LABYRINTH ENCHANTS ARE RIGGED

It ain't rigged folks.

See my thread.

https://www.pathofexile.com/forum/view-thread/1613293

"

DalaiLama wrote:

"

ScrotieMcB wrote:

There aren't 362 skills in the game. Just count them.

Not sure how valid the listing is, but not counting the level 75 ones, there are 348ish (349?) listed here:

http://pathofexile.gamepedia.com/List_of_helmet_enchantment_mods

Makes my eyes water trying to count those silly things.

I used PoeDB's datamined list which has a total (but which may be high on account of items that are turned off for whatever reason).

"

ScrotieMcB wrote:

There are 345 possible enchantments, with many skills getting 3 enchantments each. So 3/345=1/115 chance of getting one of these per attempt.

There are 82 skills with 3 possible enchantments:

Spoiler

Ancestral Protector, Animate Weapon, Arc, Arctic Breath, Ball Lightning, Barrage, Blade Vortex, Bladefall, Blast Rain, Blink Arrow, Burning Arrow, Caustic Arrow, Chaos Golem, Cleave, Cold Snap, Contagion, Detonate Dead, Discharge, Dominating Blow, Double Strike, Dual Strike, Elemental Hit, Fire Nova Mine, Fireball, Fire Trap, Firestorm, Flame Golem, Flame Surge, Flame Totem, Flameblast, Flicker Strike, Freezing Pulse, Frenzy, Frost Bomb, Glacial Cascade, Glacial Hammer, Ground Slam, Heavy Strike, Ice Crash, Ice Golem, Ice Shot, Ice Spear, Ice Trap, Incinerate, Infernal Blow, Kinetic Blast, Leap Slam, Lightning Arrow, Lightning Strike, Lightning Tendrils, Lightning Trap, Lightning Warp, Magma Orb, Mirror Arrow, Molten Strike, Orb of Storms, Power Siphon, Puncture, Rain of Arrows, Raise Spectre, Raise Zombie, Righteous Fire, Searing Bond, Shield Charge, Shock Nova, Shrapnel Shot, Siege Ballista, Spark, Spectral Throw, Split Arrow, Static Strike, Stone Golem, Storm Call, Summon Raging Spirit, Summon Skeletons, Sunder, Sweep, Tornado Shot, Viper Strike, Wild Strike.

This means there are 82 different ways to get a 9-peat.

We can estimate the chance of 9 repeats of a single skill occurring over 110 attempts as 82*110!/(9!101!)*(1/115)^9*(114/115)^101

This estimates the odds of the OP situation at about .0045%. (The chance to 9-peat off skills with only two available enchantments is negligible.) That is very, very odd luck, but not impossible. It should happen once in about 22322 110-enchant trials. So not a "one in a billion" or even "one in a million," but one in about twenty thousand. Given the amount of enchantments being ran daily, it isn't impossible, but inevitable that something like this would happen eventually.

Spoiler

Not literally inevitable, I understand randomness too, kthx.

OP is in the "under 0.1%" luck club, though, and making a RNGesus weirdness thread seems fair to me, provided he retracts his ridiculous statement that things are rigged when we have GGG staff saying it's not.

That most skills had more than one pair of effects is something I missed. I redid my poor mans Monte Carlo simulation with RANDBETWEEN(1,115) generating a longest repeat table of:



Which brings my results into the same ballpark as your estimate. I then did 9 more runs (didn't save results) and had a total of eight runs where the highest repeat was an 8x, one with a 9x, and one with both a 9x and a 10x.

All of this combined suggests that while the OPs odds of getting a 9-peat are very low; they're no longer so freakishly low to make me suspect something's wrong. A few thousand to one per user becomes plausible to happen to someone if you look at the whole set of forum/redit posters.

I stuck with a simple 1 to 115 roll instead of trying to handle the more complex case with variable numbers of results/skill you did because setting something like that up is pushing my Excel-fu to its limits and I don't have access to any real stats tools (nor the ambition to try coding from scratch).


What really surprised me is that while it took Excel less than a second to generate all 5000x110 random values every time I triggered a reroll (by editing any cell), the search/replace took over a minute to complete.

"

Summoner wrote:

What really surprised me is that while it took Excel less than a second to generate all 5000x110 random values every time I triggered a reroll (by editing any cell), the search/replace took over a minute to complete.

Shouldn't be surprising at all. Even a single edit triggers recalculation (unless you turn auto-calculate off, which you can). A search/replace is a collection of individual edits, therefore triggers multiple recalculations (again, unless set to manually calculate).

"

Perq wrote:

...

Nah, it was some random ACT 4 SUCKS CUZ I RAN 2 ACT 4 MAPS, AND THEN I RAN 2 ACT 3 MAPS, AND I GOT DOUBLE THE DROPS, thus ACT 4 HAS 50% OF DROPS kind of thread. Fun stuff.
But yeah, box threads are also fun! :D

Look, as long as Dread Thicket is still 50%, I'm good.

Also, box threads incoming this weekend, take cover/cowboy up.

"

ScrotieMcB wrote:

...

This estimates the odds of the OP situation at about .0045%. (The chance to 9-peat off skills with only two available enchantments is negligible.) That is very, very odd luck, but not impossible. It should happen once in about 22322 110-enchant trials. So not a "one in a billion" or even "one in a million," but one in about twenty thousand. Given the amount of enchantments being ran daily, it isn't impossible, but inevitable that something like this would happen eventually.

Spoiler

Not literally inevitable, I understand randomness too, kthx.

OP is in the "under 0.1%" luck club, though, and making a RNGesus weirdness thread seems fair to me, provided he retracts his ridiculous statement that things are rigged when we have GGG staff saying it's not.

Aye, this is why I gave OP a pass after seeing the 9-peat. Even without any clarifications I knew it was in the 'under 0.1% club' so his thought process wasn't entirely flawed, he was just overestimating the unlikelihood of the occurrence.

The box people QQ about 5% stuff, for a point of comparison.

"

Wruntjunior wrote:

You know, in programming, in order to make people feel like results are random for some cases we'll actually use structured randoms - there is a list of possible results, the random result is then temporarily removed from the list of possible results. Another concept occasionally used is excluding the current item from the possible list, if the previous structured random isn't used (I believe, though I easily may be mistaken, that this is done for chroms).

Ipod shuffles were blowing people's minds when a song was repeated back-to-back or even in close succession.

"

...

That said, as this IS programming...if multiple people are correlating that there are weird evidences of enchanting being weighted in some way, it probably would need dev attention, even if only to verify that it is actually working as intended. There could even be unexpected affects causing weighted results if there really is a problem, such as a problem with the RNG itself or even cosmic rays flipping a bit (outlandish sounding, sure...but it literally can happen); however, the plural of anecdote is not data - that is to say, one person reporting that they had bad results, or even several people saying they had bad results, is not evidence in and of itself.

TL;DR: RNG can be hard (to understand and do), so collect data if you believe there to be an issue

Right... To put it another way, one would compare the unlikelihood of an event to the likelihood of a different explanation, such as a glitch. The problem is we can't strictly quantify the latter so it comes down to examining red flags.

That's the irony in this post,

"

Real_Wolf wrote:

...



Omg, he got 9 so many times in a row, its rigged!

That if the 'so many' passed the subjective threshold, this is exactly the correct reaction (well, not 'rigged' but 'possible errors occurring')

"

Summoner wrote:

That most skills had more than one pair of effects is something I missed.

...

Don't knock yourself, OP actually didn't specify initially.

guys, really?

"

Qarl wrote:

Helmet enchants are evenly weighted. The repeats you report are certainly not mathematically impossible.

While getting that precise set of results again is not going to happen, getting a set of results with large number of repeats of things you don't want certainly can happen.

Could Qarl clarify something as it is unclear from the wording and the original OP?


This could mean every single individual enchant has the same weighting.

It could mean every single skill has the same weighting.

For example, you could have the same chance to get an Arc enchant as you have to get a Glacial Cascade enchant or a Kinetic Blast enchant. But, is the weighting the same for every enchant v every other enchant? For example, is it the same chance to get 25% Arc damage as it is to get +2 chain Arc?

This is quite important.