Spiral Knights - Let's science: Black Kat spawn ratio

Let's science: Black Kat spawn ratio

24 replies [Last post]

Mon, 09/22/2014 - 06:58

Mtax-Forum

As Autofire pointed in other topic - RNG Manipulation may work on Black Kat spawn chance. This topic was created to collect data from multiple users and compare which way to farm is better and I think it's worth checking. I think that number of Kats killed is raising chance of Black Kat spawn and it decrases/resets when you meet Black Kat.

Why Three Rings would do that?

To limit speedrunners. If you speedrun something - You are supposed to take minimal loot. And for now technically only speedrunning matters - And that's kinda cheating.

Did Three Rings used RNG Manipulation before?

Yes and I'm 100% sure, but I won't say where and what you can gain with that, because of marketing.

Okay, so how can I help?

Basically I need data of two types of T3 players that grind pages solo - Those who speedrun and those who play in normal way. For our testing place let's use Tier 3 Candlestick Keeps, because it's most runned place with most number of Kats. And let's use event prestige mission as resetting factor, because you always meet here Black Kats, so it probably "cleans" your chances.
All of yours job is to calculate your averange numbers of floors you have to make to meet Black Kat, post stratum where you farm and on which style - Speedrun or normal killing of everything. This will allow us to compare which option is actually better to farm pages and not waste time.

Thanks for everyone who will contribute.

Mon, 09/22/2014 - 07:23

Bopp

regular kats between black kats?

If your hypothesis is that killing regular kats "fills a meter" that causes a black kat to spawn when it's full, then maybe you should ask people how many regular kats they will between each random black kat spawn? Just a suggestion.

Mon, 09/22/2014 - 07:37

Skepticraven

↓

From my past observations, both the number of kat spawns and where they spawn affects if they turn into a black kat. Specifically in Scarlet Fortress, only certain kats would transform into black kats. I attempted to use this depth last event to study the rate at which kats turned into black kats because each depth had the same number of kat spawns.

This is the data I took. Despite encountering over 1k kats [not all runs were recorded], my BK encounters were limited to only two spots in one type of the scarlet fortress levels - specifically grim gallery where status kats spawned [start and end of depth].

Mon, 09/22/2014 - 08:23

Mtax-Forum

@Bopp: That's an option, but I guess that most of people (like me) are too lazy to count all those kats, so I prefered to keep floor as a counter.

@Skepticraven: You are on something here. There is possiblity that some kats are not able to "transform" into BK or have less chances for it, but idk for now. Maybe I'll check this stuff by the way later.

Mon, 09/22/2014 - 08:27

Gbot-Vtwo

Dude.

"@Bopp: That's an option, but I guess that most of people (like me) are too lazy to count all those kats, so I preferred to keep floor as a counter."

If you are too lazy then you might never find the answer.

Tue, 09/23/2014 - 13:07

Flawedknight

As far as I can tell; it doesn't matter

I've been only speed running and so far over the event have come across dozens of black kats. So far -- when I've been metering -- it seems I get 1 BK per every 75 or so regular kat spawns (averaged based on about 3000 kat spawns). Mind you it's just the running average I have so far; sometimes I'll see them on back to back runs or two in the same run, sometimes it'll occur 10 or 15 runs. I've been trying to maximize the number of kats I can see spawn in a minute. So far the best average I have is roughly 5 KPM (Kats Per Minute) when party load times and other values are factored in.

Please note that only a few hours have been tracked and I've calculated the total number kats spawned based on two hours of counting cats * the number of hours played. Black cat spawns were all counted. Certainly not definitive data

Oddly enough it seems more likely that a block will randomly be a jelly lol

Wed, 09/24/2014 - 13:20

Cheshireccat

Here is all of the data you requested

Estimation:_____________________________
I can help you with your "too lazy to count all those kats". Between last year and this year I counted 104 Candlesticks. The average number of Spookats-per-Candlestick is 20.94 (let's call it 21).
Please note that this average can only be used to estimate against LARGE SAMPLE SIZES. My observed Spookat-per-Candlestick range is from 9 to 30. Results from a small sample size may reflect a series of small Candlesticks or a series of large Candlesticks, thus biasing your data.

Speed Running (with boxes):______________
http://forums.spiralknights.com/en/node/100847
This is my data from the last event. It was all done with speed running and only stopping to pop boxes.
I have kept track of my data from this current event. It is completely in line with the last event--it is safe to accept that nothing has changed between the two events.

Speed Running (without boxes):____________
Not published yet. I have 201 runs collected this event. I anticipate I will have 300+ unless I get a book in the next few days.
Check the forums late next week.

Killing Spookats:_________________________
I do not have this data.

THE PROBLEM WITH YOUR RESEARCH:___
This is something I have hit upon myself: Black Kats are too rare. You are going to need a sample size of over 1200 in order to reduce the sampling error to less than 0.5%.
From the 5/14 Katastrophe: My BK rate was 1.4% (1:72 Spookats) but my standard error was 0.9%. That means the BK rate is somewhere between 1:44 and 1:199.
From the 9/14 Katastrophe: I do not want to share my incomplete data but my standard error is 0.7% at the moment. That is also a HUGE ratio.

The only way you could prove there is a difference between WITH killing and WITHOUT killing is if your standard error ranges are exclusive. Ex: without killing is 0.6% to 1.4% and with killing is 1.8% to 2.5%.
With BKs being so rare OOO would need to do something huge like triple the BK from killing vs not killing. Otherwise... you will need thousands and thousands of samples each way.

Thu, 09/25/2014 - 06:33

Jendar-Nox

...

> From my past observations, both the number of kat spawns and where they spawn affects if they turn into a black kat.

My own observations seem to indicate that only a themed Kat can become black. Candlestick Keep has a lot of predefined spawn patterns, and I never saw a black Kat appearing in places where you usually find normal Spookats.

That might also explain the results you received in Scarlet Fortress. If I remember right, Grim Gallery is the only place where you have Pepperkats in addition to Spookats, and they indeed spawn in the beginning and in the end of the level.

Thu, 09/25/2014 - 08:12

Fehzor

I wonder if that's actually a bug, the spawning over only themed kats.

Thu, 09/25/2014 - 08:42

Cheshireccat

A bug?

God I hope so.

Have you reported your theory to support? Might be worth filling out the form.

--Chesh

Thu, 09/25/2014 - 09:37

#10

Jendar-Nox

...or a feature

Most probably it's just the way the algorithm works. First it decides whether a Kat will be normal or themed, and if it picks themed, it chooses between depth theme and blackness. After all, blacks are themed Kats too (Curse-themed, to be exact).

Fri, 09/26/2014 - 04:56

#11

Skepticraven

↓

@Cheshireccat

I have not. I find it unlikely that any efforts will be made to make this rare event item to appear more common.

@Jendar-Nox

I specified the location of the spawns: specifically grim gallery where status kats spawned [start and end of depth].
I will also note that there is a possibility for kats to spawn in random monster boxes. I recall posting stuff last event with no one really catching onto my data that disagreed greatly with the 1% rate. In fact, one of the comments I made can be found here.

I was not interested in taking the effort to divide up a candlestick keep by types of kats. I would estimate that the frequency of a black kat appearing is around 2%, but limited status kats only. Candlestick keeps have roughly 50% status kats and 50% normal kats, thus giving the appearance of the 1% spawn rate.

@Topic

Scarlet fortress and arenas have either none or almost no status kats. Clockwork tunnels and candlestick keeps have around 50% status kats. Some dark cities have almost 100% status kats, but significantly fewer spawns (due to fiend themed instead of undead).

Edit:
Upon further inspection of dark cities, the status kat spawn rates are closer to 50%.

Thu, 09/25/2014 - 12:03

#12

Avihr

Hmm, i've fought worse

I have to say that my first 2 black kat encounters have been in a scarlet fortress where supposedly there is no themed kats, Cravat hall 2. (actually not sure if it was cravat hall 1 or 2 but none of those have themed kats anyway) and I also would like to mention that those 2 i found were also on the same level. haven't found any more in scarlet fortresses since that time though.

Thu, 09/25/2014 - 12:25

#13

Cheshireccat

@Skeptic

For all of our sakes, you really should try. If it can't hurt and might do a world of help... pretty please with sugar on top? :D

One of the best things about pudding is that you can share it. So go share some with SK Support. ;)

Thu, 09/25/2014 - 13:23

#14

Enfeebler

Lots of number, yay (? ._.)

I really doubt that killing monsters or Kats specifically has any influence on spawn rates (it'll kinda explain itself later in post). The other theory, with Black Kats only taking place of status Spookats could be true (doubt it as well, to be honest). Easiest way to check it is to find a normal-themed area or a room with no chance for status Kat to appear and grind it 'till the Black Kat appears (or not). For how long? You need 500 Kats to have 99.3% chance of getting 1 (or more) Black Kats (if the rate is 1%).

Since I don't think that this^ helped in any way I'll just share my data from the current event (yes, I've speedrunned through 1400 CSK depths...), maybe this will be helpful. Unfortunately I was too lazy to count all the Kats that spawned (just from time to time, out of boredom, I didn't write anything down).

__________________________________________________________________________
(9-2) Alone in the Dark (1st depth only = D24, shock themed Candlestick Keep, was playing solo, every single time)

(Day: Number of BKs/Amount of depths completed, percentage)
Day 1: 11/50, 22%
Day 2: 53/190, 27.9%
Day 3: 38/200, 19%
Day 4: 39/200, 19.5%
Day 5: 49/170, 28.8%
Day 6: 54/190, 28.4%
Day 7: 32/150, 21.3%
Day 8: 37/150, 24.7%
Day 9: 28/100, 28%
Overall: 341/1400, 24.36%
_____________________________________________________
Books: 1/341, 0.29% - 246th Black Kat/1012th depth (during day 7: 2nd BK/12th run)
1x BK: 273/1400, 19.5%
2x BK: 31/1400, 2.21%
3x BK: 2/1400, 0.14%
_____________________________________________________
0x BK: 1094/1400, 78.14%
1+ BK: 306/1400, 21.86%

[N]x BK - Amount of depths with N Black Kats
1+ BK - Amount of runs with 1 or more Black Kats, or simply amount of "successful" runs.

I used Bernoulli distribution to try to get the amount of Kats per depth. I did it in the simpliest way (could be wrong, you'll judge), took the result from "1x BK" line (273/1400, 19.5%) and tried to find the number of Kats required to get same/similar percentage. (For example: 1 success [1 BK] in 20 tries [20 Kats in depth] gave 16.5%). Turned out that the closest number was 25. Unfortunately it just means that I have been getting around 25 Kats per depth(19.6% vs. 19.5% from my data). With 1400 depths completed the number I got could be actually pretty close to what it is in general though (for Candlestick Keep). How close is it to the truth? No idea.

Then I took the number (25) and compared it with the rest of my data. I did it just to verify how close my numbers are to those from binomial distribution.
0 BKs in a group of 25 Kats : 77.78% vs. my 78.14% (1094/1400)
2 BKs/25 Kats: 2.38% vs. my 2.21% (31/1400)
3 BKs/25 Kats: 0.18% vs. my 0.14% (2/1400)
1 or more BKs/25 Kats: 22.22% vs. my 21.86% (306/1400)

My numbers for 2 BKs and 3 BKs are not too accurate; getting two BKs 31 times and three BKs twice is obviously not enough to get the accurate number, yet they're still pretty close to what the binomial distribution says, in fact all are almost spot on. It still doesn't tell us more than: I've been finding ~25 Kats per depth and the chance to get a Black Kat seems to be around 1% (that's the probability of success(BK) that I used for the Bernoulli distribution). Well, this and that the data is quite legit, which was the point of comparison.

It's much worse with books though, due to its random, really random nature it was hard for me to get enough data (0.5%(?) on top of 1%, cruel joke by OOO ._.). I've found 1 Book after killing 341 BKs, which gives ~0.3% (0.29). This 0.2% difference is quite big, for example the chance to get 1 or more books after killing 341 BKs: 64.1% (0.3% rate) vs. 81.9% (0.5% rate). All I can say is that there's very slim chance that drop rate is (much) bigger than 0.5%. One of the drop rate that we believed to be true was 1%. Well, after killing 341 Black Kats I should have 85.56% chance to get 2 or more books, possible, but not really likely. So I'd go with 0.5%, the number seems to be pretty accurate (and how could Spiral-Spy be wrong? :3).

Well, that's all I've gathered. You can use all the data freely, maybe you'll find better ways to get some conclusions, informations or simply anything interesting. There's still few days left and I'll keep grinding, not as much as before, but still - quite a bit, hoping to get 2nd book (maybe more? :3). I'll update this post after the event's done, with complete numbers.
___________________________________________________________________________________________
TL;DR: After 1400 Candlestick Keep depths: I've been getting ~25 Kats per depth with ~22% chance to get 1 or more Black Kats. Additionally I've killed 341 BKs and got 1 Book of Dark Rituals giving the 0.3% rate, anyway I still believe that 0.5% drop rate is the real one.

Thu, 09/25/2014 - 13:23

#15

Skepticraven

↓

@Aviri

If you looked at my data, none of the cravat halls were in my data. Spiral Courts were and have 0 status kat spawns. Grim Gallery was as well and has 6 or 7 if I recall correctly.

Cravat hall III is used in the owlite mission and has 0 predetermined status kat spawns. However, that is why I added the extra note in my last comment - random monster boxes have a chance of spawning black kats. Cravat halls each have a significant number of random monster boxes. In fact I've seen strange things come out of these monster boxes.

Sun, 09/28/2014 - 09:12

#16

Ankung

I was going to save this for after I got my black Seraphynx...

...but I'll just tell you about my discoveries now.
The first day I tend to find more Kats than usual
The chance of a Kat appearing skyrockets when I play in a party
The less ...erm..."grinded" a level have been, the greater are the chances of a black Kat appearing, thus I always seem to find a Kat at the current prestige mission but never in "Alone in the Dark"
Some players are "luckier". Partying with them will result in lots of them to appear.

Conclusion:
If you, on the first day, party with three other "lucky" players at "Scared to Death", you might just replace every normal Kat with Black Kats.

Also, Scepticraven, I found a black Kat in a normal Clockwork Tunnel - Haunted....Passage? While your theory might be wrong, it wouldn't surprise me anymore if the developers went through every single knight and manipulated the random numbers differently for EVERY. SINGLE. KNIGHT.

Sun, 09/28/2014 - 10:04

#17

Skepticraven

↓

@Ankung

Interesting, but not my theory. There are spawns all over depths. Of these spawns, they are either designated to a normal kat or a status kat. In the case of a status kat and the depth having no status designated, a normal kat spawns.

My theory is that those normal kats (on a designated status kat spawn) can also become a BK. Additionally, the kat spawns that are restricted to be only normal will never transform into a black kat.

Sun, 09/28/2014 - 10:05

#18

Bopp

data

Ankung, you're making some pretty strong claims about the distribution of black kats. To be convinced, I need to see some pretty detailed data. Otherwise, I suspect that you're imaging patterns in random phenomena, because human beings tend to do that.

Sun, 09/28/2014 - 10:13

#19

Forum-Brady

Some players are luckier? Aside from the fact that luck is a completely abstract term in which yes: some people are luckier than others, if any player ever somehow found out that OOO was actually giving some players (randomly or otherwise) higher rates of finding event items/monsters, there'd be complete outrage.

It's not something I expect OOO actually code into the game, that some players are simply more likely to see event monsters. And if you're only talking about whether or not the human behind the character is luckier than you are, then...well, the thread is called "let's science", so let's leave that out of it.

Sun, 09/28/2014 - 11:14

#20

Mtax-Forum

Well, thanks for all "science" you posted here so far. Something of this might be true and I'm gonna test few things of this later, though preferably on next event, because of lack of time for now.

I'm still pretty sure that there is possiblity to manipulate BK spawn ratio and searching of this impulse still have sense.

Sun, 09/28/2014 - 12:21

#21

Forum-Brady

When the spawn rates are that

When the spawn rates are that low, and the numbers are all done through RNG anyways, I don't think it'd make any tangible difference if you could manipulate the spawn rate.

Like, if you suss that there is a cap from regular kat killing, at the end of the day you're only ever going to be so much faster than just speedrunning; which has shown to be effective.

Shaving off five minutes from every three or four hours isn't going to save you any real time when you spend more than that just leaning back and yawning or watching loading screens anyways.

Out of interests sake, sure, go ahead and discuss, but I doubt it'll ever make any real impact on gameplay or farming.

Sun, 09/28/2014 - 23:22

#22

Round-Shinigami

I gave up on this even on

I gave up on this even on Saturday, so I can share my info. Sadly, I don't have exact numbers right now, but I've found about 25 kats on elite difficulty on mission 9-2 Alone in the Dark, while killing most monsters and picking out prize boxes, and some 25 kats on 8-1 Weight of Darkness (1st mission in both cases), normal, speedrunning.

In total, my Kat spawn rate was always around 30% (total number of Kats killed - 54, total runs - 178, all number are written from my memory, if you want, I'll post more detailed info later on). Elite killing every Kat and speedrunning on Normal ignoring anything I can ignore spawned me approximately the same amount of Kats (but speedrunning on normal is at least 2 times faster).

In general, if you don't play SK a LOT, this even is not worth your time. Its too limited.

PS: was only running solo. Got 0 books.

Mon, 09/29/2014 - 15:05

#23

Cheshireccat

Done

I might make a handful of runs before the end but I'm practically done.

Keep checking the forums for the next "Cheshireccat's Classes" with my updated BK data. If you managed to collect a few hundred data points you could use it as a comparison

Tue, 09/30/2014 - 09:14

#24

Modicu

My speculations...

I've grinded "Rank 9-2: Alone in the Dark (Depth 24)" on Elite, mostly solo, quite a lot during the time of the event, and I got some decent Black Kat spawns; enough, at least, to get about 3 Kat Tribe Fetishes, with a few ancient pages to spare.

Oddly enough, from what I saw, they seemed to spawn after a certain range number of Spookats, including status-inflicting Kats such as the mission's Statikats, were spawned within the level.

Just as a refresher:

1 Black Kat found within Tier 3 yields 3 Ancient Pages, with a very low chance of a Book of Dark Rituals.
Roughly 17 Black Kats yield 1 Kat Tribe Fetish (50 Pages).
1 Book of Dark Rituals yields 1 Black Kat Cowl for the owner of said book, and a Wicked Whisker to them and allies.
5 - 9 minutes (Speedrun, treasure only, Spookats spawned, not killed) | 8 - 15 minutes (100% Depth 24 completion).

After about 14 full-level runs within the first day (100% explored, enemies slain, and loot collected), 6 of those runs had a Black Kat appear, and one of those 6 runs had 2 of them appear at once.

Now, up to the point, I had spawned about 32 - 53 Spookats, prior to an appearance of a Black Kat being triggered. Now, the chances of getting 2 Black Kats at once likely fall within the lines of pushing sheer luck, but that's as far as I can guess. From that range, give or take a few, as it can not be precisely in that range for everyone, but it's what I had witnessed at the least.

However, interestingly enough, when the server maintenance occurred recently, I found that the Black Kats had an alarming increase in their spawn rate. The original amount of Spookats that needed to be met in order to trigger a spawning of them... it was virtually halved. Post-maintenance, I found every 11 - 27 Spookats that had spawned would next trigger a Black Kat spawning.

Now, in the time of the event, I've done this 246 times, and I have not gotten a Book of Dark Rituals drop from the Black Kats I've slain, but I've slain enough of them to acquire enough ancient pages for 3 Kat Tribe Fetishes; roughly equal out to 51 of them being killed.

If you want the dumbed down version of this, I'll simply sum it up in the fact that finding them outside of the end of the Kataclysmic Confrontation's prestige missions... is almost a play of pure luck! Yep!

Despite my findings, this does not always mean it's the same for everyone else who's hunting for them as well. Not everyone's going to find enough Black Kats to get what they want, and some who may have a streak of luck might find a nice few of them, but never get the rarest of drops once they have been killed off, no matter the numbers and numbers of them slain.

Plus, as others have noticed... this is like other events from the past; you can't get everything in the first implementation of new items in the event you grind endlessly, but much rather you collect what you have, and await for the return of said event. Give yourself an advantage when it returns!

Search

Let's science: Black Kat spawn ratio