The math behind Magic Jewel Corruption - Tribute to the Great Vaal and The Adorned Diamond. 1164 Magic Jewels Corruption Results
Hi everyone,
We all love when our items become better via Corrupting don't we? But Corrupting expensive items is Dangerous. And on of the most expensive item in every build are the Jewels. A specially if you use the Adorned Diamond.
So I've been trying to get some for myself yet in the end ended up just burning almost all of my divs. This ofc resulted in my getting quite salty about it. I've tried all over internet to find the odds of Jewel corruption options yet both poe2db.tw and craftofexile.com do not have that information which resulted in me spending last couple evening buying Magic Jewels of the market for the remaining currency I've had in order to run my own tests.
So let's start a bit with explanation:
Introduction and Experiment Explanation
Why would you like to corrupt Magic Jewels?
Because of the existence of this item:
What it does is the same as Ingenuity except it does it for every corrupted Jewel you have allowing you to obtain combinations and amounts of certain stats present of Jewels otherwise Unavailable.
For example with 75% Adorned this item becomes 7% Attack Speed 26% Bonuses from Quiver, and 10 Fire Resistance (YES IF YOU ARE LUCKY ENOUGH TO HAVE ENCHANT ON YOUR GEM IT ALSO GET'S MULTIPLIED BY ADORNED EFFECT).
This allows us to get strongest Jewel stats combination in the game.
However corrupting Magic Jewels is a very risk process which very often results in bricking a Jewel making it worthless.
What you need to know is that there are 5 possible outcomes of Magic Jewel Corruption:
1. Reroll of Affix(es) (Brick Jewel)
2. Losing an Affix (Bricks Jewel)
3. Rarefying Jewel (Ads third mod but also turns Jewel into rare which disables Adorned effect) most of the time the price of such Jewel is around 20% of Initial Magic Jewel Price - for the sake of tests I do consider it as bricking.
4. Just Corrupting (No Enchant) -> Nothing changes. Jewel becomes corrupted. Unlike for normal items this is a POSTIVIE OUTCOME which greatly increases price because it makes Jewel a Target for Adorned.
5. Corruption with Enchant (Jewel get's successfully corrupted and gains an Enchant). Possible Enchants are (5-10 Str, Dex, Int, Fire Res, Cold Res, Light Res, 4-7 Chaos Res, Maim Immunity, Cannot be Hindered, Corrupted Blood Immunity and Cannot be Blinded).
Since no place currently lists the outcome chances for Corruption outcomes I've gathered 1154 Magic Jewels in order to corrupt them.
Experiment Assumptions|
1.There are 5 possible outcomes of Magic Jewel Corruption: 1. Reroll of Affixes 2. Losing an Affix, 3. Rarification (with gaining an affix), 4. Successfull Corruption without Additional Effect, 5. Corruption while gaining Additional Enchant|
-
To ensure possibility possibility of losing an affix each Jewel needs to be upgraded to 2 affixes using the Orbs. 3. Losing an affix, Rarification and Enchant are effects which cannot be mistaken with any other outcome and can be easily sorted grouped by after corruption of all the Jewels
-
No effect corruption and Affix Reroll can only be told apart during item corrupion. As such it's of paramount importance to immediatly separate items with one of the outcomes right after corruption.
-
Since Jewels do not have tiers of affixes (which require given ilvl) we assume that item level of the Jewel has no impact on corruption outcome (Enchanted Corrupted Jewels can be seen on trade website with level as low as 24)
Experiment Goals
- Finding out the chance for each of the outcomes
- Finding out the chance for each of the enchants occurring (NOTE: our sample size is huge enough for to get a good results for Goal 1. Depending on the overall chance of an Enchant occouring it may turn out that out sample size is not big enough to give us creditable results. Also if corruption enchants treat different Jewels bases differently our sample size for Emeralds and Sapphires probably won't be big enough even if chance for an enchant is equal 15%. [This also shows which kind of Jewels people use the least
Experiment Methodology
- Every single Jewel taking place in Experiment must be Magic Non Corrupted Jewel with 2 Affixes
- The Jewels are placed in separate stash tabs divided by base (this is done in order to make counting outcomes per type easy).
- Prepare destination tabs for following outcomes 1. Reroll of Affixes 2. Losing an Affix 3. Rarefication 4. Enchant
- The Jewels are corrupted without Omen using Vaal Orb following this steps:|
4a) Memorize the affixes of the Jewel you're going to corrupt|
4b) Corrupt the Jewel|
4c) If the affixes on the Jewel have been rerolled (meaning it still has 2 affixes but different) immediately take out such Jewel into your inventory otherwise proceed to the next Jewel
4d) While corrupting keep dumping Jewels with rerolled affixes into their designated tab (as your eq gets full)
- One all Jewels of given type are corrupted:|
5a) Use filter "Rarity: Rare" to quickly identify all Jewels which underwent Rarefication and move them to their designated tab
5b) Use filter "Enchant)" to quickly identify all Jewels which gained an Enchant
5c) Manually go through remaining Jewels and take out the ones with 1 Affix and move them to designated tab for the Jewels which lost an Affix|
5d) Tidy up remaining Jewels into columns so they are easy to count. (The only remaining Jewels are the ones which successfully corrupted without additional effect)
-
Repeat steps 4-5 for remaining Jewel Bases
-
Count number of Jewels for each base and each outcome which resulted in given outcome
-
Split Enchants per type and base and gather data
Experiment Results:
Outcomes for corrupting over 1154 Jewels
Conclusions:
1. The differences between outcomes probabilities are within margin of error for given base size.
2. Most likely the exact in game percentages are 25% for Affix Reroll, 12.5% for Losing an Affix, 12.5% for Rarefication, 25% for Simple Corruption, and 25% for gaining an Enchant
3. There is approx. 50% chance of Bricking the Jewel.
4. You shouldn't use Omen of Corruption while corrupting the Jewel (if you care about the money) since No Effect is a POSITIVE OUTCOME resulting in increase in value. Removing it from the pool changes the probability from 50/50 into 37,5% of Gaining an Enchant and 62.5% of Bricking the Jewel.
5. If you have a good Magic Jewel and you are considering corrupting it before sale you SHOULD NOT DO IT. The reason is that there is 50% chance of reducing the value of Jewel to 0. That is unless you are wealthy enough to average the outcome by doing multiple times (and assuming the Corrupted version without enchant of the Jewel sell for at least approx 2x price of non corrupted one).
Enchants Results:
Now as for enchants themselves the test resulted in 271 Enchanted Jewels while the number is not small given the number of possible enchant outcomes it's not a big enough base to draw conclusions and should be treated as curiosity
Enchant Outcomes for Corrupting 1154 Jewels of which 271 Gained an Enchant
Corruption Outcomes Based on 1154 Corruptions
1. Given the sample size the enchant % are very likely to be off by few %.
2. However it can be clearly seen that Chaos is rarer then other resistances.
3. Attributes and Elemental Resistances seem to share the same chance.
4. Main, Hinder, Corrupted Blood and Blinded effects seem to be the rarest of enchants with similar chance of occurring.
However those theories are drawn based the sample size which is not sufficient? Can we somehow do better or confirm them? Yes we can use the market.
The Market Results
While the perfect Jewel Selection with Enchants is likely to be severely misrepresenting the odds (Intelligence and Resistances are more likely to be bought then other enchant outcomes) most Corrupted Jewels on the Market are not perfect. Therefore while the number of Jewels with given enchant can slightly differ given it's desirability with huge enough sample size we should be able to approximate the odds of given enchant occurring by comparing number of the Jewels with given enchant on the market with the total number of enchanted Jewels at the market.
Market Conclusions:
- Fire, Cold, Lightning, Strength, Dexterity and Intelligence seem to share the same chance of occurring as an enchant. [The lower number of int Jewels is likely due to demand]
- Chaos is twice as rare as Elemental Resistances [Also may be affected by high demand]
- Maim, Hinder, Corrupted Blood and Blinded seem to be least likely to occur.
Thank you for your attention :) If you want to provide some feedback in game you can find me under nick MadderThanMad - I've few days of currency farming and exping to make up for.
Comments
Useful, thanks
Thanks for the data!
Just an aside here: I already assumed but your analysis all but confirms it. There are four corruption outcomes, nothing, lose/gain mod, enchant, reroll.
This language is much clearer and coincides with other corruption items: i.e., gems, 4 outcomes (lose/gain level, lose/gain socket, lose/gain quality, nothing) and as such each outcome is equally weighted. The enchant mods being weighted isn't new, but seeing specific numbers is nice; I did not expect the ele res to be so much higher than the stats and figured those would be about equal. I will point out however, that vaal corruption implicits (enchants) in the past HAVE cared about ilvl, so that may be affecting your results. It might not be possible to get certain enchants if the ilvl is too low.
EDIT: I wished that gaining an affix as a magic item would keep it magic that would've been very cool, but probably too OP. Interestingly enough, I wonder if you took a 3 affix rare jewel and corrupted it to lose an affix if it would turn magic. If that were the case, depending on price you might be able to make decent money that way if rare jewels with 2 affixes you care about are cheap.
Since reddit has compressed a bit tables here is a link to Google Sheet with Results
https://docs.google.com/spreadsheets/d/1ucugETAsgmegcXVYWZxrgTMN1NUEePvn1PwdIzvI5So/edit?usp=sharing
In some cases, I always Vaal magic jewels due to the change in value. For example, a high roll on sapphire with %es/crit roll can be about 1d+, and successfully corrupting for adorned can turn it into 7d+.
Nice work.
This is amazing! Thanks!👍🏻
This is awesome. Thank you
Holy moly! I only started Act 3 last night, so I'm not reading all that. I may circle back in 80 hours more playtime to digest it all. lol. I just wanted to show my appreciation for your enthusiasm and dedication. :)
Nice, I appreciate that the market numbers reinforce your numbers with pretty similar percents of enchants.
4. You shouldn't use Omen of Corruption while corrupting the Jewel (if you care about the money) since No Effect is a POSITIVE OUTCOME resulting in increase in value. Removing it from the pool changes the probability from 50/50 into 37,5% of Gaining an Enchant and 62.5% of Bricking the Jewel.
This is great, but as an FYI, if a particular enchant does not exist, and you are targetting enchant specifically without care for the monetary value behind, then you need to use the Omen
Several int stackers have done this!
I still dont understand the hype for the adorned. I feel like equipping all rare jewels with good stats for your character is the same or better than the adorned diamonds effect on a magic jewel. For example yesterday I was very lucky and found an emerald with 4% attack speed, 4% quarterstaff attack speed, 14% elemental damage, and 21% increased damage with herald skills. I corrupted it, and it added a 5th mod 10% lightning penetration. That jewel has 5 good stats for a monk. I dont think a single corrupted magic jewel with a 100% adorned diamond can be better than that. I can see the adorned being better for rangers though with the quiver bonus but that's about it.
I guess it's worth saying for the 95% of peasant gamers with no diamond build, the negative outcomes are in the minority.