Personally I’d rather a currency system, where you win some kind of filthy lucre in events then you trade them in for orbs. I’d be fine with “rarer” orbs like ascension orbs costing more, or randomized chaos orbs being sold at a discount. Then players could decide what orbs they want and move at a pace metered only by their guild or personal progress.
Instead we have, “It’s fair, but it only looks fair if you can see the numbers we can’t show you, and actually if you try and use math it’s not really fair.”
This topic has been discussed ad nauseam in the past, and while academically interesting, unlikely to produce any real traction. I’ve worked as a game developer in various capacities, and here are some things to note:
You don’t generally use true random number generators when you write game code. They’re way too slow, and the requirements for “true randomness” are simply not there. Instead, you use pseudo-random number generators (pRNG), which are essentially just fancy math that produces results that feel random to humans. This is good enough for just about everything but (real-money) gambling scenarios, where a harder degree of randomness is mandated by law.
You don’t generally create your own pRNG. Almost all development frameworks and programming languages come with built-in pRNG libraries. If there is fault here, it is usually in those libraries.
The game client was written in C# (language) and Unity (framework), both of which have built-in generators that have passed muster for thousands of applications and games. Client-side pRNG is used for things like in-battle randomness (% chance to activate a skill, how the gems are seeded in skydrops, etc.)
We know nothing about the language that the servers were written. All bets are off there. Server-side pRNG is used in things like chest opening, PvP matchmaking, and the like.
Streakiness isn’t ideal in a pRNG, and there are tests that measure it in various ways. It’s almost unavoidable that some streakiness will exist in any given implementation. When I researched this in the past, C#’s pRNG measured quite favorably. It’s possible that this has changed, though unlikely.
I’m only writing this because someone I respect did math I respect that found problems with the GoW RNG, which:
We know nothing about, and isn’t client-side, so isn’t necessarily guaranteed to be the built-in C# or Unity pRNG.
Further, we don’t know how the RNG output is being used in the game code and there are various manipulations or interpretations that can accidentally contribute to a bias.
I generally agree with you and trust the C# RNG for most things, but we’re looking at very strange numbers that suggest an RNG with a more severe bias than the one you’re reassuring us is OK. So like, uh, sure. The possibly different RNG you’re defending is fine, I’m not quite sure I believe its output is being used unmolested by GoW.
Well, also to be fair, akots was discussing events generated by the client-side pRNG, so we actually have no indication either way about whether server RNG (chest openings and the like) are streaky based on the test methodology he used.
As for how the client-side numbers are “massaged,” if at all? You have a point. I suppose it comes down to whether you believe the devs or not when they indicate that an event is truly “fair” (they know what the code says; at that point it stops being about the random numbers themselves and becomes a matter of trust).
But yet we are left in the dark over basically the same thing just titled Chaos Orbs and Legendary Tasks.
Having both hands behind your back is a sign of respect. Having one hand in front palm open while the other hand is behind your back causes instant trust issues.
I respectfully did a LT thread asking for clarification months ago. But since it was posted here and not asked in a stream it therefore becomes a very low priority to be answered if at all. I would say “ignored”… But I believe at least one dev bothered to read thread and then decided not to answer… Or wasn’t allowed to answer.
Everything we know is based on business research that decides that it is in the devs best interest for us to know.
Everything we don’t know is based on business research that decides that it is in the devs best interest for us not to know.
I read both your previous replies before this one, as well as the one from @Koromac, and had a similar idea. Remove the Chaoas Orb as a reward and introduce the Chaos Token, or possibly the Orc’oin.
Then players could visit the new Global Goblin Bank of Zaejin, and access their Orb Banks. Every 4th and 8th would be Clan, the 11th would be Ascension, the other 8 an even split of Frustration and Disappointment.
There would also be slot machines available (Goblins are gamblers by nature) where unwanted orbs could be inserted with a chance (for a nominal fee) to become a token of Cedric (or the other Medal components).
And perhaps also: Orc’oins could be used in a similar progression system for Medal components, tho I don’t know what an acceptable ratio might be. I ran this idea by Brian the Lucky, and he liked it. High King Irongut was also onboard.
Ah you’re right, I didn’t think hard enough about if akots was discussing client-side or server-side.
As to your second point, the problem is if we have to believe the devs, the current sanctioned dev response is “the RNG is streaky but it works out in the end”. Hence my response: the phenomenon they call “fair” actually results in a distribution of “winners” and “losers” that is not the same one described by their posted drop rates.
It would be really complicated, but I could do the math and set up 100 players pulling 4 chaos orbs and tell you, with error bars, how many players are expected to get 0, 1, 2, 3, and 4 ascension orbs. (If I had the rates, I can’t remember if these are posted. Maybe I should’ve picked chaos shard treasures.)
The “streaky” RNG will award the orbs at a different rate. More players will get 0, fewer will get 1, more will get 2, and more will get 3. That sounds great but it’s possible it means there’s a much greater proportion of players getting 0 than the posted rate indicates. And I don’t think those players are excited that they very likely lost 1 orb so someone else could get 3.
I think that’s why the response to the Chaos Shard portal drop rates is to change the posted rates. My guess is they saw the thread and ran the data, found that they were indeed giving out things at a different rate, and determined it’d be easier to post the rate their data suggests than fix the RNG to do what they’re asking it to do.
