Armorsmith Archive

I don't see why it matters, with perfect res (no loot) the best you can get is 8200, and so far all loot we have seen have been health bonuses, nothing that raises any protection. I very much doubt we will ever see loot that give +1500-2000 because if that loot was put into a lower level armor we might end up with the kind of armor the devs try their best to avoid.

---

MrGouda wrote:

Come on people. We have no idea what 10000 extrapolates to. name a number. I'll give you a polynomial that fits existing data provided by Okram and through that point with an R^2 of 0.99.

NO ONE KNOWS what 10000 relates to our data stops at the mid 50's percentage. And those of you looking at the natural curve of the data are silly. Mathematics is now used to fit the data. More data is needed above 53% resists to determine the answer to this questions.

By some equations with excellent fits, 10000 is only in the 80s% resists.

---

There are a few problems with this line of thinking:

1. Blixtev already came in here and said something about being glad that people were able to reverse engineer the equations.

2. Another poster (can't remember who) was able to reverse engineer our regression equation a little more and came up with a pretty standardquadratic diminisherexpression:

60 * (AP / 10000) + { [ (10000-AP) / 10000 ] * (AP/200) }

---

Brutus_Krylop wrote:

---

MrGouda wrote:

Come on people. We have no idea what 10000 extrapolates to. name a number. I'll give you a polynomial that fits existing data provided by Okram and through that point with an R^2 of 0.99.

NO ONE KNOWS what 10000 relates to our data stops at the mid 50's percentage. And those of you looking at the natural curve of the data are silly. Mathematics is now used to fit the data. More data is needed above 53% resists to determine the answer to this questions.

By some equations with excellent fits, 10000 is only in the 80s% resists.

---

There are a few problems with this line of thinking:

1. Blixtev already came in here and said something about being glad that people were able to reverse engineer the equations.

2. Another poster (can't remember who) was able to reverse engineer our regression equation a little more and came up with a pretty standard quadratic diminisher expression:

60 * (AP / 10000) + { [ (10000-AP) / 10000 ] * (AP/200) }

---

That was me, and it verified the more popular equation, and more or less proved that this is the equation in use. It is a simple linear combination/diminishing returns equation, and is easily scaled, fitting the data, all of which supports the view that this *is* The Formula. This expression agrees with 10,000 = 60% (and 11,000 = 60.5%, before the result starts decreasing. ie 12,000 = 60%)

---

Soljin wrote:

Eeek! *sits alonehuddled in cornerin confusion*.

---

Yuo and me both