LK /p3-4 Superchest Drops (The Search for Lo Is Over) ifra

Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

Negative results: did maybe 300 /p3 runs so far, no runes from Io and upward found.

As for probability, either the seeds are biased or they aren't. The whole reason we get fixed repeatable drops is the small number of seeds. If they aren't biased, it works as we all suspect. If they are biased, I haven't yet seen evidence of that. Quoting several hundred runs or even several thousand as representative of the overall behavior is simply not correct -- it would be like me saying no good runes exist on /p3 because I never found any in 300 runs. Some people get lucky and others don't, that's the basic nature of probability. The only cure to make your experience more closely match the expected is to do more runs. And by the way, the good LK map has 6 special chests. 65,536 / 6 = 10923 which is where 11k runs comes from. Doing 11k 6-chest runs doesn't ensure you see all 65,536 drops. But doing 11k 6-chest runs will produce ON AVERAGE a full representative set of drops. But for any given person, it's this:

If P(Ber in 1 special chest) = 1/65,536 then:
P(Ber in 11k 6-chest runs) =1 - (1 - 1/65,536)^(11,000 * 6) = 63.47%

So statistically, if everyone in the world right now did 11k 6-chest LK runs, 63.47% would end up with one or more Ber runes, and 36.53% would end up Ber-less. Them's the breaks. And if you want a higher chance for yourself:

For 90% chance of Ber: (1 - 0.9) = (1 - 1/65,536)^(n * 6); n = 25,150 6-chest runs
For 95% chance of Ber: (1 - 0.95) = (1 - 1/65,536)^(n * 6); n = 32,721 6-chest runs
 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

Code:
Rune Item#1              Item#2              Item#3              Item#4              Poster
Io
-    Crossbow            Tulwar                                                      Phosheez
-    Falchion            Giant Axe           Wire Fleece                             Neck Romancer
-    Sacred Globe                                                                    Neck Romancer
-    Broad Axe           Gauntlets           Military Pick       Serpentskin Armor   Neck Romancer
-    Ballista            Long Sword          Sabre               Small Crescent      Neck Romancer

Lum
-    Composite Bow       Ghost Armor         Trelissed Armor                         Phosheez
-    Ceremonial Javelin                                                              Cygnus
-    Glaive              Militairy Axe       Ring                Sallet              Neck Romancer

Ko
-    Ceremonial Javelin  LQ Large Shield                                             wwzerozeroone
-    Chu-Ko-Nu           Greater Claws       Scissors Quhab      War Axe             Neck Romancer
-    Sacred Globe        Shadow Bow          Short Siege Bow

Fal
-    Short Spear         War Dart                                                    Neck Romancer
-    Cudgel              Light Belt          Tulwar              Siege Crossbow      Neck Romancer

Lem
-    Broad Sword                                                                     Neck Romancer

Pul
-    Gargoyle Head                                                                   Neck Romancer

Um
-    Cedar Bow           Scale Mail          War Sword                               Phosheez
-    Battle Axe          Gilded Shield       Great Maul          Pilum               Nightfind
-    Bec-De-Corbin                                                                   Neck Romancer

Mal
-

Ist
-    Glaive                                                                          wwzerozeroone

Gul
-    Large Charm                                                                     Reborn2k
-    Large Charm          Flawless Sapphire (does that count as an item?)            Broncobiv2
Vex
-

Ohm
-

Lo
-    Balanced Axe                                                                    Phosheez
-    Balanced Axe                                                                    Reborn2k
-    Balanced Axe                                                                    erics

Sur
-    War Hammer          Battle Axe                                                  Neck Romancer

Ber
-    Hawk Helm                                                                       Fishpowered
-    Hawk Helm                                                                       Neck Romancer
Gul rune from /players3.
 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

What if your ultimate goal in running Kurast was to get lots of gg charms/jewels and gems? Would there be an expected diffenece in drops on /players 3 vs /players 8??
 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

What if your ultimate goal in running Kurast was to get lots of gg charms/jewels and gems? Would there be an expected diffenece in drops on /players 3 vs /players 8??
The chests do drop less "stuff" on /p3 than /p8. But since I've switched to /p3, I really haven't noticed a let-down in the amount of gems, jewels, and charms. I still get plenty, and it seems as much as on /p8. I have no hard evidence of the numbers, but it's just a feeling.

Maybe the difference in chest pop amounts from /p3 to /p8 is mostly in gold piles or something, while gems, jewels, and charms stay constant? Because like I said, I still get plenty.



 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

Maybe the difference in chest pop amounts from /p3 to /p8 is mostly in gold piles or something, while gems, jewels, and charms stay constant? Because like I said, I still get plenty.

I believe that there is some truth to that. The special chests ALWAYS drop equipment (gear/charms/jewels), its part of the special chest algorithm.

I'm certain that p7 drops more stuff that p3 but especially the items part (charms and jewels) will be muted because the cycles will continue until they drop something, countering some of the effect of higher no-drop rates.

I would be tempted to argue that I've seen less runes overall but seriously, I've only done a few hundred p3 runs and couldn't possibly make an assumption about the mechanics with such a tiny amount of data.



 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

I did 1k p7, and 1k p5 LK runs. 1k p3 in progress. I documented every rune i pick up: Tal and better plus Nef. There seems to be more runes overall on p7(39) comparing to p5(24). I will have better results when i finish 3 series of 1k*(p3+p5+p7) runs.
 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

Sorry, it is correct.

I can't possibly be bothered to prove that the game works the way its supposed to when the evidence you have is only observation over ~150 runs. Please feel free to assume those chests are Special and do 100K HotD runs.

I'm going to have to disagree with you, both about the Halls of Pain chest being Sparkly and there not being enough data. I ran a known Sparkly Chest (Act3 Sewers 1) and the apparently Sparkly Chest in the Halls of Pain 50 times each and counted the number of drops that would be impossible for a Special Chest. These include set/unique drops as well as any drops containing non-magical/non-rare weapons or armor.

From the Sewers Sparkly Chest, 12 drops were Special Chest incompatible (and 38 were not obviously distinguishable). From the apparently Sparkly Chest in the Halls of Pain, 0 drops were Special Chest incompatible, i.e., the chest was at all times indistinguishable from a Special Chest. According to a Fisher Exact Probability test, the probability that the two chests are both truly Sparkly and that this distribution is the result of chance is 0.00011 (one-tailed). I think that level of significance would satisfy any scientist. At the very least it calls into question the rule of thumb that all chests that sparkle behave as Sparkly Chests, at least in Act 5.



 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

jjscud, I don't mean to argue with your knowledge of chest-hunting, but I remember we discussed this topic once before:

http://www.purediablo.com/forums/showthread.php?p=6458126&postcount=18

After looking at the screenshots, I'd risk saying it's possible that
(a) what sparkles in Act 5 could well be a special chest
(b) Act 5 mechanics is simply different
(c) we don't have enough data points

Would be interesting to find out. BTW: I think nice method of proving that would be taking screenshots of 'rare' drops. And compare between people. (It's only 2% of situations, which is not all that many). Or maybe it's a stupid idea? Need my :coffee:

--Greebo
 
Last edited:
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

Strictly speaking, according to the drop calc, you get 1/3 less "stuff" from LK chests on /p8 compared to /p3. So 1/3 less gems, 1/3 less jewels, 1/3 less charms, etc. In other words, you are giving up 1/3 of all of your regular /p8 drops in order to go for Lo and Gul (and might be giving up Ohm, Vex and Mal as well, unless they can be found). In 11k runs, that's a LOT of stuff. A small upside is that the giant chest drops are probably a bit less crowded, so less good runes will get lost in the pile. Now, if we can prove that there are MULTIPLE instances of a HR in /p3, it would be totally worth it. Otherwise... I don't personally think it's worth it.
 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

And compare between people. (It's only 2% of situations, which is not all that many). Or maybe it a stupid idea? Need my :coffee:
--Greebo

Hm or rare drops with one particular item (say perserved head or something similar, since theese are 3 times rarer). But we can as well run those for runes.

Problem here is, it would actually require someone to run them.

btw> its 5% of situations I believe ... I guess running say 1000 runs (5 people x 200) would be enough to see the ratio of magical/rare items and if it would be those 5:95 we can safely predict those are really special chests.

Oh my, if true, it would definately throw chest running upside down :jig:


 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

OK, since it's not an early morning anymore, I might be able to express myself clearly.

About the 'rare drops' (i.e. 'all yellow'):

I'm not sure whether (rare_drop/all_drops) ratio is 1:20 or 1:50, I somehow remember 1:50 (might well be wrong, too lazy to check atm), but that doesn't matter all that much. The point is, it's a low ratio. Assuming 65536 drops, that's 1300 or 3200 drops with rares. Those are relatively small numbers.

The nice thing is, probabilities work kind of funny in such situations: In a classroom full of people (say 40), the probability of 2 people having birthday on a same day (1 in 365) will be over 90%.

So if I were to check that a chest is a special chest and that seeds repeat themselves, I'd keep taking screenshots of rare drops from it. There should be a 'double' rare drop fairly soon (in comparison with waiting for a rune-drop to repeat itself anyway).

That doesn't necessarily mean it's a smart thing to do, of course. Just something I came up with when half-awake.

--Greebo

@Milb: Good to know. I haven't seen such screenshot, thus my confusion and lack of knowledge.
 
Last edited:
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

OK, since it's not an early morning anymore, I might be able to express myself clearly.

About the 'rare drops' (i.e. 'all yellow'):

I'm not sure whether (rare_drop/all_drops) ratio is 1:20 or 1:50, I somehow remember 1:50 (might well be wrong, too lazy to check atm), but that doesn't matter all that much. The point is, it's a low ratio. Assuming 65536 drops, that's 1300 or 3200 drops with rares. Those are relatively small numbers.

The nice thing is, probabilities work kind of funny in such situations: In a classroom full of people (say 40), the probability of 2 people having birthday on a same (1 in 365) will be over 90%.

So if I were to check that a chest is a special chest and that seeds repeat themselves, I'd keep taking screenshots of rare drops from it. There should be a 'double' rare drop fairly soon (in comparison with waiting for a rune-drop to repeat itself anyway).

That doesn't necessarily mean it's a smart thing to do, of course. Just something I came up with when half-awake.

--Greebo


I don't think that that is the way it works.

The chest will decide which items drop, then there is a 1 in 20 chance that the items will drop all rare. I'm sure I have seen a picture with a Ber/Sur with rares but the same item types, this doesn't mean there is a extra drop for either of those runes.



 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

Would be interesting to find out. BTW: I think nice method of proving that would be taking screenshots of 'rare' drops. And compare between people. (It's only 2% of situations, which is not all that many). Or maybe it's a stupid idea? Need my :coffee:
This is one way to do it, but a much easier way requiring much less data is to test the hypothesis that the chest is Sparkly, because that is falsifiable. All you need is one run in which the chest drops something non-magical or non-rare, or rares and magicals together. And keep in mind that the chance for uniques and sets is not the only way in which the two chest types differ. In my little study I looked for drops of low quality and otherwise white or gray weapons/armor. My sample size was too small to get a precise estimate of the actual probability of the Sparkly chest dropping a "white" item, but it seems to be between 1/4 or 1/5. More precision than that afforded by higher sample sizes is unnecessary. All that is required is that the probability is relatively high, because that makes any streak of more than a few dozen runs without such drops extremely unlikely.

I think this is a very important question (relatively speaking of course :wink:). First, because I think the common knowledge might be wrong in this case. And also because if Act 5 Sparkly chests have a 2% chance of dropping uniques (and Sparkly Chests seem to drop multiple uniques - I've seen up to 4 but I think they can go much higher), than running them for TC87 uniques might be even more lucrative than for the elusive ZOD (assuming those chests could theoretically drop items from that TC). The Halls of Pain chest has a fixed location about 6-7 straight-line teleports from the WP. I did fifty runs in about 20 minutes. Now, I don't think it will ever drop a unique, but if I've just been incredibly unlucky, I should continue to do as many runs as possible.

EDIT: I looked back over the previous thread that touched this issue and discovered I had posted a comment on this exact same issue. I had reported data from 180 runs, so now I've done 230 runs and never gotten a drop that couldn't have come from a Special Chest. The Halls of Pain chest is a Special Chest. It's probably a good bet that all chests in Act 5 that look like the Ark of the Covenant are Special Chests, even though they sparkle. I wouldn't be surprised if Sparkly Chests don't even exist in Act 5.



 
Last edited:
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

I believe that there is some truth to that. The special chests ALWAYS drop equipment (gear/charms/jewels), its part of the special chest algorithm.

Are you sure of this? I had this drop (pic1, pic2) 10 minutes ago, and a similar one this morning. All whites, no rare or magic items.

It was on p3.


 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

Are you sure of this? I had this drop (pic1, pic2) 10 minutes ago, and a similar one this morning. All whites, no rare or magic items.

It was on p3.

They are still there, simply moving a little isn't always enough to show them.



 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

I actually picked up some gold and potions to make sure there wasn't anything behind. That was all that dropped.
 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

Well, quite the day.

I've ran a lot of Act 5 Sparkly chests and dug through the best special chest descriptions I can find.

First, the all white drop, the special chests drop 11 or 12 cycles (unlocked or locked) but they abort after something magic or rare drops. It is however possible that all cycles do not drop a magic/rare item and then the process ends. I've talked to a few others and looked over some things and I would guess that this never happens at p7/p8 but yes, it does appear possible.

Second, I must agree, at least some of the sparkly act 5 chests are special. I blame Blizzard for doing things bass-ackwards in act 5. That probably means a whole new Act 5 chest thread needs to be made and a lot more tests done to determine all the special chests. (Not to mention even more confusing descriptions for "special chests, sparkly chests, and super chests".)

Now, anyone seen the foot in mouth smiley...


While reading, I ran across a few other things.

First, as I mentioned above, special chests continue drop cycles (to a limit) until magic or rare items drop. This means two things.

1. As I already posted, player settings will be muted in special chests, if there's a no magic drop, it just adds another cycle. We really can't assume that you get less charms/jewels just because the player settings are lower. I suspect that may still be true but the differences should be much smaller than they would with normal chests.

2. Lower player settings means higher chance of a no drops, obviously. In normal objects/monsters thats the end of it BUT with Special Chests that MIGHT mean more cycles on average which MIGHT mean more junk Tc AND Good Tc and that means MORE runes, not less.

It all depends on how it plays out in the numbers but there is a very real possibility that lower player settings will produce more runes.

Well, that's all I can glean from Urlik's guide, I think I will see if I can find some better sources for special chest drops.
 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

Well, that's all I can glean from Urlik's guide, I think I will see if I can find some better sources for special chest drops.

Good luck with that jjscud! That is a pretty indepth, and complicated writing!



 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

The special chests drop 11 or 12 cycles but they abort after something magic or rare drops. Lower player settings MIGHT mean more cycles on average which MIGHT mean MORE runes, not less.

Now that's fascinating. I believe that this hypothesis is supported by the current evidence that /p3 appears (so far) to have all of the /p8 rune drops, and more, which seems very improbable otherwise.

This also explains how Sur & Ber can exist on /p3 and /p8 yet have completely different magic items associated with them. The magic items are generated after the runes, so it's likely that the /p3 and /p8 Sur and Ber drops start identically, come from identical seeds, and diverge when the drop shifts into magical items. The seed which produces Lo on /p3 fails to produce Lo on /p8, because a shift into magical items stops it from doing so. And if the special chests are dropping 11 or 12 (why 11 or 12?) times then I'd have to rework all my probability calculations as well.



 
Re: LK /p3-4 Superchest Drops (The Search for Lo Is Over)

2. Lower player settings means higher chance of a no drops, obviously. In normal objects/monsters thats the end of it BUT with Special Chests that MIGHT mean more cycles on average which MIGHT mean more junk Tc AND Good Tc and that means MORE runes, not less.
I've been absent for quite some time. I'm glad to see these findings somewhat validate my past need to investigate other players settings. It also seems that its smarter to run on p3 than p7 because of ber, sur and lo being available.



 
Back
Top