MMO News and theorycrafting for advanced MMO gamers. News and articles that relate to your gameplay. World of Warcraft, SWTOR, Guild Wars 2, Rift, TERA, Eve Online, Star Wars the Old Republic, Diablo3, The Secret World and all Western AAA MMOs

Hello There, Guest! Register

 Softcaps discussion
03-13-2012, 01:10 AM (This post was last modified: 03-13-2012 01:12 AM by Kaedis.)
Post: #93
 Kaedis Super Moderator Posts: 1,587 Joined: Sep 2011 Reputation: 7
RE: Softcaps discussion
Quote:The extent in wich stats scale with each other isn't explored yet and without a parsable combatlog (wich is needed for large enough samples) it won't be any time soon.
Parses might not be a must have for stat weights, but without them I wish you best of luck getting a significant sample size - and shooting a target 100 times doesn't cut it for a solid statistical test.

Really starting to get tired of claims like this. Parses are qualitatively inferior to simulations for the purposes of analytical calculations such as stat weights and comparative dps output. With parses, it is impossible to factor out such variables as player skill, RNG, relative gear quality, and unequally favoring boss mechanics. Simulators are by far the more reliable. Gear and "player" skill can be rigidly controlled, as can boss mechanics. RNG can be eliminated as a factor by running many thousands of iterations of combat per reports, compared to the single iteration present in a log parse.

On a side note, the simulators take into account the interaction between stats. Even though Alacrity might increase the value of your other stats somewhat due to that interact, you are still losing dps stacking it (if it's your worst stat). This is mathematically irrefutable. It's a basic law of derivatives. Let's take an example: you're on a plane which has a Z/X slope of 0.1 and a Z/Y slope is equal to X. If you are at X = 5, your Z/X slope is 0.1 and your Z/Y slope is 0.5. Increasing X by 1 increases the "effectiveness" of Y, by increasing your Z/Y slope to 0.6 (20% increase!). However, increasing X by 1 only increases Z by 0.1, whereas increasing Y by 1 increases Z by 0.5. In fact, as with all systems, the optimization point for an interconnected system (in particular one in which increasing variable T either decreases the value of further increases in T, increases the value of increases in all other variables, or both) is found by increasing the highest return-per-point variable until all variables have identical marginal values. In the case of the above plane, if you're starting at (0,0), you increase X until the Z/Y slope equals to Z/X slope, which occurs at X = 1, then increase Y from then on.

Even Angels must kill from time to time...
 « Next Oldest | Next Newest »