Ancient Items Roll Like Crap

General Discussion
If you ask me any Topic that discuses RNG good or bad is totally ILLOGICAL mathematically, when the RNG can be adjusted at any time, either by whim, fancy or given order, by the Dev's or whom ever is overseeing this game.
05/23/2019 06:46 AMPosted by Windigo
If you ask me any Topic that discuses RNG good or bad is totally ILLOGICAL mathematically, when the RNG can be adjusted at any time, either by whim, fancy or given order, by the Dev's or whom ever is overseeing this game.

Ooh, great! We're talking about how the RNG might be on items after the RNG has been altered? Well, in which case, I think the drop rate of one in 17.3 bazillion is too low for primordial super mega ancient legendaries of awesomesauce and the affix values, even though the ranges have been increased by 5786% compared to standard legendaries, always roll in the lowest 10% of the range, except when they don't and disprove that.
it'll be perfect if Blizz increase legendary affix for ancient items
for example :
YANG'S RECURVE
Normal:
Multishot attacks 50% faster and its damage is increased by 150%. (Demon Hunter Only) [150 - 200]%
Ancient:
Multishot attacks 50% faster and its damage is increased by 175%. (Demon Hunter Only) [175 - 225]%

Just add range of normal item (50%) starting from normal+ half of range (175%)
05/23/2019 06:46 AMPosted by Windigo
If you ask me any Topic that discuses RNG good or bad is totally ILLOGICAL mathematically, when the RNG can be adjusted at any time, either by whim, fancy or given order, by the Dev's or whom ever is overseeing this game.

Ooh, great! We're talking about how the RNG might be on items after the RNG has been altered? Well, in which case, I think the drop rate of one in 17.3 bazillion is too low for primordial super mega ancient legendaries of awesomesauce and the affix values, even though the ranges have been increased by 5786% compared to standard legendaries, always roll in the lowest 10% of the range, except when they don't and disprove that.


I will admit, that's a lot of Brazilians. I have no idea how I would get that many in one place at a time.
05/23/2019 07:19 AMPosted by Meteorblade
Ooh, great! We're talking about how the RNG might be on items after the RNG has been altered? Well, in which case, I think the drop rate of one in 17.3 bazillion is too low for primordial super mega ancient legendaries of awesomesauce and the affix values, even though the ranges have been increased by 5786% compared to standard legendaries, always roll in the lowest 10% of the range, except when they don't and disprove that.

LoL, Math has always been gobbledygook to me especially when there is a Wife, Child or a evil and or sneaky Developer factors of change involved. You know the W+C<D formula ;)
05/22/2019 01:00 PMPosted by StoneOld
05/22/2019 02:32 AMPosted by Meteorblade
Of course I believe his "story" more. Your "story" is based on how you feel about items. His "story" is based on empirical data about thousands of items.
OK, this one more try.

How do you know his story is based on anything? Have you seen the sheet? Have you verified its accuracy? Why are you willing to take that guy's story about property value distribution and not mine? Perhaps because he agrees with your opinion? So even though me and that guy present to you the exact same proof* - words on this web page - you are inclined to say he is right and I am wrong.

Why?

*actually my proof is better because I recorded the stats of an item to demonstrate my claim.
05/22/2019 03:50 AMPosted by Amonra
If the differences are as large as OP is suggesting then a few thousand items would be more than sufficient evidence for such a significant bias to be obvious.

But we would need to actually see this evidence.


There's actually a very specific reason I didn't bother making the sheet available: you would just claim I made the data up.

I didn't contribute to this post to try to change your mind because I know all too well you would just twist any data anyone else provides. I contributed to this post just to clarify for any users with any sense of reason that that affix rolls on ancients don't appear to be weighted to favor either higher or lower rolls.
@Dreyda

As a DB Guru, I'd love to see the raw data.

From my tests, the D3 RNG converges faster then the RNG built into the DB.
(Ie Smaller window is needed to match "%crit hit" 95% of the time)
05/24/2019 11:39 AMPosted by Dreyda
There's actually a very specific reason I didn't bother making the sheet available: you would just claim I made the data up.
What, are you dense?

How about this? I own a 7-second quarter-mile drag car I made out of a 2008 Dodge Challenger. Do you believe that?

Now, If I showed you a video of me pulling my driver's license out of my pocket and showing that and the vehicle title to you, then driving a 7.28 pass, then having someone bring me a laptop and video me logging onto the game and this forum, would you be convinced then that I own the car then?

I'm not here to call your data fake. I am saying that there is no reason whatsoever for anyone to believe your claim unless you provide some basis for it. So, show me the money.

My claim is that after watching tens of thousands of these Ancient items drop, that their properties roll in the lower half of the possible range, and most of them very near the lower third of the range, if not lower. I then presented a representative item. I can leave out my guess on the dev's philosophy of this condition, because who knows? But I can't believe that anyone who has played this game for any length of time can't see the same trends.

So I am sharply interested in seeing data that puts the mode in the center of the range. Or, according to you, mean = median = mode.

Show me the money!
Mode === most common value

Mean === average

I just want to see the raw data because I dont feel like hacking bot software to collect the data myself.
^^ Yeah, and according to him, all the property values averaged out right down the middle, which means, practically speaking, mean=mode=median (this is assuming that the values did not roll half at the minimum value, and half at the max).
05/24/2019 05:11 PMPosted by Slamboney
My claim is that after watching tens of thousands of these Ancient items drop, that their properties roll in the lower half of the possible range, and most of them very near the lower third of the range, if not lower.

I too have seen tens of thousands of items and do not concur with your claim.

05/24/2019 05:11 PMPosted by Slamboney
I then presented a representative item.

I present Crusader, Will of the Light gloves with the following affixes / values...
Primary
7% Attack Speed
47% Critical Hit Damage
9.5% Critical Hit Chance
20% Area Damage

Secondary
178 Lightning Resist
3268 Thorns
According to your argument, a single item can be seen to be representative of an entire statistical distribution, so the single item I have presented not only means that ancients must roll with stats in the top 10% of their ranges but that ancient gloves will always be quadfecta items. Except that's just as invalid an argument as the one you're presenting.

Oh, and in case you think I made up those stats, they're on this hero...
https://eu.diablo3.com/en/profile/Meteorblade-2266/hero/107355817
05/24/2019 08:05 PMPosted by Slamboney
^^ Yeah, and according to him, all the property values averaged out right down the middle, which means, practically speaking, mean=mode=median (this is assuming that the values did not roll half at the minimum value, and half at the max).

Let's take a population of 100 items which can have values from 0 to 100.
    50 of the items have a value of 6.
    30 of the items have a value of 90.
    20 of the items have a value of 100.

The mean for this population is 50.
The mode for this population is 6.

So, that's a population where the mean is exactly 50% of the possible value range, but the mode is at 6% of the possible value range, i.e. mean =/= mode.
05/24/2019 11:17 PMPosted by Meteorblade
According to your argument, a single item can be seen to be representative of an entire statistical distribution, so the single item I have presented not only means that ancients must roll with stats in the top 10% of their ranges but that ancient gloves will always be quadfecta items. Except that's just as invalid an argument as the one you're presenting.
I don't see anywhere where you've made the claim that after watching tens of thousands of these items drop, that the property values tend to roll in these ranges, and that Ancient items roll very well.

You got a good pair of gloves there. No question. But do you believe all Ancient items roll so well, and are you prepared to make that claim? If you did, you could present those gloves as an example. But you haven't made that claim. You have simply showed me a well rolled Ancient item, and said that because it exists, my argument is invalid.

[Edit for clarity]

In other words, had you seen thousands of Ancient items drop like that, you could make the claim that they all tend to, and show those gloves as a representative. If, hypothetically, Ancient items did tend to drop like your gloves there, I would not be complaining, and I wouldn't have made my post. But Ancient items don't tend to drop like your gloves, they tend to drop like my belt, and I have watched thousands of items over the last 4 years do so; thus, my claim.

Do you begin to see the problem with your premise?

*And MB, I was just figuring something like in the range of 6-10, 8 rolled the most times, eight happened to land right in the middle of the set, the average of the set happened to be eight. And yeah, I know, practically speaking that would never happen.
05/25/2019 09:11 PMPosted by Slamboney
Do you begin to see the problem with your premise?


05/25/2019 09:11 PMPosted by Slamboney
You have simply showed me a well rolled Ancient item, and said that because it exists, my argument is invalid.


Read the last sentence of the paragraph you quoted.

05/24/2019 11:17 PMPosted by Meteorblade
Except that's just as invalid an argument as the one you're presenting.
05/25/2019 09:11 PMPosted by Slamboney
I don't see anywhere where you've made the claim that after watching tens of thousands of these items drop, that the property values tend to roll in these ranges, and that Ancient items roll very well.

That's correct, I have made no such claim. Had I done so, I would need to provide evidence for that claim. Instead, what I have done is said that I do not believe your claim. Of course, me not believing your claim is not the same as me claiming the opposite of your claim.

05/25/2019 09:11 PMPosted by Slamboney
You got a good pair of gloves there. No question. But do you believe all Ancient items roll so well, and are you prepared to make that claim?

I do not believe all items roll so well.
I have not claimed that they do so.
I will not claim that they do so.

05/25/2019 09:11 PMPosted by Slamboney
If you did, you could present those gloves as an example. But you haven't made that claim. You have simply showed me a well rolled Ancient item, and said that because it exists, my argument is invalid.

I said that if I had made the argument that all items roll with fantastic stats and used those gloves as an example to back up that claim, then that argument would be as bad as the one you're presenting, i.e. it was an example of how bad a single item is as evidence of an entire statistical population. I did this to demonstrate how bad your argument was, not to present one of my own.
Are SlamB and StoneO family related...?
05/26/2019 03:05 AMPosted by Jazz
Are SlamB and StoneO family related...?

He's stated in the past that he has six accounts.
Slamboney / StoneOld / StoneCold are three of them.

He seems to chop and change which account name he's posting under so that when someone thinks to themselves "Hold on, that sounds familiar" it's harder to find the previous occasions where he's said the same thing because you can't use post history to find it unless you also know his aliases.
Amazing! Stone Cold Crazy! (Queen title).

Join the Conversation

Return to Forum