Shadowcraft looks broken - where to reforge?

[Amaroq]

An alternative to Shadowcraft would also be AskMrRobot.

[Shimitsu]

Yay for heuristics!

[Deleted]

Ahh but we're ignoring part of Sorazza's post;...

No, you are ignoring a BIG part of my post.

I give you, shadowboy and the other, I guess either mathematicians or, more likely computer scientists (or at least decent programmers) here one thing: You are right about the math. I am neither, I am a microbiologist. I don't care about complex statistics, thats what the guys in the department for computational biology are for. I wear a lab coat and do funny things to the DNA of bacteria.

What you are ignoring is, that neither nature nor reforging cares for a mathematically perfect solution...nature cares for what works. My reforge method is very simple: I approximate a stat priority list, then shuffle around stuff from its lower end until I get to hit and exp cap. Once thats done, I reforge into my strongest sec. stat, and again shuffle things around with hit and exp until I am satisfied. And yes, I do that after every raid if I get a new item, while having fun in vent.

I have used a simulation tool (I don't even recall which) ONCE in the last two years, out of curiosity. It shuffled around some of the refs I did, with an end result that would have netted me a whooping amount of 71dps...thats right, 71dps.
Consider now the fact, that no one is a machine, therefore we make mistakes, and even doing ONE RvS when you already have 5 CPs flushes more damage down the drain than that reforge would give you over the course of 2min...That greatly diminishes the reasonability behind reforge-simulation. Now why is that...because, while shadowboy is right about 13 billion possibilities, the reforges that actually make sense are very limited...this game is simple, it's simple to understand, and so is the stat-system. Yes, with complex simulation tools you can calculate an optimum, but unless the guy who did the reforging is inexperienced in understanding the system and/or playing a rogue, the end result (and by that I mean the actual numbers the PLAYER, not the simulation, pulls in a fight) will usually be close enough to what the calculated reforge does, that it makes no real difference.

What my post did NOT try to tell anyone, and I think thats the part you misunderstood, is that our brains were capable of doing the calculations to find the optimal reforge in a reasonable amount of time...as you have so elaborately pointed out, thats a challenge for which computers are much better suited. But, and I will say it once again, with some time and enough knowledge its absolutely possible to find a competitive reforge just using the tools nature (and blizzard) equipped you with.

I did ALL my reforging manually, from the day it became available, and I will continue to do so in the future. I was always able to play my rogue successfully and I never had to worry about defending my raidspot. The reforging is a fun part of the game for me, sometimes I spend hours speaking to that etheral in OG, spending several hundred gold in reforging and re-reforging, until the numbers look like I want them to. I wont let a piece of code take that away from me, just so I can say: "Hey look, in theory I can do a better job than you, because the computer said so!!!"

Oh, and you know what else is cool? I am not lost if the reforge tools go haywire

[shadowboy]

Ahh but we're ignoring part of Sorazza's post; you're not just supposed to decide on the manner to use SimC to generate your own perfect reforging stat weights to brute force a method, you're supposed to both manually do the reforging, and manually run all simulations thereof in your head or on paper. I can only imagine, given even a single iteration, how long it would take me to write out 5-10 minutes of combat based on what I believe to be the best method of playing my rogue, applying dynamic changing values to each of the attacks I was making, and checking my energy generation every second.

It's just not going to happen. Using sample weights from a good source (see: EJ or any other site that has collected information from the BiS and pre-BiS profiles to see if there are often inflection points which the poster is likely to mention, as reviewed through the sim process at SimC or another well-written simulation engine), you can personally write a "decent" reforging scheme manually to meet your requirements, but using any tool at all is a lot more efficient, and finding the "best possible" can use a more greedy algorithm if someone bothered to write one for certain specifications, but would probably miss inflections where dropping just 1-10 points under hit/expertise caps would allow a different reforge to get more of the better secondary stats for your character.

Story short: if you want really good stat weights, run yourself through SimC repeatedly to check for changes in relative stat weights, but realize that each step you take closer takes more work for less reward. The initial (manual reforge) run should be within a couple % of maximum DPS, and any reforging calculator with sample weights from a good source should be MUCH closer (<1% from ideal, and probably really, really close).

From a manual, not computer source, however, you will not reforge perfectly in a relevant timeline... and with it, you should just get as close as it's worth your time to get.

I'm not arguing that you SHOULD re-scale your stats at every reforge iteration. In fact, I'm arguing the OPPOSITE. I'm saying that doing an exhaustive global maximum search is not feasible. I'm arguing that just using existing reforge tools (such as wowreforge), and the approximations that come with them are quite good enough.

---------- Post added 2012-09-07 at 10:15 PM ----------

I give you, shadowboy and the other, I guess either mathematicians or, more likely computer scientists (or at least decent programmers) here one thing: You are right about the math. I am neither, I am a microbiologist. I don't care about complex statistics, thats what the guys in the department for computational biology are for. I wear a lab coat and do funny things to the DNA of bacteria.

My background is actually theoretical chemistry (protein folding, in particular), but yes, it was computational in nature.

I also agree with your post. In fact on the beta I felt fine without having a reforge tool.

[Deleted]

My background is actually theoretical chemistry (protein folding, in particular), but yes, it was computational in nature.

I also agree with your post. In fact on the beta I felt fine without having a reforge tool.

I jumped to the conclusion because your posts read exactly like the handouts we got in our "Basic Bioinformatics" lecture last year. Real fun actually, 50% micro/molecular biologists who were fast asleep half the lecture, and 50% guys from structural and computational biology listening intensively

[shadowboy]

I jumped to the conclusion because your posts read exactly like the handouts we got in our "Basic Bioinformatics" lecture last year. Real fun actually, 50% micro/molecular biologists who were fast asleep half the lecture, and 50% guys from structural and computational biology listening intensively

I find biology dreadfully boring. I find physics a much more interesting scientific endeavor. Even though I studied proteins it was very far removed from the biology side of things. Even when I was doing experimental biophysics during my lab rotation it was quite distant from biology.

[Verain]

1)- You put your stats to mastery.
2)- You put your stats in the boss's butt.

Once September 25th hits, see the rest of the thread.

[Deleted]

I find biology dreadfully boring. I find physics a much more interesting scientific endeavor. Even though I studied proteins it was very far removed from the biology side of things. Even when I was doing experimental biophysics during my lab rotation it was quite distant from biology.

Same here with physics, math and computer science. I guess there is a cup for every tea.
For me, nothing compares to the sheer beauty, ingenuity and the capacity to learn and adapt that is innate to biological systems.

[Kael]

I find biology dreadfully boring. I find physics a much more interesting scientific endeavor. Even though I studied proteins it was very far removed from the biology side of things. Even when I was doing experimental biophysics during my lab rotation it was quite distant from biology.

I was an English Lit. major with a minor in writing that I pursued into poetry but my work is in computers. What? Always enjoyed working with numbers, though, so it's a side-hobby I guess.

I did my own reforges when all of the tools were (briefly) down, but it does take more time for a worse solution, when I could take 10, maybe 15 seconds and toss a machine at the stats and use the best results it picks up, heh. Inevitably I find I've overshot hit or expertise and I was going back over my gear going "where would a 40% result be ~45 rating closer to the cap..." and ended up wasting almost as much time as trying to reformat my UI... and then switched gear anyway because the T12H set bonus isn't worth much compared to what it was before ><

Granted none of this matters much (to most of us) with gameplay changing significantly in a short time until ~T16 and WoLogs rankings for T13H having been locked, but I still take some odd pride in doing things the right way on my rogue. My bear, ironically, I take far less seriously and rank much better with. /Sadface?

[Nicedood]

I would say askmrrobot but I am unsure my self.