Budikah wrote:

Grunklestank wrote:

Budikah wrote:Quite a few of the Ultima Online freeshards ended up putting in a pseudorandom system like is being talked about here.

Beforehand, you'd lose a fight because you might have missed 5-6 melee swings in a row when the enemy only needed a stiff breeze to die. People freaked out, much like we see here - and the devs on these servers saw fit to try the pseudo random system - I can't recall it's design, but I think each miss would weigh the next random number towards being a hit - it was still fairly random, but long miss or crit streaks went the way of the dodo and they truly became an event and something to talk about rather than a regular occurrence.

They realized it wasn't fun even if it was right, and they fixed it. Nobody much spoke of it after they made their changes.

This is exactly what I'm talking about and the kind of coding I'm referencing. I don't understand why this would be so hard to implement on Uth...???

Tbh true random RNG usage in a multiplayer game seems archaic af at this point. Even IF humans do suffer from negativity bias, it's been shown to create downright "not fun" elements over the years, and fixed in many other games. Uthgard doesn't need anymore "not fun" elements to chip away at it's already "not large" population

It probably isn't hard.

But it's work.
Work that needs to be done by somebody who probably disagrees with you.

LOL omg what a perfect summary of Uthgard 2.0

13k casts from live 1on1 comparison to Uthgard with matching results.
ppl still dont believe it.
stop feeling statistics, understand it.

Can probably neved be proven to be mot livelike although i still belive mythic changed something on the resist rates somewhere after toa (making resists more common to balancd out toa bonuses) but since it was a nerf prolly never documented

Last uthg i went live and shoed proof that on Uthg vs yelow u get resisted like vs oorange on Pendragon , Uthg VS orange like Pendragon VS red and so on , i got resisted by blues a couple of times here .. twice ! on same mob .. how the f is that even posible?:D and guess what no1 believed me.. they said its live like..but its not ..rng s ****** .. because 6 resists in a row VS a yelow mob should be imposible.

Budikah wrote:

Grunklestank wrote:

Budikah wrote:Quite a few of the Ultima Online freeshards ended up putting in a pseudorandom system like is being talked about here.

Beforehand, you'd lose a fight because you might have missed 5-6 melee swings in a row when the enemy only needed a stiff breeze to die. People freaked out, much like we see here - and the devs on these servers saw fit to try the pseudo random system - I can't recall it's design, but I think each miss would weigh the next random number towards being a hit - it was still fairly random, but long miss or crit streaks went the way of the dodo and they truly became an event and something to talk about rather than a regular occurrence.

They realized it wasn't fun even if it was right, and they fixed it. Nobody much spoke of it after they made their changes.

This is exactly what I'm talking about and the kind of coding I'm referencing. I don't understand why this would be so hard to implement on Uth...???

Tbh true random RNG usage in a multiplayer game seems archaic af at this point. Even IF humans do suffer from negativity bias, it's been shown to create downright "not fun" elements over the years, and fixed in many other games. Uthgard doesn't need anymore "not fun" elements to chip away at it's already "not large" population

It probably isn't hard.

But it's work.
Work that needs to be done by somebody who probably disagrees with you.

No its just incorrect and not live like. Thats why it doesn't get implemented. It skews probabilities. Its wrong.

Blue wrote:

Budikah wrote:

Grunklestank wrote:[quote="Budikah"]Quite a few of the Ultima Online freeshards ended up putting in a pseudorandom system like is being talked about here.

Beforehand, you'd lose a fight because you might have missed 5-6 melee swings in a row when the enemy only needed a stiff breeze to die. People freaked out, much like we see here - and the devs on these servers saw fit to try the pseudo random system - I can't recall it's design, but I think each miss would weigh the next random number towards being a hit - it was still fairly random, but long miss or crit streaks went the way of the dodo and they truly became an event and something to talk about rather than a regular occurrence.

They realized it wasn't fun even if it was right, and they fixed it. Nobody much spoke of it after they made their changes.

This is exactly what I'm talking about and the kind of coding I'm referencing. I don't understand why this would be so hard to implement on Uth...???

Tbh true random RNG usage in a multiplayer game seems archaic af at this point. Even IF humans do suffer from negativity bias, it's been shown to create downright "not fun" elements over the years, and fixed in many other games. Uthgard doesn't need anymore "not fun" elements to chip away at it's already "not large" population

It probably isn't hard.

But it's work.
Work that needs to be done by somebody who probably disagrees with you.

No its just incorrect and not live like. Thats why it doesn't get implemented. It skews probabilities. Its wrong.[/quote]

Sure, but if the player experience is frustrated by a more absolute RNG system, then what does it matter what is right or wrong?

It can skew probabilities all it wants - nobody is sitting there casting a spell going "well, 12.5 percent chance this gets through..." and relying on that - because it's a fools errand when it's just as possible for that and the next four cast spells to be resisted.

I don't really have any skin in this discussion. I don't expect it to change. I was just sharing my experiences from another game and another freeserver experience - people simply ended up liking the the faux-RNG better even if it was "wrong" and "not livelike". The same discussion occurred there, and the devs over there largely had the same response you did.

For you, yeah, statistics and proper data do matter - for your average player, they're just ****** that they blew 40% of their Mana pool trying to cast one spell, irregardless of if the numbers match up or if you properly emulated the live environment.

RNG streaks/series have a gaussian distribution which includes far ends, which means extremely lucky or extremely unlucky. In the end that creates emotions. Sure negative emotion feels more intensive. But thats the way random works. You can't say I played 1000 times in the lottery, now I deserve a win. You also can't say I made 150 craft attempts, now I deserve a MP with 100% certainty. You would also take 2 MP's in a row instead of preventing that.

I hear ya Blue, here is something that wouldn't go with my post for some reason.

http://www.uoguide.com/Random_Number_Generator

Long story short, humans are flawed in their observations, and nobody has conjured up an RNG system that works flawlessly.

My only reply to your above post would be that the RNG takes the role of a dungeon master in games like this - and while nobody "deserves" a MP success after 150 failures, a human dungeon master might eventually take pity, simply for the players enjoyment, and fudge some numbers - because it's a game played for enjoyment after all. Math and algorithms in place currently are indifferent to your suffering.

Budikah wrote:I hear ya Blue, here is something that wouldn't go with my post for some reason.

http://www.uoguide.com/Random_Number_Generator

Their explanation about streaks is just wrong. If something has a 92% probability the check would be randomNr < 0.92. Doesn't matter if the number is 0.920035622, 0.92546323 or 0.93343437. If its above or equal to 0.92 its a miss. If they round random numbers then they are for sure in trouble.

If someone tested it on live and the results were the same it should stay this way, I hate getting 3 resists in a row but thats what we call RANDOM.

Blue wrote:RNG streaks/series have a gaussian distribution which includes far ends, which means extremely lucky or extremely unlucky. In the end that creates emotions. Sure negative emotion feels more intensive. But thats the way random works. You can't say I played 1000 times in the lottery, now I deserve a win. You also can't say I made 150 craft attempts, now I deserve a MP with 100% certainty. You would also take 2 MP's in a row instead of preventing that.

I love RNG gaussian function induced emotions, negative or positive! Gets the blood flowing . Generating a number that is truly random is the age old question and I feel like people are trying to get the DEVs to answer it here. PRNG is not ideal because its not truly random. Like Blue said, getting many resists in a row just proves that it is random. There has been many tests done with multiple DEV responses. IDK what else there is to be said, but this will probably come up again and again.

Okay,

This will help put the issue to bed.

I did an analysis of a trailing 15, 25, 50, and 100 casts on the daoc chat logs i accumulated and lurker provided, 50 casts was the best representative statistic

These numbers are (from lurker's logs)
-----------------------------------------------------------------
Resist Rate (Trailing 50 casts)
0% - 2
2% - 22
4% - 37
6% - 119
8% - 243
10% - 307
12% - 346
14% - 405
16% - 467
18% - 546
20% - 414
22% - 251
24% - 138
26% - 74
28% - 24

I generated 10 sets of random numbers, and assuming a resist rate of 15% here is the trailing 50 cast distribution based on an independent RNG.
-----------------------------------------------------------------
4% - 37
6% - 119
8% - 231
10% - 288
12% - 343
14% - 402
16% - 467
18% - 546
20% - 415
22% - 258
24% - 143
26% - 75
28% - 25
30% - 2
32% - 1

0% - 2
2% - 22
4% - 37
6% - 119
8% - 231
10% - 288
12% - 343
14% - 402
16% - 467
18% - 546
20% - 415
22% - 258
24% - 143
26% - 75
28% - 25


4% - 41
6% - 143
8% - 339
10% - 467
12% - 502
14% - 594
16% - 459
18% - 335
20% - 230
22% - 121
24% - 79
26% - 54
28% - 35
30% - 9
32% - 2

2% - 17
4% - 63
6% - 175
8% - 260
10% - 343
12% - 455
14% - 525
16% - 562
18% - 396
20% - 265
22% - 176
24% - 95
26% - 43
28% - 18
30% - 2

4% - 60
6% - 236
8% - 302
10% - 333
12% - 397
14% - 518
16% - 542
18% - 399
20% - 222
22% - 173
24% - 113
26% - 67
28% - 11
30% - 9
32% - 3
34% - 7
36% - 3

0% - 11
2% - 8
4% - 20
6% - 33
8% - 158
10% - 299
12% - 420
14% - 484
16% - 547
18% - 500
20% - 435
22% - 235
24% - 123
26% - 71
28% - 47
30% - 4

4% - 51
6% - 147
8% - 273
10% - 449
12% - 452
14% - 459
16% - 541
18% - 373
20% - 231
22% - 137
24% - 146
26% - 80
28% - 44
30% - 1

14% - 509
16% - 488
18% - 498
20% - 362
22% - 204
24% - 114
26% - 37
12% - 497
10% - 354
8% - 161
6% - 89
4% - 20
28% - 41
30% - 20
32% - 1

4% - 60
6% - 236
8% - 302
10% - 333
12% - 397
14% - 518
16% - 542
18% - 399
20% - 222
22% - 173
24% - 113
26% - 67
28% - 11
30% - 9
32% - 3
34% - 7
36% - 3

0% - 0
2% - 1
4% - 23
6% - 94
8% - 276
10% - 388
12% - 522
14% - 560
16% - 545
18% - 377
20% - 282
22% - 174
24% - 98
26% - 47
28% - 8

I was casting a lvl 47 dd on mobs which were 50-53. My resist rate so far after 1250 casts is 15.54%. I did have 1 streak of 34% resists. I need to get more data, but i had turned on logging AFTER the worst of it came. It is feasible to have a 34% resist rate in 1200 casts, but reaching that rate on 1/3 of the casts as the prior tests makes me skeptical, and it seems it is just VERY unlucky (considering the really bad streaks were behind me).

In reality 10% of the 1250 rng stretches I ran hit the 34% resists (or higher) over 50 casts, and 40% resist over 50 casts about 2% of the time.

This proves that the rng is correct, but mastery of focus is DESPERATELY needed to balance out the resist rate. If spell resists are reduced by even 3% it has a HUGE impact on the skew and streakiness going from 15% to 12%, makes the upper bound of resist streakiness go from 40% to 34% over a trailing 50, and only 6% of the sets of 1250 will have a trailing 50 casts meet or exceed the 30% resist threshold.

Interesting stuff Entrancementx.

Very similar results looking at your simulated RNG vs Live data.
Because graphics are lovely for humans to understand data, I popped that all in a chart...



Certainly seems to be a common pattern between live and your test results using a 3rd party RNG.
If there was some form of smoothing you would expect to see a noticeable difference for the 'Lurker' set vs any others

looks like some of you are missing Pweet's point. please refer to
https://www.surveysystem.com/sscalc.htm - determine your sample size to get statistics you learn in school and college. casting 100 times same spell in game - is NOT statistics.

for "population" i would probably use total number of spells casted daily on average.

gl

vadox wrote:looks like some of you are missing Pweet's point. please refer to
https://www.surveysystem.com/sscalc.htm - determine your sample size to get statistics you learn in school and college. casting 100 times same spell in game - is NOT statistics.

for "population" i would probably use total number of spells casted daily on average.

gl

I don't think anybody has been looking at samples of 100.

They are between 1,250 and 13,000 at present.

Obviously larger sample sizes are better, but smaller sample sizes can provide interesting data too, with a lower degree of accuracy...