Great Depth Worm's Spawning Mechanic Needs Completely Rework

[Bumber64]

On 9/10/2025 at 6:50 AM, yyyyyyyyyyyyyyy said:

I use AI to calculate the probability of spawning GDW in a Worm Attack. And meanwhile I independently calculated the probabilities for the first six attacks using mathematical knowledge of random variables

First off, never trust an AI (LLM) with numbers. They can't actually do math. They're making educated guesses on what looks believable.

You didn't explain your independent method, but your interpretation is incorrect based on your numbers. Saying it caps at 50% and then spontaneously stopping at 0.15866 (15%) is a bright red flag. Why does the probability decrease after wave 7?

The chances are just these:

1	0.00
2	0.05
3	0.10
4	0.15
5	0.20
6	0.25
7	0.30
8	0.35
9	0.40
10	0.45
11	0.50
12	0.50

Spawning of the GDW is blocked before day 25. That's not strictly the world day count, IIRC, but time spent in the caves. Regardless of the particulars, the chance still increments for the next wave.

Are your numbers actually the chance for a given wave number to contain the GDW, considering you may have already got one? That's a misleading statistic given the thread title.

Or are they pure AI hallucination?

Edited September 12, 2025 by Bumber64
The 25 day limit isn't based on player age. Just cave day.

[Mike23Ua]

Can’t you literally just craft a whole bunch of hostile flares, bundle them up into winter feast gift packages take them with you down into caves and then use them as needed?

If someone who works at Klei wants to step in and correct me here I would appreciate it but I was under the Assumption that the hostile flare could be used to force summon incoming Depth Worm Waves? (which would also include a chance at the GDW)

[Bumber64]

1 minute ago, Mike23Ua said:

I was under the Assumption that the hostile flare could be used to force summon incoming Depth Worm Waves?

You know what they say about assumptions, Mike.

[Dingle]

7 minutes ago, Mike23Ua said:

Can’t you literally just craft a whole bunch of hostile flares, bundle them up into winter feast gift packages take them with you down into caves and then use them as needed?

If someone who works at Klei wants to step in and correct me here I would appreciate it but I was under the Assumption that the hostile flare could be used to force summon incoming Depth Worm Waves? (which would also include a chance at the GDW)

Hostile flare only does 3 things off top of head: Deerclops and Mactusk summon chance in Winter, and pirate monkey raid on a boat.

I was always sad it never got anything else.

[Mike23Ua]

2 minutes ago, Dingle said:

Hostile flare only does 3 things off top of head: Deerclops and Mactusk summon chance in Winter, and pirate monkey raid on a boat.

I was always sad it never got anything else.

That’s what the Wiki says anyway but I throughly read the games actual patch notes (which can change overtime) and I’m pretty sure somewhere along the way the hostile flare increased the chances of force spawning hound & depth worm waves.

Now please note: this does not spawn them EARLY or as soon as the flare is set, it simply summons them to the player when you already hear or see the on screen warnings.

If they do not actually do this, then I’ve been wasting the holy high heck out of hostile flares trying to aggro Hound & Worm waves onto myself so they avoid spawning on my significantly less skilled friends who are trying to learn and enjoy the game.

[yyyyyyyyyyyyyyy]

OK, I find some people confuse total probability with conditional probability. Let me explain, we mark regular attack as 0, GDW attack as 1. Take the 4th attack as example, my results are the sum of probability of two occasions: 0001 and 0101. My results are total probability, which means they consider all the occasions. The 50% means that first 10 times attack are regular ones, under this "condition", the 11th attack has a  50% probability to be a GDW one, which is 00000000001.This is conditional probability.

40 minutes ago, Bumber64 said:

First off, never trust an AI (LLM) with numbers. They can't actually do math. They're making educated guesses on what looks believable.

You didn't explain your independent method, but your interpretation is incorrect based on your numbers. Saying it caps at 50% and then spontaneously stopping at 0.15866 (15%) is a bright red flag. Why does the probability decrease after wave 7?

The chances are just these:

1	0.00
2	0.05*
3	0.10*
4	0.15
5	0.20
6	0.25
7	0.30
8	0.35
9	0.40
10	0.45
11	0.50
12	0.50

*Actually 0 if before day 25. (Possibly cumulative time spent by the oldest player in the caves during wave setup? IIRC, it's not the world day count.)

Are your numbers actually the chance for a given wave number to contain the GDW, considering you may have already got one? That's a misleading statistic given the thread title.

Or are they pure AI hallucination?

 

9 hours ago, Semind said:

Anecdotally, it takes me hundreds of days to get my first GDW in a world. I can almost certainly count on one hand the number of times I've encountered it so far. Also I do not understand at all how you get 16% when it's directly stated to be 50%

 

[Bumber64]

3 minutes ago, yyyyyyyyyyyyyyy said:

OK, I find some people confuse total probability with conditional probability. Let me explain, we mark regular attack as 0, GDW attack as 1. Take the 4th attack as example, my results are the sum of probability of two occasions: 0001 and 0101. My results are total probability, which means they consider all the occasions. The 50% means that first 10 times attack are regular ones, under this "condition", the 11th attack has a  50% probability to be a GDW one, which is 00000000001.This is conditional probability.

Why do we care? Are you planning on missing waves? Waves don't progress without players, which just means someone else got the GDW.

What's more interesting is the probability of seeing a GDW by a given wave, which would actually be higher than the chance per wave. Therefore,

43 minutes ago, Bumber64 said:

That's a misleading statistic given the thread title.

[yyyyyyyyyyyyyyy]

Based on the description of depth worm attacks from the wiki, I independently mapped out all possible scenarios for the first six attacks using tree diagrams (calculating beyond this would become extremely complex due to the rapidly increasing number of scenarios). Using knowledge and methods of random variables, I computed the corresponding total probabilities. Meanwhile, I provided the same description to the AI and asked it to independently calculate the total probabilities for the first 20 attacks. The results showed that the data for the first six attacks matched between both calculations, which is why I trust the AI's results for the remaining 14 attacks.

[Bumber64]

But the data isn't meaningful to any point.

The only thing wrong with the spawn mechanic is that you can technically fail a 50% chance forever.

Edited September 11, 2025 by Bumber64

[yyyyyyyyyyyyyyy]

2 minutes ago, Bumber64 said:

The only thing wrong with the spawn mechanic is that you can technically fail a 50% chance forever.

I suggest you learn the Law of Large Numbers, or be quiet.

7 minutes ago, Bumber64 said:

Why do we care? Are you planning on missing waves? Waves don't progress without players, which just means someone else got the GDW.

What's more interesting is the probability of seeing a GDW by a given wave, which would actually be higher than the chance per wave. Therefore,

Do you really understand what do I mean? 0 means regular worm attack, not missing a worm attack.

[Bumber64]

17 minutes ago, yyyyyyyyyyyyyyy said:

I suggest you learn the Law of Large Numbers, or be quiet.

What does that have to do with anything? Are you suggesting a player can wait for the GDW for a number of waves approaching infinity?

If your thread title was "The Chance of GDW Spawning on a Given Wave" then I'd agree with you, but you're just throwing irrelevant data at people hoping to confuse them into agreement. You can't intimidate me into silence.

Edited September 11, 2025 by Bumber64

[yyyyyyyyyyyyyyy]

These numbers here, are trying to explain how rare GDW is! The truth is, we don't always stay in the cave. If we keep staying in the cave, it still takes a year for a GDW to spawn!

I'm not try to make GDW a territorial boss, it doesn't have to. All I want is that GDW spawns more steady.

[Bumber64]

19 minutes ago, yyyyyyyyyyyyyyy said:

Do you really understand what do I mean? 0 means regular worm attack, not missing a worm attack.

What is the relevance? Each regular worm wave increases the GDW spawn chance by 5%. Why do we care that we didn't get one on wave X specifically?

11 minutes ago, yyyyyyyyyyyyyyy said:

These numbers here, are trying to explain how rare GDW is! The truth is, we don't always stay in the cave. If we keep staying in the cave, it still takes a year for a GDW to spawn!

Somewhat true, but it's not a result of the mechanics. Worm waves are slow, so ten waves is a long time. A GDW every other wave wouldn't be good, however. Like deerclops showing up every 20 days year-round.

Edited September 11, 2025 by Bumber64

[yyyyyyyyyyyyyyy]

1 minute ago, Bumber64 said:

What is the relevance? Each regular worm wave increases the GDW spawn chance by 5%. Why do we care that we didn't get one on wave X?

Somewhat true, but it's not a result of the mechanics. Worm waves are slow, so ten waves is a long time.

May I ask do you have any knowledge related to probability theory? Expected value? Law of Large Numbers? Conditional probability? Random variables? If not, I won't waste my time explaining these to you as my time and English knowledge are limited. I should spend them on the discussion of how to improve spawning mechanic of GDW.

I should clarify that this topic is meant to explain why GDW is so rare. In fact I support all the methods to summon or spawn GDW as long as they can make GDW spawns regularly and controllable.

[viblym]

I'd like it if they just increased the odds of the boss spawning but the "territorial boss" idea sounds very boring, we already have several bosses like that, and every cave boss functions this way. I much prefer the worm actually targeting the player

[Bumber64]

I took statistics in college. My issue is that you're misrepresenting your statistics to show something they don't.

This reminds me of the anecdote of a WWI study that showed a new helmet was increasing the number of soldiers sent to the hospital with head injuries. So what was happening? Well, soldiers shot in the head wearing the old helmet weren't ending up in the hospital. They were being tracked as combat deaths.

So when you list attack number 10 as a "0.15889" chance, you're misrepresenting how many GDWs you've fought so far. The outcome 0101 represents two GDW encounters, but you don't account for that data. It's treated the same as 0001.

Edited September 11, 2025 by Bumber64

[yyyyyyyyyyyyyyy]

11 minutes ago, Bumber64 said:

So when you list attack number 10 as a "0.15889" chance, you're misrepresenting how many GDWs you've fought so far. The outcome 0101 represents two GDW encounters, but you don't account for that data. It's treated the same as 0001.

0101's probability is 0.05²＝0.25%. Like I say these numbers showed how rare GDW is from a expected value view, which means GDW spawns once a year on average.

You can give your probabilities by the way. Instead of keep saying I'm making up numbers.

[Chesed]

12 hours ago, Mike23Ua said:

That’s what the Wiki says anyway but I throughly read the games actual patch notes (which can change overtime) and I’m pretty sure somewhere along the way the hostile flare increased the chances of force spawning hound & depth worm waves.

Now please note: this does not spawn them EARLY or as soon as the flare is set, it simply summons them to the player when you already hear or see the on screen warnings.

If they do not actually do this, then I’ve been wasting the holy high heck out of hostile flares trying to aggro Hound & Worm waves onto myself so they avoid spawning on my significantly less skilled friends who are trying to learn and enjoy the game.

From what I can actually tell quickly looking at the item, no it doesn't.

Here is everything programmed to respond to the activation of a Hostile Flare.

They are pirate waves, Deerclops, and MacTusk. There are no mention of hound waves or depths worm waves interacting with the item at all.

Even if the activation of the flare had a delayed effect as you're claiming it does, the waves would still likely respond to this same function to keep track of when the flare was used and who used it, and it doesn't. So saying it doesn't activate immediately doesn't have much of an impact in that case.

Edited September 11, 2025 by Chesed
a word

[ThemInspectors]

 

5 hours ago, yyyyyyyyyyyyyyy said:

0101's probability is 0.05²＝0.25%. Like I say these numbers showed how rare GDW is from a expected value view, which means GDW spawns once a year on average.

You can give your probabilities by the way. Instead of keep saying I'm making up numbers.

 

6 hours ago, Bumber64 said:

Why do we care? Are you planning on missing waves? Waves don't progress without players, which just means someone else got the GDW.

What's more interesting is the probability of seeing a GDW by a given wave, which would actually be higher than the chance per wave. Therefore,

Goodness, OP's math is strange. These are the actual Chances for seeing a GDW. Individual means the chance of seeing GDW on wave N, and Any means the chance of seeing it before or on wave N

Wave	GDW (Individual)	GDW (Any)
1	0.0%	0.0%
2	5.0%	5.0%
3	10.0%	14.5%
4	15.0%	27.3%
5	20.0%	41.9%
6	25.0%	56.4%
7	30.0%	69.5%
8	35.0%	80.2%
9	40.0%	88.1%
10	45.0%	93.5%
11	50.0%	96.7%

Median time to happen is at around wave 6, (As in, you can expect that half of your encounters will happen before wave 6 and half after wave 6)

 

Math for those interested:

Spoiler

P(X) = Individual GDW
P(!X) = Not seeing GDW = 1 - P(X)
P(!Y) = Not seeing ANY GDW = P(!X_1) * P(!X_2) ... * P(!X_n)
P(Y) = Seeing ANY GDW = 1 - P(!Y)

 

Wave	P(X)	P(!X)	P(!Y)	P(Y)
1	0.00%	100.00%	100.00%	0.00%
2	5.00%	95.00%	95.00%	5.00%
3	10.00%	90.00%	85.50%	14.50%
4	15.00%	85.00%	72.68%	27.33%
5	20.00%	80.00%	58.14%	41.86%
6	25.00%	75.00%	43.61%	56.40%
7	30.00%	70.00%	30.52%	69.48%
8	35.00%	65.00%	19.84%	80.16%
9	40.00%	60.00%	11.90%	88.10%
10	45.00%	55.00%	6.55%	93.45%
11	50.00%	50.00%	3.27%	96.73%

 

OP's math is strange, I can't really follow it at all. It tries to calculate the individual success of GDW spawning based a chain result, but it doesn't make sense to mix the probabilities that way.

 

Edited September 11, 2025 by ThemInspectors

[Bumber64]

6 hours ago, yyyyyyyyyyyyyyy said:

You can give your probabilities by the way. Instead of keep saying I'm making up numbers.

I wrote a Lua program:

Spoiler

local chance = {0.0, 0.05, 0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.40, 0.45} --else 0.5
local results = {}

local function test(s) --Analyze probability for a string like '010'
  local idx, gdw, t = 1, 0, {} --Chance index, GDW count, equation
  for x = 1, #s do --Iterate chars
    if s:byte(x) == ('1'):byte() then --GDW wave
      table.insert(t, chance[idx] or 0.5) --Store current chance in equation
      idx = 1; gdw = gdw+1 --Reset chance, count the GDW
    else --Normal wave
      table.insert(t, 1.0-(chance[idx] or 0.5)) --Store chance of not GDW
      idx = idx+1 --Increase next chance
    end
  end
  local prob = 1.0
  for _,v in ipairs(t) do --Multiply probabilities
    prob = prob * v
  end
  local result = {s = s, p = prob, t = t, gdw = gdw} --We save these
  result.text = ('%s: %.4f (%d GDW)'):format(s, prob, gdw) --Display text
  result.math = table.concat(t, ' * ') --Show work
  return result
end

local function recurse(s, waves) --Branch out possibilities
  local w = #s --Length is current wave
  if w > waves then return end --Exceeded desired limit
  results[w] = results[w] or {} --Group results by wave
  table.insert(results[w], test(s)) --Test combination and store results
  recurse(s..'0', waves) --Next is not GDW
  if s:sub(-1) == '0' then --'11' is invalid
    recurse(s..'1', waves) --Next is GDW
  end
end

local function analyze(waves, show_work) --Analyze all combinations and print
  recurse('0', waves) --First wave never GDW, do recursive analysis
  for k,v in ipairs(results) do --Get results by wave
    print('-- Wave #'..k..' --')
    local tot = {} --Track probability per GDW count
    for _,r in ipairs(v) do --Iterate each combination in wave
      print(r.text) --Display text
      if show_work then --Show the equation
        print(' = '..r.math)
      end
      tot[r.gdw+1] = (tot[r.gdw+1] or 0.0) + r.p --Add to total; Lua arrays start at 1!
    end
    local out = nil --GDW count text
    for x,y in ipairs(tot) do --First entry is a special case
      out = (out and ('%s, %d: %.4f'):format(out, x-1, y) or ('GDW Count: [%d: %.4f'):format(x-1, y))
    end
    print(out..(']\nHas GDW: %.4f\n'):format(1.0-tot[1])) --Chance of any GDW; tot[1] is 0 GDW chance
  end
end

analyze(10, false) --10 waves, no equations

You can run it from an online interpreter like this one. On the last line you can change 10 to however many waves you want to calculate. (It prints all valid combinations, so don't go too crazy.) Changing false to true makes it show its work after each combination like:

010100: 0.0024 (2 GDW)
 = 1.0 * 0.05 * 1.0 * 0.05 * 1.0 * 0.95
010101: 0.0001 (3 GDW)
 = 1.0 * 0.05 * 1.0 * 0.05 * 1.0 * 0.05

(Disabling printing of each combination should be as easy as putting -- before print(r.text). I might add an option later.)

You wanted the probabilities? Here they are:

-- Wave #1 --
GDW Count: [0: 1.0000]
Has GDW: 0.0000
-- Wave #2 --
GDW Count: [0: 0.9500, 1: 0.0500]
Has GDW: 0.0500
-- Wave #3 --
GDW Count: [0: 0.8550, 1: 0.1450]
Has GDW: 0.1450
-- Wave #4 --
GDW Count: [0: 0.7268, 1: 0.2707, 2: 0.0025]
Has GDW: 0.2732
-- Wave #5 --
GDW Count: [0: 0.5814, 1: 0.4066, 2: 0.0120]
Has GDW: 0.4186
-- Wave #6 --
GDW Count: [0: 0.4361, 1: 0.5301, 2: 0.0337, 3: 0.0001]
Has GDW: 0.5639
-- Wave #7 --
GDW Count: [0: 0.3052, 1: 0.6220, 2: 0.0719, 3: 0.0008]
Has GDW: 0.6948
-- Wave #8 --
GDW Count: [0: 0.1984, 1: 0.6702, 2: 0.1282, 3: 0.0031, 4: 0.0000]
Has GDW: 0.8016
-- Wave #9 --
GDW Count: [0: 0.1190, 1: 0.6716, 2: 0.2006, 3: 0.0087, 4: 0.0001]
Has GDW: 0.8810
-- Wave #10 --
GDW Count: [0: 0.0655, 1: 0.6312, 2: 0.2834, 3: 0.0197, 4: 0.0003, 5: 0.0000]
Has GDW: 0.9345

We'll take the breakdown of wave 4:

-- Wave #4 --
0000: 0.7268 (0 GDW)
0001: 0.1283 (1 GDW)
0010: 0.0950 (1 GDW)
0100: 0.0475 (1 GDW)
0101: 0.0025 (2 GDW)
GDW Count: [0: 0.7268, 1: 0.2707, 2: 0.0025]
Has GDW: 0.2732

You said it's 0.13075. That seems to line up with 0001 + 0101. Did you forget 0010 and 0100 as combinations?

6 hours ago, yyyyyyyyyyyyyyy said:

GDW spawns once a year on average.

It's actually worse than that. A default DST year is 70 days.

First year you can get at best 8 waves. 8 waves is (0.6702+2*0.1282+3*0.0031+4*0.0000 = ) 0.94 GDW.

After that, the best case drops down to 7 waves per year. That's (0.6220+2*0.0719+3*0.0008 = ) 0.7682 GDW.

If we ignore years, we can just say it's actually around 100 days per GDW. Average 9 waves for (0.6716+2*0.2006+3*0.0087+4*0.0001 = ) 1.1 GDW.

1 hour ago, ThemInspectors said:

Goodness, OP's math is strange. These are the actual Chances for seeing a GDW.

You sniped me while I was editing my post. Our numbers seem to line up.

Edited September 11, 2025 by Bumber64

[yyyyyyyyyyyyyyy]

34 minutes ago, ThemInspectors said:

 

 

Goodness, OP's math is strange. These are the actual Chances for seeing a GDW. Individual means the chance of seeing GDW on wave N, and Any means the chance of seeing it before or on wave N

Wave	GDW (Individual)	GDW (Any)
1	0.0%	0.0%
2	5.0%	5.0%
3	10.0%	14.5%
4	15.0%	27.3%
5	20.0%	41.9%
6	25.0%	56.4%
7	30.0%	69.5%
8	35.0%	80.2%
9	40.0%	88.1%
10	45.0%	93.5%
11	50.0%	96.7%

Median time to happen is at around wave 6, (As in, you can expect that half of your encounters will happen before wave 6 and half after wave 6)

 

Math for those interested:

  Reveal hidden contents

P(X) = Individual GDW
P(!X) = Not seeing GDW = 1 - P(X)
P(!Y) = Not seeing ANY GDW = P(!X_1) * P(!X_2) ... * P(!X_n)
P(Y) = Seeing ANY GDW = 1 - P(!Y)

 

Wave	P(X)	P(!X)	P(!Y)	P(Y)
1	0.00%	100.00%	100.00%	0.00%
2	5.00%	95.00%	95.00%	5.00%
3	10.00%	90.00%	85.50%	14.50%
4	15.00%	85.00%	72.68%	27.33%
5	20.00%	80.00%	58.14%	41.86%
6	25.00%	75.00%	43.61%	56.40%
7	30.00%	70.00%	30.52%	69.48%
8	35.00%	65.00%	19.84%	80.16%
9	40.00%	60.00%	11.90%	88.10%
10	45.00%	55.00%	6.55%	93.45%
11	50.00%	50.00%	3.27%	96.73%

 

OP's math is strange, I can't really follow it at all. It tries to calculate the individual success of GDW spawning based a chain result, but it doesn't make sense to mix the probabilities that way.

 

I can understand what do your numbers mean. But I don't get what do you want to explain using them. What do we discuss here is not just encountering the first GDW, right?

Let me explain my numbers in another way. Consider them as the number of GDW you encounter in each wave, instead of probability. Then you can understand why I say GDW spawns once a year. And I do omit the limitation that GDW spawns after day25 and the probability of first 3 attacks as they are too small. The omitting aims to simplify calculation. So in fact it takes even longer for a GDW to spawn.

8 minutes ago, Bumber64 said:

If we ignore years, we can just say it's actually around 100 days per GDW. Average 9 waves for (0.6716+2*0.2006+3*0.0087+4*0.0001 = ) 1.1 GDW.

Wait a minute, are you trying to prove my numbers are smaller than real ones? ALL THE TIME?

Have you noticed that the sum of my probabilities of first 9 attacks is also 1.1?

"A YEAR" is not "THE FIRST YEAR". OK, now I know where the problem is, let me edit this topic.

[yyyyyyyyyyyyyyy]

Wait, this is not my fault. I said "To simplify, let's consider the occasion after Day 101. " at the very beginning. Now it is Bold. 

[Bumber64]

2 hours ago, yyyyyyyyyyyyyyy said:

Let me explain my numbers in another way. Consider them as the number of GDW you encounter in each wave, instead of probability.

I think I get it. You're saying it's each wave's contribution to getting a GDW. Or what fraction of a GDW you get per wave. This wasn't indicated well, and never showing how you "independently" calculated any of the numbers doesn't help. My initial impression was that you were claiming the GDW spawn rate itself mysteriously drops. (It doesn't, but the waves do get farther apart.)

At first I thought you were trying to show the odds of a single wave triggering your first GDW spawn, and misinterpreting the wiki to mean a random 5%-50% increase in the current rate each wave. (These numbers obviously didn't match your data, so that wasn't it.) Many people claim they've never even seen GDW.

Later you clarified that you were calculating each set of possibilities, which led me to believe you were trying to calculate the odds of seeing exactly 1 GDW by a given wave. (This wouldn't be useful info, and the numbers still didn't match.)

So the actual thing you're trying to show is how many GDWs you can get over a number of waves, which is useful data. You list the individual value per wave, but it wasn't clear that these weren't cumulative values, and still needed to be added to the previous waves.

My numbers show the odds of getting each number of GDWs after a given number of waves. Taking a weighted average of these gets you the average number of GDWs after a given number of waves. (My program doesn't do this step.) To get the contribution per wave, you'd have to subtract off the previous waves' contributions.

2 hours ago, yyyyyyyyyyyyyyy said:

Have you noticed that the sum of my probabilities of first 9 attacks is also 1.1?

This was the fundamental misunderstanding for me. It wasn't clear that the previous values needed to be added together to get the total GDWs.

That said, I'm still puzzled why the value suddenly decreases after wave 7. Unless we're factoring in the time between waves, the waves themselves surely shouldn't be getting worse at spawning a GDW? I'd have to take the weighted averages and see how they're changing between waves. Maybe later.

2 hours ago, yyyyyyyyyyyyyyy said:

"A YEAR" is not "THE FIRST YEAR". OK, now I know where the problem is, let me edit this topic.

I wasn't necessarily assuming this. The data is provided in waves, not days, so it didn't seem to matter. Converting from waves to days is an extra step, and we can just multiply by 11 in the long run.

I think we've got the confusion cleared up, but IDK if your numbers are correct yet. I spent almost 3 hours on the program itself, so I'm tired of numbers.

Edited September 11, 2025 by Bumber64

[ThemInspectors]

1 hour ago, yyyyyyyyyyyyyyy said:

I can understand what do your numbers mean. But I don't get what do you want to explain using them. What do we discuss here is not just encountering the first GDW, right?

Let me explain my numbers in another way. Consider them as the number of GDW you encounter in each wave, instead of probability. Then you can understand why I say GDW spawns once a year. And I do omit the limitation that GDW spawns after day25 and the probability of first 3 attacks as they are too small. The omitting aims to simplify calculation. So in fact it takes even longer for a GDW to spawn.

1. Fractional number of GDW is an odd way of quantifying it. 

2. Since the median amount of waves is 6 and you are assuming 11 days for Depth worm attack, we should expect a GDW every 66 days. A year is 70 days, so it's less than a year which is rather reasonable for me. The main issue is that the probabilities you use are just weird and aren't well explained?

3. The 25 Day GDW Limitation isn't a big deal. As the first probabilities of the earlier waves aren't high, it's not super likely that the GDW will spawn then. The Median of 6 waves still holds for waves 1,2,3 not spawning a GDW

Spoiler

Special case for day starting at 0 (25 day buffer, approx 3 waves)
Wave	P(X)	P(!X)	P(!Y)	P(Y)
1	0.00%	100.00%	100.00%	0.00%
2	0.00%	100.00%	100.00%	0.00%
3	0.00%	100.00%	100.00%	0.00%
4	15.00%	85.00%	85.00%	15.00%
5	20.00%	80.00%	68.00%	32.00%
6	25.00%	75.00%	51.00%	49.00%
7	30.00%	70.00%	35.70%	64.30%
8	35.00%	65.00%	23.21%	76.80%
9	40.00%	60.00%	13.92%	86.08%
10	45.00%	55.00%	7.66%	92.34%
11	50.00%	50.00%	3.83%	96.17%

4. Due to the probability distribution, it's really unlikely that it takes more than 110 days to spawn. Additionally, since half of GDW waves happen on or before wave 6, it wouldn't be uncommon to see 3 GDW every 2 years or 4 GDW every 3 years.

[Bumber64]

I want to say that I don't think the mechanic fundamentally needs reworking, just tweaking.

IIRC, setting GDW to Often changes it to +10% per wave and caps at 100%. Should we just make that Default? We can run the numbers on it.

Edited September 11, 2025 by Bumber64