Let’s be honest, RNG in WoW is down right broken. This should be the first thing Microsoft fixes.
3 Likes
What does that mean?
Are you just saying that you think they apply elements of chance in inappropriate places in the game design, or are you literally suggesting that the algorithm Blizzard is using for pseudorandom number generation is nonfunctional in some way?
1 Like
why do people still think Microsoft will run wow
6 Likes
Could you give some examples?
Imagine an mmorpg with no RNG, mobs could only drop 1 item and it would drop every time. The only way they could make some items more rare then others is to make some mobs super hard… 50 mil HP, charm half the raid into attacking the other half, anyone who doesn’t dance in the green circle gets insta-ported to Ironforge etc…
1 Like
It didn’t give them what they wanted so it must be broken!
3 Likes
I’m legitimately not sure what the OP is saying. I am in game design for slots, so that runs purely off of RNG. If an engineer told me our “RNG was broken” the only way to interpret that is that our RNG algorithm is consistently failing to uniformly cover the output interval within some tolerance interval, which would require a lot of dedicated testing.
Furthermore, this would never happen. RNG algorithms usually aren’t even proprietary. There’s a dozen or so industry standard algorithms named after smart guys that everyone uses. It’s not really something that you point towards as the source of potential issues.
I’m not dismissing the OPs sentiment, it’s just that whatever that sentiment is, the way he is describing it does not make sense or say what he thinks it does. That’s why I asked for clarification.
2 Likes
Yaaaaaa then I can do the raid/dungeons once in a 6 month season and have bis gear. I love the way this guy thinks
I ask myself this too…Microsoft won’t say boo unless wow just goes to pot and loses a lot of money…
All Microsoft wanted was king and call of duty
1 Like
As a web dev who spent more time at my last gig than I really want to admit trying (and failing) to make our unit tests less flaky, I don’t know how you’d even start testing something that’s random by definition! 
1 Like
There are many things one might do, depending on the nature of the suspected flaw.
The most simple example is that if the RNG is functional, its output set will follow the law of large numbers. If I take an algorithm and tell it to generate 1,000,000,000 numbers between 0 and 50 (inclusive), and I take the average of those 1,000,000,000 numbers, and it doesn’t look something like 25.000000164, that’s not a good sign.
Though again, problems like that have always been a result of programmer code doing something with RNG outputs after they have been generated. It is never just that the direct output of any RNG algorithm itself is failing to give us usable pseudorandom numbers. These algorithms are just really well known and standardized, it would be like finding out that the Pythagorean Theorem doesn’t work.
1 Like
RNG has been a thing forever and will continue to be a thing to some capacity.
There’s hardly any problematic RNG right now anyways.
Which direction are we talking about?
More RNG?
Less RNG?
1 Like
You no longer want it to be random?
1 Like
If I could find a picture of a slot machine and edit “RNG” on it…
Not saying more or less. I’m just saying fixed.
Let’s use golden mane for example. .02% drop chance. Logically speaking 100/.02= 5k Theoretically that should drop within 5k. I got it 2 years ago around 15-20k.
Rng needs a tweak to have a player advantage. It shouldn’t be be a 1 in 5000 chance every time.
1 Like
While you could get it within 5k kills, that is NOT how probability works.
Sometimes you get lucky and you get it before the expected value. Sometimes you are just unlucky and you get after the expected value.
Given that each event is independent, the chance of it dropping remains at 0.02%.
So you want bad luck protection? aka make it a cdf? Where the probability gets added after every unsuccessful event? Also not everything that as a random drop chance is a 1/5000 chance lol. Stop being so dramatic.
The odds of getting a .02% drop in the first 5,000 kills is 1 - the odds of not getting the drop 5,000 times in a row. Since these events are independent, you just multiply them together, so this chance is 99.98% (the odds of not getting the drop once) to the 5000th power, or about 37%. Correspondingly, the odds of getting the drop by then are only about 63%.
The “expected value” is indeed 1.0 after 5,000 kills, but this is a weighted value. A decent chunk of the 63% of players who got the drop in 5,000 kills would actually have gotten multiple drops, if they bothered to keep running the full 5,000 (which they don’t, of course). The sum of products - 1 * the odds of getting exactly 1, 2 * the odds of getting exactly 2, and so forth - does equal 1.
There’s about a 1.8% chance to not get the drop 20,000 times in a row, so statistically, of all the players who started farming at the same time as you, you got it ahead of about 2% of them.
These numbers are not in aberrational or a sign that the RNG is broken - just that people don’t always have a great intuition for how random chances work.
4 Likes
they are the same clowns who thought Bobby Kotick ran wow.
1 Like
Casinos were built for people with your level of understanding on probability.
4 Likes