Non uniques forging

I m glad that people used that much my previous topic, but sadly it suddently went all wrong ^^
Anyway, i still have tones of newbie questions, so i am opening a new thread (and one day, i will maybe figure out what you were talking about).

I am still crafting my very first silver item, so sorry for my unexperienced questions…
I am trying to figure out what i will get if i build a gold ring.
Is this sheet correct ? (i consider a base perk value arbitrarly at 1000 for green 1* item; and an increase coefficient of 6.13 between green 1* and titan 5*:
When i go from bronze titan 5* to silver green 1*, i use a normalization coefficient of titan5*_value/3x2 as i switch from a 2 perks bronze ring to 3 perks silver item; and i also use a “reduction” coefficient of 4.37 as i switch from a titan5* to green1* item
i also used a normalization /4x3 for silver to gold and same reduction coeffcient which gave me that following sheet:

(perks 3 and 4 are from a bronze titan 5* item disassembled)

=> is it correct that for the final gold item i will get a slightly lower value for perks 1, 2&3 compared to the bronze 5* items used?
=> it is better to use as final disassembled item the one with the most valuable perk (demo for exemple)?
=> from your point of view, do you consider that it is better to craft first to gold a ring or another item (not taking into account that a ring will be usefull for many heroes while another item will be usefull just for one)

I read something from the dev about unlimited loop of refinement. Never understood that part. Is it like the unique forging system? if i use an old gold 5* item with 4 perks on another gold 5* item with exact same 4 perks more up to date, i will get a gold 5* item with even higher perks value ?

1 Like

That’s wrong.
The last perk you add to the ring will be the lowest value of the 4.
Each max forge cycle increases value by around a factor of 1.38
The first 2 base perks will be around 92% value, with the 3rd perk around 90%(there’s some give or take here)
Normalized by 2/3*1.38=92%.
On gold it will be around 95% for the base 2, 91-93% for the 3rd one and around 69-70% for the last one.
Your mistake is not taking into account that the values on the gold item are normalized by 1/2 since there will be 4 perks where there should be 2, but since the first 3 perks went through the max forge cycle once as a bronze and again as a silver item, and as a gold item, whereas the 4th will only have went through it as a bronze and as a gold item, so it lacks around 38% increase to reach those other ones, which would be .69x1.38=~95%
So the first 3 perks were halved because of gold refine, and then multiplied by ~1.38per forge cycle, so around a 90% increase(1.38x1.38). Thus .5x1.9=.95
The 4th perk was multiplied only once, so .5x1.38=.69

1 Like

Ok, so for perk 1 and 2, i am quite close for value of green 1*: 934 (93.4% of previous green 1* value), correct?

But then, I don t get it… the first 3 perks are halved (though it will be 3/4) during silver to gold refine… but they already were normalized by 2/3 when they moved from bronze to silver
so if i calculate like you did:
for perks 1-2-3:
(1.38)x(2/3)x(1.38)x(1/2)x(1.38)=0.87 (87% of initial value) (coefficients used are: (bronze_forge_cycle)x(normalizeBronzeToSilver)x(bronze_forge_cycle)x(normalizeSilverToGold)x(bronze_forge_cycle)

for last perk (n°4):
1.38x1/2x1.38=0.95

You’re multiplying by 1.38 one extra time. You’re comparing green values but multiplying by 1.38 one extra time when it goes from gold green 1* to titan 5*, but that 1.38 can’t be counted because you’re not turning it into green again.
You either compare green with green which would be 100%x1.38x1.38x1/2 for the first 3 perks
(Bronze green 1* to silver green 1*)x(silver green 1* to gold green 1*)x(gold half normalization)
As for the last perk it’s just 100%x1.38x1/2
(Bronze green 1* to gold green 1*)x(half gold normalization)
And the normalization for the first 3 perks is not half of the 2/3, it’s half of the full bronze 5* equivalent values multiplied by the 1.38 factor twice

ok, so my sheet was correct, except for perk 4:
i get about 98.2% on gold ring of initial perk for perks 1 2 and 3 (i think the correction factor depend on the type of perk, not because it is the third one, anyone can confirm?)
For perk 4, i get a final green value of 70.1%, close to your value… But to get thoose number, i need to normalize by 3/4 for perks 1 2 and 3 when going from silver titan to gold green, while i need to normalize by 2/4 (0.5) for perk 4 when going from bronze dismantled for the gold green… this is pretty strange!?

And now i added the sheet for a vest or a boot etc…

I assume base health/cmd/damage on green gold are correct (no normalisation factor, so it is pretty easy), but based on calculation you shared for ring, i get super low values for final perks :o (normalization factor of 1/3)

What is this 1.38 you are referring to? Can you give an example with values please?

That’s 1.0051^64=1.38482= 38.482% increase if you compare a min forge to a max forge.
Which is 64 extra stars dismantled for .51% increase each
I’m not 100% sure if it’s exponential. If it’s not then my calculations are wrong and it should be .51%x64=32.64% increase per max forge cycle.
Example would be let’s say I min forge an item from green 1* to titan 5* and it has 100k value perk, it’d have 138482 value if I max forged it

Edit: it might also be exponential, but only every forge, that would mean it’s actually 1.0153^16x1.0204^4=1.38227=38.227%

You’re normalizing by 3/4 because when you went to silver you normalized by 2/3, then when you went to gold you normalized by half. But this half is based upon the full value, not the 2/3.
To factor that in you need to multiply the silver value by 3/2 to cancel with the 2/3 and then divide by 2
3/2 ÷2 = 3/4

And no, the correction factor is not because of type of perk. I’ve seen an item being refined with nearly the same values for resistances, fire on base ring and lightning as 3rd perk, lightning was like .94 on gold and fire was .96 or something. Lightning was def lower percentage.

The problem, is that if everyone uses his own terminology, we get lost into the meaning and have disagreements. And you didn’t answer my question actually.

Are you saying that 38% is the difference between a normal forge and a max forge? So that 100k is potentially a 22600 green, which can be forged either a 100k or 138k if max forged item ?

Whats the 0.51 and why are you counting it with stars to multiply it?

Sorry for asking, I don’t want to insult or say your thought process is wrong, so i am looking to understand your way of thinking and calculating.

From what I’ve seen and heard, each forge step adds 5% stats if you use a 1*, and for each extra star it increases by .51%, so using a 4* would be 6.53%, using a 5* 7.04%
The only question is how does the exponential work?
Also, when it jumps color there’s a different % which is also increased by the same %/star dismantled. I’m still testing tho, have a couple max forged underway that I want to compare to my screenshots
It might actually be different, could be per forge step which would be like 36% And change. I’m still trying to figure it out

@Zhadess i think you are not accurate on several topic. The +5%forge is indeed a forge increase for most of perks, but not all of them. For superforge (forging with 4* items) you get +6.5% for each forge step (color change use different values)
Anyway, the total amount of perk increase is roughly x6.1 to x6.3 for most of perks from a green 1* to titan 5*
Then, when you refine, you have a normalization factor (the one i m trying to get with 100% certitude), and a “don t have a name for it” (anyone have a standard name for it ^^ ?) factor due to the regression from titan 5* to green 1*

Can someone confirm/correct my sheet?
I am not much interested with the effect of forging/superforging, has i have most of coefficient for this topic and it will only affect the “6.13” value of my sheet
I am more willing to understand clearly how the normalization factor works, so i can use the best perk order…

I didn’t see that you mentioned the items that are not rings. Yes, you get a normalization of 1/2 for 2nd perk and 1/3 for 3rd perk. On gold item it will be 1/3x1.38x1.38 for base and 2nd perk and 1/3x1.38 for the last perk.
BUT
What you didn’t factor in is that the base for the item will go up, I’m not sure by what %, but it’s around 12-14% to silver and 18-20%(compared to bronze) to gold
So for ex you have a weapon with loh, refine it with FB and then refine again with demolition

Loh and FB will have around 1/3x1.38x1.38 of their bronze values and demo will have around 1/3x1.38 of its base bronze value
The damage on that weapon will increase by around 13% On silver compared to bronze and around 19% On gold compared to bronze
Same goes for vest, boots, cape, gloves and shield

The normalization is very simple:
Rings should have 2 perks
If they have 3, 2/3x3=2perks
If they have 4, 1/2x4=2perks

Normal hero items should have 1 perk
If they have 2, 1/2x2=1
If they have 3, 1/3x3=1

If you refined a weapon with 1000 attack speed with a 1000 stun item and a 1000 stun glove with a 1000 attack speed item and used 1* for all the forges you should have a weapon and glove with 500AS and stun each

That’s why you need to max forge, otherwise refining is a waste

Oh and about the other perks that don’t increase by 5%, those are even more affected by max forging as they get 2.3% 1*, 3.83% 4* and 4.34% 5*

You should research first, all the parameters, is my advise. Most of the data is correct, but there are some differences that we have also mentioned. Actual examples, will help others (me too) understand more what you want to say.

For me the 38% has no real value (to use it), but we have to make sure to explain it and refer to it with the correct way, so people are not miss-led. It is however pretty close to the actual number.

When you say exponential, are you referring to the leveling up, or the actual cycle?

And also, you said I’m wrong and that it’s 6.5%
It is 6.5%, but we’re not comparing green items with titan items if were talking about the factor of normalization. You need to compare titan 5* with titan 5*, which is what actually matters

The cycle, not levelling up. I don’t know how the bonus per star will add up. If it’s per forge step like
Green 1 to 2
Green 2 to 3
Etc…

Or if it adds each star separately. That doesn’t make too much of a difference, what I’m mostly wanting to find out is if the .5% per star adds additively or multiplicatively to each other. I’m 99% sure it’s multiplying, so it should be around 38.5-38.8% per cycle

Ok let’s make an example
I get a ring with 100k fire and ice resist, a vest with 100k poison resist and boots with 100k lightning resist, all of them bronze titan 5*
I’ll refine the vest into the ring to silver then the boots into the ring to gold
The final item will have
APPROXIMATELY
Fire and ice resist(base)(100K)x1/2x1.385x1.385=95900
Poison resist should have the same 95900, but for some reason always has a little less. I’d guess between 91500-93500
Lightning resist(4th perk)
Should be
(100K)x1/2x1.385=69250

Now on a NOT ring item

Lets say I have a vest with 100k fire res, one with 100k poison res and one with 100k ice res

I’ll use the fire as base, refine the poison into the fire one, then the ice into the silver fire+poison one

Final values on gold
SHOULD BE APPROXIMATELY
Fire (100K)x1/3x1.385x1.385=63940
Poison should have the same 63940, but will probably have a couple % less
Ice (100K)x1/3x1.385=46167

All max forged ofc since I’m using the 1.385 factor, which could be wrong by a % or 2

Of course its multiplying to each other. Which is why you have to use real values and test it. Its fairly simple, and cheap to test on greens.

That’s why I said I’m 99% sure, so it should be the ~38.5%