LK 1.13beta Special Chest Drops

[jjscud]

LK 1.13beta Special Chest Drops

With the rune drop odds changed in 1.13 there's (well a given at this point) that we will be looking at new drops.

Because the drop rolls happen in rounds for each rune pair tc, and because TC roll down odds were unchanged for the lower rune tcs, there should be no new patters for the runes Um and lower. Above that though, who knows, its very possible to see multiple new hr drops like there is in RoF.


I'll add my confirmed existing patters later when I have more time.

Rune  /pX Item#1              Item#2              Item#3              Item#4              Poster
Io
-

Lum
-
 
Ko	    
-

Fal
-
 
Lem
-
 
Pul
- 
 
Um
-
 
Mal
-

Ist
- 
 
Gul
-
 
Vex
      8   Flanged Mace        Heavy Belt          Heavy Crossbow      Suwayyah            Post #140

 
Ohm
-
 
Lo
-
 
Sur
     
 
Ber
-

edit: Thanks nubikoen, I deleted Jah and Cham somewhere along the way but must have went back after.

 

[Nubikoen]

Re: LK 1.13beta Special Chest Drops

You've added runes up to Cham in the table. They don't drop from objects in act3, do they? (Ignoring the truncated seed, that is)

Edit. I was a bit unsure since you had LK in the title, but also mentioned the RoF in the post.

Anyways. I've restored yet another character to interfere with my HC career, and have done just over 100 /p3 runs. No good stuff yet.

 

[Kimppi]

Re: LK 1.13beta Special Chest Drops

Any plans on checking out RoF too? I rolled a map there, ran it a little on p7 and found this. Doesn't look like to fit the previous patterns. The shot is pretty terrible but I was trying not to get killed.

After S&E and looking at the picture I realized that I could have just killed the mob and cleared the ground a bit. On top of that I forgot to pick up the rune.:whistling: Well, good thing it was "just" a Mal.

I also did 100 or so p3 LK-runs but only found an existing Ko-pattern.

 

[Kefir-Tribe1]

Re: LK 1.13beta Special Chest Drops

I've a question here, perhaps a bit OT. If the new patterns give roughly the same chance of getting desirable runes as in 1.12, will LK running still be the best way of farming them, considering the increased drop rates for monsters?

 

[skiffcz]

Re: LK 1.13beta Special Chest Drops

I've a question here, perhaps a bit OT. If the new patterns give roughly the same chance of getting desirable runes as in 1.12, will LK running still be the best way of farming them, considering the increased drop rates for monsters?

No. Here is why.. Also note that counciliers have drop chance up to Cham. I recomend the original pre 1.12 thread, highly interesting stuff.

 

[Fabian]

Re: LK 1.13beta Special Chest Drops

jjscud, is there any sort of way to predict which player setting would have a higher chance of having more high rune drops? Or is p7 and p1 just as good in theory at this point, for instance?

 

[sirpoopsalot]

Re: LK 1.13beta Special Chest Drops

I could be wrong on this, but isn't it somewhat probable that the chest patterns are unchanged? If Blizzard didn't change the mechanisms for objects-dropping-runes, then I can't see how changes to the odds of high rune drops would supersede that. I mean...

... the patterns are already consistent and they don't conform to the rune-drop odds anyways. So what's changed?


Of course, that's contrary to what Kimppi's Mal implies , though it could just be a previously undiscovered pattern (or he could be a major haxxor ).

 

[maareek]

Re: LK 1.13beta Special Chest Drops

@poops: That's more or less what I thought after reading about the new patch. Best I can figure it maybe the new formula introduces more possible drops that include those runes across the board.

For example (just for example, these numbers should be considered meaningless), in 1.12 there are, say, 2 outcomes that would produce a Ber then in 1.13 there may be 6 of them. In that case, assuming these changes are across the board and not just for drops from monsters (I would guess this would be the case but I have no clue), 4 new drop patterns would have to be introduced. I don't see where the original 2 would be different, but maybe my not-very-math-inclined brain doesn't understand the full scope of those changes.

 

[Ohomemgrande1]

Re: LK 1.13beta Special Chest Drops

I believe the chance for new drops goes something like this:

The RNG hasn't changed. There are ~65k possible sequences for the special chests. What has changed are the drop odds for the top runes. The runes drop something like this:

The game decides to drop a rune and starts at some TC depending on area. rune17 contains zod and all rune TC's below 17 contain 2 runes or a drop to the next lower TC.

There are 5 chances to drop a rune in each TC and a generally larger chance to drop to the next TC. The chance to drop in TC is what has changed in the table.

So, let's say that in patch 1.12, the RNG rolled a number which translated to 6 out of a possible 100 choices (5 for rune, 95 for next TC), then it would move to the next TC. Now in 1.13, the 95 has dropped to say 45. Now, the same RNG number that translated to 6 before now translates to 3. Instead of dropping to the next TC, the game drops a rune at a higher TC than in 1.12.

Therefore, we will definitely see most of the old drops but there is a chance to see new ones. I wouldn't be surprised if we actually see drops where everything is the same but the rune has upgraded to the "higher" of the two runes in the same TC. i.e. Ber instead of Sur.

We just need to do a lot of runs to find out

 

[maareek]

Re: LK 1.13beta Special Chest Drops

Therefore, we will definitely see most of the old drops but there is a chance to see new ones. I wouldn't be surprised if we actually see drops where everything is the same but the rune has upgraded to the "higher" of the two runes in the same TC. i.e. Ber instead of Sur.

Isn't that somewhat unlikely since each rune in a TC has been increased by the same amount? Seems more likely that more instances where 1.12 would have dropped to Gul or Ist, etc, would stop instead at Ber/Sur or Lo/Ohm. Which is still a very nice improvement.

Or is there some part of the selection process that I'm missing (I'm assuming that the process after a rune is chosen to drop is roughly the various formulas are put into place giving you an integer that either correspondes to a Sur, a Ber or a drop-down sequence, which wouldn't be affected by the overall chance because both Sur and Ber have had the same increase in [ie, if there are now 5 checks which correspond to a Ber there are also 5 checks that correspond to a Sur])?

 

[jjscud]

Re: LK 1.13beta Special Chest Drops

Therefore, we will definitely see most of the old drops but there is a chance to see new ones. I wouldn't be surprised if we actually see drops where everything is the same but the rune has upgraded to the "higher" of the two runes in the same TC. i.e. Ber instead of Sur.

Nice explanation Ohomemgrande.

It does depend a bit on exactly how its coded but logically, that makes sense. The old HR patterns should remain with some new ones. However, the order in the TC file is low rune, high rune, change to roll down. This means that it should be the higher runes that may shift down to the lower runes.

My short version: The truncated seed does not free us from the probabilities, it simply makes the sample size too small to show a definite pattern. The special chest drops are still governed by the probabilities.

Isn't that somewhat unlikely since each rune in a TC has been increased by the same amount? Seems more likely that more instances where 1.12 would have dropped to Gul or Ist, etc, would stop instead at Ber/Sur or Lo/Ohm. Which is still a very nice improvement.

No, not at all. Each level (runes 14, runes 13, etc...) should be its own selection. The fact that some drop ended on Gul tells us absolutely nothing about the rolls that happened the two rounds before so we can't guess anything. Well, actually we can, we can guess that since what happened in later rounds doesn't tell us anything abut the early rounds, then probability rules. The new hr patterns will most likely be evenly distributed and so, mostly from the low rune patterns that we've never bothered to track.

 

[maareek]

Re: LK 1.13beta Special Chest Drops

By that I meant it seemed more likely that a Gul/Ist or lower process would be more likely to become a higher TC rune drop than that a Sur would become a Ber or an Ohm become a Lo. In other words, that the odds are better of a lower TC rune stopping instead at a higher TC than a rune pattern that used to be a Sur should become a Ber.

Your point that the lowest rune in the TC gets first check makes it even more likely that the lower rune patterns known won't become the higher runes in their TC. Right?

I'm not trying to say that there are good odds of lower rune drops rolling as higher runes, but that the odds of old patterns that resulted in the lower rune in a TC changing to the higher rune are very low (I was basing this on both runes checked at the same time; with the lower rune being checked first I think it becomes even less likely).

Granted, I may still be missing something, or my word selection may be lacking, but I'm not seeing where there's much chance of old rune patterns getting an upgrade within their own TC.

 

[RegalStar]

Re: LK 1.13beta Special Chest Drops

I remember rune drops work something like this: A rune TC that contains up to Vex will go like:

Vex: 2
Gul: 3
(Lower rune TC that goes up to Ist): [a large number]

Blizzard could have easily raised the rate of the high runes directly, or lowered the pick rate of a lower rune TC, or both. If they did lower the rate of picking a lower rune TC, depending on how Diablo's RNG works a pattern that used to pick Vex may very well pick a lower rune TC post 1.13.

 

[Cius]

Re: LK 1.13beta Special Chest Drops

Before doing 10K runs to check I would recommend waiting for the final 1.13 patch. Imagine doing all that work and then they tweak the numbers some more in the final release! Still, the possibility of this changing is far more interesting to me than just about any other change they implemented.

 

[Kefir-Tribe1]

Re: LK 1.13beta Special Chest Drops

No. Here is why.. Also note that counciliers have drop chance up to Cham. I recomend the original pre 1.12 thread, highly interesting stuff.

That’s very interesting!

In his calculations Fabian makes the approximation that if you have a say 1/3000 chance to get a HR in a run, then you will find one HR on average each 3000 runs you do. My knowledge in probability calculations is really dusty, so I might be talking gibberish here, but isn’t that a bit off? Shouldn’t 3000 runs give a 1 – (2999/3000)^3000 = 63.2 % chance off acquiring said rune? Which mean fewer runes on average?

How about good old Cow level or perhaps Chaos Sanctuary now when IM is out of the picture? Maybe updating Urlik’s guide with the new numbers would answer these questions. Sorry for the OT.

 

[Nubikoen]

Re: LK 1.13beta Special Chest Drops

That’s very interesting!

In his calculations Fabian makes the approximation that if you have a say 1/3000 chance to get a HR in a run, then you will find one HR on average each 3000 runs you do. My knowledge in probability calculations is really dusty, so I might be talking gibberish here, but isn’t that a bit off? Shouldn’t 3000 runs give a 1 – (2999/3000)^3000 = 63.2 % chance off acquiring said rune? Which mean fewer runes on average?

That's true if you only do 3000* runs total.

When talking averages, you're looking at 1 every 3000* runs.

*I went with the 3000 you used, I haven't checked if that's accurate.

 

[Fabian]

Re: LK 1.13beta Special Chest Drops

Kefir, I'm talking about expected value in those posts. During 3000 runs, there's a ~63% chance you'll have found at least 1, 37% chance you've found 0. Sometimes, you'll find more than 1. On average, you'll have found a number of runes equal to 1.

Like you say, two things can be calculated. How many runes you'll find, on average, over a long time (expected value, EV), or the chance that an actual high rune will drop in X runs.

For instance, in my high rune spreadsheet, there are two formulas for 1.12a LK p3 running. The first one calculates the chance of finding at least one rune in X LK runs:

1 - ((65533/65536)^(X*6))

This is the forumla you're thinking of (although adjusted for LK; we're in an LK thread after all ). The chance of NOT finding a high rune is 65533/65536 for any given chest, and each run we open 6 chests. When I plug in X = 200, I get 0.0535, or 5.35%. When I plug in 3641, I get 63.21%.

The second formula calculates what the expected value is in X number of runs:

(3*6/65536) * X

For each run, there are 6 chests with 3 good drops, out of 65536 total. After 1 run, you will have found an average of 18/65536 high runes, and this scales with each run you do. So after 3641 runs, you'll have found 1 rune, on average. After 10923 runs, you'll have found 3 runes on average. This is the same thing I'm talking about in the Travincal thread.

Any other questions/comments about it, please ask.

 

[Kefir-Tribe1]

Re: LK 1.13beta Special Chest Drops

Ah, okay that makes sense. I should crawl back into my cave now.

When the 1.13 LK patterns are clear it will be interesting to see what the best rune target is. I sincerely hope it's one where you get to fight monsters.

 

[Shagsbeard]

Re: LK 1.13beta Special Chest Drops

The "best" method will still be the one you can tolerate doing over and over. That factor has far more significance than people give it credit. Even if the council method is twice as fast, if you can't tolerate doing as many runs as you could in LK it might not be as practical.

 

[Ohomemgrande1]

Re: LK 1.13beta Special Chest Drops

However, the order in the TC file is low rune, high rune, change to roll down. This means that it should be the higher runes that may shift down to the lower runes.

Ah yes, you are correct. That would be sad.