Skillup Rate RNG

[PabloEdvardo]

I've been maxing casting skills on my Ogre SK every level for a while now.

I recently got 46 which is when casting skills finally cap at 235.

I spent about 10 minutes getting Abjuration from 230 to 235.

I spent about 10 minutes getting Divination from 230 to 234...

I've now cast Locate Corpse over 1421 times, and I have yet to get the last point in Divination.

Anyone else seen some extremely bad / strange skillup RNG?

This is the stuff that makes me feel like streaky RNG isn't just a rumor, but statistically evident.

[zillabunny]

Yeah jc is a nightmare I will get 3 skill ups in a row then dump 500 plat and get nothing on the next 15 combines

[fadetree]

the skill RNG being streaky is not just a rumor. You can see a repeated pattern even when it's moving along - you'll see 15-20 fails, then a short period where you get one to three skillups, then off on another long fail sequence.
I looked at the core RNG function that the EqEmu code is using, and it's just the usual perfectly good RNG call from the c++ std library using the "Mersenne Twister" algorithm. That is a good quality (i.e., not streaky) generator.
Assuming that the P99 code is probably using the same call, it means that the observed streakiness is not coming from the RNG. I think there almost certainly is some other processing going on underneath the hood that is affecting the results. I am planning to trace what all happens on a skill roll from the time the rng generates its number all the way up to when the skillup result happens to see if I can see any other stuff going on.

[Frug]

the skill RNG being streaky is not just a rumor. You can see a repeated pattern even when it's moving along - you'll see 15-20 fails, then a short period where you get one to three skillups, then off on another long fail sequence.
I looked at the core RNG function that the EqEmu code is using, and it's just the usual perfectly good RNG call from the c++ std library using the "Mersenne Twister" algorithm. That is a good quality (i.e., not streaky) generator.
Assuming that the P99 code is probably using the same call, it means that the observed streakiness is not coming from the RNG. I think there almost certainly is some other processing going on underneath the hood that is affecting the results. I am planning to trace what all happens on a skill roll from the time the rng generates its number all the way up to when the skillup result happens to see if I can see any other stuff going on.

Check fizzles while you're there. I know confirmation bias is hard to overcome, and I need to do some proper data analysis, but fizzling seems WAY MORE common after having fizzled once. I get 2, 3 fizzles in a row MORE often than just one, it feels like. Which for any RNG seems suspect, given the chance of a fizzle is less than half.

[Teppler]

If skill up rng isn’t so random and we accept this(and I agree) what does this mean for loot drops?

[cd288]

the skill RNG being streaky is not just a rumor. You can see a repeated pattern even when it's moving along - you'll see 15-20 fails, then a short period where you get one to three skillups, then off on another long fail sequence.
I looked at the core RNG function that the EqEmu code is using, and it's just the usual perfectly good RNG call from the c++ std library using the "Mersenne Twister" algorithm. That is a good quality (i.e., not streaky) generator.
Assuming that the P99 code is probably using the same call, it means that the observed streakiness is not coming from the RNG. I think there almost certainly is some other processing going on underneath the hood that is affecting the results. I am planning to trace what all happens on a skill roll from the time the rng generates its number all the way up to when the skillup result happens to see if I can see any other stuff going on.

Lol how is that not the definition of streaky? Streaky is just that, streaks of good and bad rather than it being randomly interspersed throughout. So you have a streak of fails, then you have a streak of a few skill ups.

[Tecmos Deception]

Lol how is that not the definition of streaky? Streaky is just that, streaks of good and bad rather than it being randomly interspersed throughout. So you have a streak of fails, then you have a streak of a few skill ups.

That IS the definition of streaky. The issue is that p99's RNG is disproportionately streaky when you look at the average % for different things to occur.

If something has a 50/50 chance of happening, then it will tend to not have streaks of 5+ very often. They'll happen, but you'll be able to predict quite accurately how often a streak of any given length will happen.

On p99, something that has a 50/50 chance of happening has streaks of whatever length too consistently for the RNG to actually be even close to actually random.

[Teppler]

Lol how is that not the definition of streaky? Streaky is just that, streaks of good and bad rather than it being randomly interspersed throughout. So you have a streak of fails, then you have a streak of a few skill ups.

It’s clearly non randomized streakiness because of what pretty much everyone has witnessed.

You get 3 skill ups over 3 seconds then nothing over minutes and that happens somewhat consistently. Who hasn’t experienced this... a lot? That is clearly not totally randomized over the longer time period. There’s some formula with variables.

[cd288]

That IS the definition of streaky. The issue is that p99's RNG is disproportionately streaky when you look at the average % for different things to occur.

If something has a 50/50 chance of happening, then it will tend to not have streaks of 5+ very often. They'll happen, but you'll be able to predict quite accurately how often a streak of any given length will happen.

On p99, something that has a 50/50 chance of happening has streaks of whatever length too consistently for the RNG to actually be even close to actually random.

Ah I misread their post and thought they said streakiness IS just a rumor. Gonna edit mine now lol

[Thelion]

The streakyness seems to occur on drop rates aswell. And FD fails. steaky rng

[Smellybuttface]

There are other strange things with skill ups that seem to defy any sort of RNG.

Like the fact that some skills you need to have one point in before you can really start seeing gains (referring of course to non-trained skills, like Sense Heading, Bind Wound, Swimming).

I’m also going to assume Devs were aware of the streakiness of RNG, considering on Live they implemented the following change in a patch:

“ - Dramatically increased the chance to gain skills when your current skill level is less than your class' skill cap from 5 or 10 levels less than your current level.”

I’ve played on Live after this patch, and it is a huge difference. If I was gaining skill ups under that level cap on say spell skills, the rate was just about 100%. This patch isn’t definitive, but I think someone recognized the fact of how absurdly slow some of the skill ups were in the game and tried to reconcile that.

[Frug]

There are other strange things with skill ups that seem to defy any sort of RNG.

Like the fact that some skills you need to have one point in before you can really start seeing gains (referring of course to non-trained skills, like Sense Heading, Bind Wound, Swimming).

I'm pretty sure swimming can start from 0, but its % chance to skillup seems absurdly low compared to live (faulty memory?) or blue (less faulty).

I think Tecmos mentioned this some time ago.

I raised Bind Wound from 0 to 1 on my mage last night, as well.

[zillabunny]

If skill up rng isn’t so random and we accept this(and I agree) what does this mean for loot drops?


We've seen multiple times where phin drops 3 kedge robes or 2 mag wands

[Wallicker]

Perhaps RnG is tied to time in some way. Be interested if anyone has a massive log of kills and drops for the same mob spanning all times of day and if those drops Correlate through ToD or Time of Engage or Time of spawn in real life or game time.

[zillabunny]

excatly my thoughts walllicker it's gotta be tied to time in some way

[PabloEdvardo]

Add a few more hundred casts of Locate Corpse, still at 234 Divination.

Meanwhile, I maxed Conjuration to 235 just from some Disease Clouds in my last grouping session.

Could my Divination cap be bugged at lvl46 somehow? Such a strange phenomenon.

[fadetree]

Talking about randomness is weird and confusing, because we actually don't know what it is. If we did know, then it wouldn't be random. We know what it ought to look like from a statistical sense, so we can say that something doesn't SEEM random, but that can change over the length of sequence you are looking at. A sequence of 1 is not random at all. A sequence of 10 really can't be said to be random. Sequences of hundreds is where you start to get statistical grasp, but all you can ever say is that it 'seems' random. Then, when you start talking about repeating subpatterns that seem nonrandom, such as the skillups pattern, the math gets deep.
So that's some philosophical B.S., but anyway the pattern I have seen in skillups and other things over the 20 years that I have been playing have got to not be random. I feel it in my bones there's something else going on in the machinery besides a D100 roll. Especially in skillups, it really feels to me like there's some kind of state being saved and used or weighting based on how many times you have tried since your last skillup...and the math is not smooth (hi there, hell levels) so you get a cluster of skillups at the end of the sequence instead of a smooth distribution.

[Teppler]

We've seen multiple times where phin drops 3 kedge robes or 2 mag wands

I’ve seen some wacky ass shit with consecutive drops that always made me believe there’s more to the formulas than people think.

[Madbad]

I’ve seen some wacky ass shit with consecutive drops that always made me believe there’s more to the formulas than people think.

Agreed

[silo32]

I’ve seen some wacky ass shit with consecutive drops that always made me believe there’s more to the formulas than people think.

this

[Naxy]

Streaky random is definitely Classic. Whilst doing a trade skill I would always space out my combines until I got a skillup and the rapidly do the next few and often get one or two more skillups. I have long suspected that the RNG is linked to the tick, but I have no data to prove it.

[PabloEdvardo]

https://i.imgur.com/qP41zgD.png

Now I really feel like my Divination cap is bugged at 46.