta - financial randomization (most European IT firms love Malta), and also in Brazil, where they should have captive center with Dev/QA, or hispanic/portuguese support staff, or both. And they have decent revenue, they have been through the acquisition, major deal on holdings including 350M investment. Comcast Interactive Capital was a big shareholder till 2011.

Not sure what is the definition of small startup in Hamburg, but in San Fran it wouldn't be called one

...

One of their career opportunity says

• Your day-to-day work will take place in close collaboration with your team and make use of agile methods (TDD, Pair Programming, Peer Reviews, …)

How do they produce that many defects including some very critical assuming they really do TDD, PP, PR and other practices from XP. Unbelievable

Disclaimer, dear Dev and project management, no offence, I'm just trying to piss you off and talk to your ego, challenge to do better job

 

Quality System doesn't work if there's no specific team/group specialized in this task only.
Looks like (guessing) quality system/assurance is managed by people with other priority activities.

If some Devs, for instance, are code writers and quality controllers at the same time ... this will not work at all, but note that this mixed system is allowed in many organizations
Deprecated way to go.

 

They do have QA. They call the QA Testers I would be freaked out with such an offence out of 49 vacancies open there 14 Dev and 6 QA Testers

And, the system without QA is popular in the Bay and works pretty well, but I agree, on earlier stages of an organization and on the projects of smaller scale.

XP practices are designed to support that kind of way of development - TDD, Pair Programming etc. But most firms will fail with the TDD as they take it as just a method, practice, and not a culture. One should truly believe into TDD to make it work. It should be in your heart, it should be your habit, your only way of doing things. You cannot TDD today but not tomorrow) so yes, BP can't live without QA, and they know that.

Back to on topic. So we don't make this thread completely useless piece of junk. How do we get this processed with the Dev? How do we ask the question, and do we have well shaped question to ask already?

I would still wait for Traki input, but wouldn't the question/request look like -

'Dear Dev, please go check your code to make sure everything works as intented, do some proper testing and perhaps revise some of the values to make the cumulative chances of getting items reasonable.' I would also ask them the following. Dear Dev and Design team, when you decide on a probability, think of the cumulative chance, not a probability. Think of in how many runs do you want to make something very probable rather than just saying let's make it 0.1%.'

And finally, Dear Dev, I respect your hard work whatever I say in some threads. I know defects happen, that is fine. That's why there is a hope the subject is defected as well)

 

There is quite a lot of posts in here.

Fist the 20% chance for julov.
Some maps I feel it is 25% based on this answer.

There are 5 copies of a map (only found 4 for great Dwarven hiest) and one map has julov. When you open a new instance it has 1 of 5 copies to pull from. This will make a 20% chance and still be called random.

There is no way random can exist anywhere. I even feel nature has a cause and effect that can be forsee able if you can provide enough information.

For a drop rate % this is my understanding. Each mobs on a death has a number 1 to 1,000,000 that is regesterend as a death number. That number release to the items you get. To double a drop rate you assign 2x more death numbers to that item. I want to say the things an mmo can use for a random base are time and client's latency. There are others that can be used but I think these are the two leading values.

 

Accidents are unforeseen events, until they are analyzed, and the contributing factors discovered.
It can be the same with occurrences that appear random, until the coincidence is determined to be concurrence or causation.
Within a program, there can only be cause and effect, even if it is not by design or intent. If it were truly random, it would not be a program.
Let's try more accurate terminology, like probability, not "randomness." Otherwise, it is too much like a confidence scam.

 

A rubber chicken and mint flavored lifesavers work every time sometimes. You have to swing the chicken clockwise though, counter clockwise would just be silly. That's clockwise looking up at it not from the top down.

 

Sounds like philosophy behind their randomizer?..

 

Humor saves the mind, or stress-induced superstition taking over, inquiring minds want to know.

 

Stress induced humor, maybe, possibly, where's my chicken?

 

Ok found some drop chance numbers in places around and beyond. This is my theory of how drops work. A database of drops for a specific mob has an item drop for these "random" numbers. I have concluded that the numbers are between 0 and 100,000. They the put items on a drop number for areas. Let's take The Deep Sea set. Let's assign 10 numbers for the ring. If any mob in atlantis dies with a number of 72,532 12,510 9,552 6,812 450 72 8 57,234 9,810 or 33,332. The ring will drop on death. This would make the drop rate at .01% Now the system can have a set drop needed per hour and would add 10 more numbers each minute or hour till a Deep Sea item dropped. That would be something people see when the server does not have a lot of people on. This drop table could also be server wide and would explain why a problem in heredur causes Agathon to have problems. I have seen only one server go down for maintenance so please take what I say with a grain of salt.

 

Go Papa, You Rock!! Your chicken is a random something generated by random something!!!!

Sorry for random post. The probability of me putting in a random post here? wish I knew it.

 

Well , for me random loot system does't not exist it is more like controlled dice . Developers done their best to give us some taste of randomness but something isn't clicking right here .
Drop depends on your build also time and map
The problem is big when u use same build on and on and farm with the same ppl I already know what I'm getting
Unique drop comes and goes in waves , only time when its really random is when the loot server is going mad .Indeed some id's are really lucky ,not mine , Devs can take certain actions to control drop for any player in the game also the system itself will prevent U from getting what U really need just in case u will become op ,system check calculate your build and release the drop U're getting
Also I would say that playing via client limit your randomness .
I wish I could have this magic.. program and assign some better drop to my char and yes how could I forget about predator farming since about 2 years and guess what ...nothing
I lol thru tears ...

"React on events (or overrides) when certain things happen
Simulate loot behavior of big players in the game "

This is how random loot system is here ....

 

I don'the know how the loot is diveded, but one of my toons, at level 30 (Normal upgrading...not xp blocking), was all equipped with either unique or legendary items. Another toon, at level 41, only had 2 legendary...and not for lack of trying!

 

There are other issues that complicates the drop equations:

Group effects is one. Group size changes the number of minions, and their offense and defense stats.
There must be a timer on some battles, because if the group takes forever, everyone gets squat. That happens a lot on dungeon bosses.
And then there is the player who was somewhere else during the battle, and gets a legendary or unique drop from it. "In the map" is good enough for random?
There may be counters involved, about relative damage give and damage taken, per person. IDK.

Another issue is location in the instance, or perhaps the instance itself. If you do enough cape quests in HoD, you know some instances and some locations are much better than others. Rich or poor on in-between is self evident.
Some maps are a waste of time. If Blackborg only gives you 67 wood per run, it is definitely time to move on.
Maybe randomness is programmed through map variation.

It really gets messy, fast.

 

I din't get this: a player didn't deal or receive damage and he dropped a unique, so there may be something involved about damage given and taken. It seems a random conclusion

There is just one thing that we know, the more you farm the more you drop, and it fits with a good random generator.

And i know that BP can change the drop chances (for example when gorga was dropping one unique every 10 runs ), but changing drop rates doesn't mean that 'their random generator is broken'. Here we have one for example: https://www.random.org/

 

You are correct about understanding how you can never deal damage and still get an item. Let me explain it this way. Some programs can be writen that if you deal the deadly blow the pool of numbers is small it pulls from. And in this new revised poll will have a higher chance to drop a certain item. I also think once we pass the level of the mob it broadens the number pool. Here is a few examples. Let's say most of the numbers 1 to 20,000 you get nothing Also so 80,000 to 100,000 you get nothing. The higher % damage you do removes a %age off the top number. Score the dead blow lowers the number 2x more for a max of -20,000. Once you pass the level of the mob you low number starts closer to 0.

So a person level 17 kills king heredur and he did 100% of the damage his number is between 20,000 and 80,000 where all the good drops are. A tag along that is level 50 that dealt 0 damage is 0 to 100,000 and It would be luck for him to get a good item...