Brockton's Celestial Forge (Worm/Jumpchain)

Can you calculate what are the probability for Joe to draw one specific 400 pts perk next chapter (assuming 900 points gained), pretty please ?

(Very quick and dirty napkin math give me 1%, but I'm not good enough at probability to calculate the exact number myself.)

The Reference spreadsheet has a "Chance to Roll" column for every perk, so if are looking for say the 400 point Armsmaster perk, that is 2.78% I have no idea how they reach that percentage, have not taken the stats class I should have.
 
The Reference spreadsheet has a "Chance to Roll" column for every perk, so if are looking for say the 400 point Armsmaster perk, that is 2.78% I have no idea how they reach that percentage, have not taken the stats class I should have.
Hmmm. According to that document the most likely perks to be rolled are the Avid Glove and a Splatoon perk that gives you breedable snails you can kill to improve the abilities of clothing items. Both are only 4.17%, so even being the most likely apparently isn't that big of an advantage, but hopefully it gives them the boost they need. I would love to see Joe suddenly have a bunch of weird pets (I doubt he'd be that willing to kill cute cartoon snails just to give his equipment a minor boost. If he wants to harvest them after they die naturally he'll have to wait fifteen years as well...).
 
Last edited:
Ok I'm working on a detailed post calculating the possibilities of rolling M/E, tell me if you meant another 400CP perk, it does affect the math
 
Done! Anyway, to fill the word tax, what are the thread's thoughts on Joe's Gundam vs his Veritech?
I'd imagine they either eventually all end up being similar to one another, with their individual quirks and advantages being integrated into one another, or else they'll all end up built to maximize their own unique benefits. Then they would either work as a team to take down huge threats, or just take on threats solo that their individual specializations prepare them for.

As for how they compare to each other in their current states... well, I'm not in the correct fandoms to be able to knowledgeably comment on that.
 
Last edited:
Done! Anyway, to fill the word tax, what are the thread's thoughts on Joe's Gundam vs his Veritech?
Duel of two ridiculously overpowered mech with Apeiron's trademarked showines?
Yes, please, just yes.

Which one would win? Well i believe most satisfying answer would be...
He needs to do the fusion dance with them.
...Verigundam? I mean friendship. Friendship in shape of very scary mech.
 
Last edited:
Would the perk that allows two items to be fused trigger Workaholic and create a total of 5?

Edit:

Because if it did, that's the way you get a giant robot army.
 
Done! Anyway, to fill the word tax, what are the thread's thoughts on Joe's Gundam vs his Veritech?
Well to start with. It's a bit of an Apples and Oranges. I mean both are still giant robots/fruit. But Mobile Suits and Veritechs are built for very different kinds of conflict.

Leading off with Mobile Suits, the big thing to remember with them, is that they're designed to behave as, what are effectively macro sized infantry. As Joe himself pointed out, Gundam physics are designed to facilitate their function as infantry. With them melee to short range combat is the primary paragram, with long distance combat being more something for battleships and static defenses, with few exceptions. They're strong, beefy, and robustly designed to go from one fight to the next with not too much maintenance needed, not requiring too complex of logistics or work crews to keep going.

Veritechs on the flipside, are Aerospace Fighters that have the capacity to transform into mecha, and this is important to note, because the fighter aspect is the primary one. VF's are high speed/agility weapons deployment platforms, meant to help deliver ordinance to the enemy in as swift, reliable, and safe a manner as possible. This makes them comparative hanger queens, yet also paradoxedly more disposable by nature, since like fighter craft, they're meant to operate in a more mission by mission basis, normally attached to dedicated supply bases (if often mobile ones).

Overall, the Veritech is probably more mobile, fast, and more 'plug and play' for ordinance, but is less adaptable to entire new weapons loadouts, where the Mobile Suit's more infantry analogous design shines brighter. The Mobile suits are also better suited to urban combat, where the Veritech is more high speed response. And from what I recall even grunt MS tend to have a lot more 'muscle' which can make them very useful for a number of situations.
 
Last edited:
Can you calculate what are the probability for Joe to draw one specific 400 pts perk next chapter (assuming 900 points gained), pretty please ?

(Very quick and dirty napkin math give me 1%, but I'm not good enough at probability to calculate the exact number myself.)
If the perk was drawn from one out of 90 possibilities the chances would 1.85%. Math stuff: a drawing at every 200 points gives a with 15/90 perks at or below 200 points means 84% of first draws will be a skip. After 1 or more skips there will be a 1/90 chance to get the perk. That means for every 2 picks there is a 1.85% chance [(75/90)x(1/90)x2] of a specific 400 point perk.

The problem is the way Lord does his calculations is by constellation rather than perk. So a perk in a constellation with fewer perks has a higher chance of a specific perk being picked. As each constellation has an 8.34% chance [1/11] of coming up, and each perk has 1/x percent chance where x is the remaining perks in the constellation the chance of a specific 400 point perk is hard to pin down.

An example for Miniaturization/Efficiency (Worm) 400pt.
The size constellation has 3 perks remaining. The chance of drawing the size constellation is 8.34 percent. There is 1/3 chance of getting Min/Eff drawn from the size constellation meaning there is a 2.34% chance of that perk being picked on each draw.

TL : DR Math sucks, the answer is complicated
 
Last edited:
I suck at stats/probabilities, although I am good at most Math (besides geometry, fuck geometry)
for a 400 we have a couple possibilites.
For 100CP to be missed, there is roughly a ~91.32%
For 200CP to be missed, there is roughly a ~94.03%
For 300CP to be missed, there is roughly a ~84.17%
For rolling the actual perk, you take into account the chance of rolling the Size constellation, 1/12, and the chances of rolling M/E in particular, 1/3. In total that is 1/36, a massive shot to our odds since the chances of specifically rolling M/E are 2.77%.
However, multiply this with the odds of previous steps and you get:
Starting with 100: 2%
Starting with 200: 2.19%
Starting with 300: 2.33%
Starting with 400: 2.77%
HEAVY NOTE: Of course, the odds will play out differently since Joe could have up to 5 tries to get different perks before this(900CP per chapter and all), which will influence odds
HEAVY NOTE 2: Odds may look depressing, but considering his high chances of saving, it could happen any time. Something with 0.00001% can happen, they are just rare.
 
Last edited:
Says:
I suck at stats/probabilities
Then proceeds to:
For 100CP to be missed, there is roughly a ~91.32%
For 200CP to be missed, there is roughly a ~94.03%
For 300CP to be missed, there is roughly a ~84.17%
And then:
However, multiply this with the odds of previous steps and you get:
Starting with 100: 2%
Starting with 200: 2.19%
Starting with 300: 2.33%
Starting with 400: 2.77%
I doubt your first statement. Also thank you.
 
Last edited:
I see getting a Diploma for Anaheim.

Does that mean the credits he earned is recognized at local Universities/

Could he use it to 'Graduate' and get his Family to stop bothering him?
 
I see getting a Diploma for Anaheim.

Does that mean the credits he earned is recognized at local Universities/

Could he use it to 'Graduate' and get his Family to stop bothering him?
Wait... I think you're on to something here. Joe can literally say he has a degree in giant robot science and be totally honest. I want him to say that to the Undersiders when Lisa is there, her reaction to something her power is saying is 100% the truth would be hilarious.
 
If the perk was drawn from one out of 90 possibilities the chances would 1.85%. Math stuff: a drawing at every 200 points gives a with 15/90 perks at or below 200 points means 84% of first draws will be a skip. After 1 or more skips there will be a 1/90 chance to get the perk. That means for every 2 picks there is a 1.85% chance [(75/90)x(1/90)x2] of a specific 400 point perk.

The problem is the way Lord does his calculations is by constellation rather than perk. So a perk in a constellation with fewer perks has a higher chance of a specific perk being picked. As each constellation has an 8.34% chance [1/11] of coming up, and each perk has 1/x percent chance where x is the remaining perks in the constellation the chance of a specific 400 point perk is hard to pin down.

An example for Miniaturization/Efficiency (Worm) 400pt.
The size constellation has 3 perks remaining. The chance of drawing the size constellation is 8.34 percent. There is 1/3 chance of getting Min/Eff drawn from the size constellation meaning there is a 2.34% chance of that perk being picked on each draw.

TL : DR Math sucks, the answer is complicated
I suck at stats/probabilities, although I am good at most Math (besides geometry, fuck geometry)
for a 400 we have a couple possibilites.
For 100CP to be missed, there is roughly a ~91.32%
For 200CP to be missed, there is roughly a ~94.03%
For 300CP to be missed, there is roughly a ~84.17%
For rolling the actual perk, you take into account the chance of rolling the Size constellation, 1/12, and the chances of rolling M/E in particular, 1/3. In total that is 1/36, a massive shot to our odds since the chances of specifically rolling M/E are 2.77%.
However, multiply this with the odds of previous steps and you get:
Starting with 100: 2%
Starting with 200: 2.19%
Starting with 300: 2.33%
Starting with 400: 2.77%
HEAVY NOTE: Of course, the odds will play out differently since Joe could have up to 5 tries to get different perks before this(900CP per chapter and all), which will influence odds
HEAVY NOTE 2: Odds may look depressing, but considering his high chances of saving, it could happen any time. Something with 0.00001% can happen, they are just rare.

Sanity check: I bute-forced the problem (well, my computer did, I just wrote the python code), and I got ~6.6% chance of getting the M/I perk next chapter and ~4.9% for Strong Spark (another story breaking perk). Disclaimer: numbers change a LOT if Joe have unspent points, with 300 unspent points, chances are around 10%.

Of course simulating the next chapter 100 000 times and taking the average only give approximate numbers.

(Assumptions in spoiler for space reason)

  • 0 unspent points at the end of this chapter. That one is probably wrong, so I'll update this post in a few days, once Lord give us the real number of points left.
  • 9 rolls per chapter. This is the average over the past few chapters.
  • 100000 simulations. That might be a bit on the low side, but I'm too lazy to wait 2min for my computer to compute.

Known problems (that probably induce very small errors in the calculations) :

  • No replacement of the Clothing constellation if it's emptied. In practice, Lord already said the first emptied constellations will be replaced with a capstone constellations, but it's very unlikely Joe would pick all Clothing THEN a capstone in a single chapter (it would cost him 300+100+300 of his 9*100 CP), so I'm not too worried.
  • Some perks exist in multiple constellations. I consider them independent. That's probably not correct, I might change it if I have time.
  • Some perks are part of a cluster, I treat them as separate rolls. I thinks the correct way to do it would be to lock the costlier perk of a cluster behind the cheaper perk (and doing a single roll for the whole cluster), but emulating that would require a lot more work. It might actually be pretty significant because of the 9 Star trek perks which are all 100 CP perks, so they are a lot more likely to get picked if they are separate.
 
Last edited:
May I ask about your algorithm? How did you calculate it is what I mean.

In pseudo code:

Code:
Do 100 000 times:
    Reset the CP and perk list
    Do 9 times:
        Increase CP
        Roll a constellation then a perk
        Buy the perk if you can
    Save if M/I and/or Strong spark have been picked

Write how many time M/I have been picked, divided by 100 000 (multiplied by 100 'cause it's a percentage). Same for Strong Spark.

In python: here (it's 200ish lines, althrough most of it is made of the perks, the actual logic is pretty short).
 
In pseudo code:

Code:
Do 100 000 times:
    Reset the CP and perk list
    Do 9 times:
        Increase CP
        Roll a constellation then a perk
        Buy the perk if you can
    Save if M/I and/or Strong spark have been picked

Write how many time M/I have been picked, divided by 100 000 (multiplied by 100 'cause it's a percentage). Same for Strong Spark.

In python: here (it's 200ish lines, althrough most of it is made of the perks, the actual logic is pretty short).
...Computer randomness is notoriously unreliable and nonrandom, and the large sample-size does not correct that
 
Praise be RNGesus and Monte-Carlo for letting us brute force complex probability calculations!
...Computer randomness is notoriously unreliable and nonrandom, and the large sample-size does not correct that
Python's RNG (mersenne twister) is good enough for statistical stuff(especially at this level). It is only when you need something that is cryptographically secure (eg secure and unpredictable against adversaries) that python's default rng fails.
EDIT: typo
 
Last edited:
Back
Top