Response to Loot Blog


"Personal Loot"

The good (sort of). This is actually not that bad (I said “a lot” of the post was bad, not all of it). Personal Loot is of course pretty dumb in any non-LFR context but in LFR it seems like a necessity. GC explains this all pretty well. I don’t like it, but only for the same reason I don’t like LFR—the raid game should be about working with people as a team. To the extent they’ve introduced a version of raiding that’s not about that, the loot system has to reflect a setting where people don’t care about each other. The concept of LFR is turning raiding into a single-player experience, and loot rules that minimize any possible impact that other people can have on your life are consistent with that.

"Bonus Roll"

The bad. This is a points system (ironic given the later section about toning down points systems) but you get paid in lottery tickets. This is what I would come up with if I were trying to make loot even more frustrating. Before you’d save up points to get gear (a system with pros and cons we’ve all gone over at great length), now you get the buildup of saving points up only for disappointment when your loot lotto scratch-off card gives you another few gold.


The ugly. This is another item modification subsystem to tack onto enchanting, gemming, and reforging. Unlike the others it doesn’t even have the illusion of skillful choice; it literally has no purpose but to be a chore. I don’t know what more there is to be said here. What player-friendly purpose does this ostensibly serve? The system doesn’t give me items now—I still have to get those the normal way—but then I have to do dailies or whatever to level up my gear. Why do I have do chores if I’m already getting the gear I need by playing the game normally?

This might have a silver lining—if it undoes the long-standing issue of giving players current-tier gear without doing current-tier content (which has been at the center of my biggest complaints about WoW in recent years). But that remains to be seen, and I’ve seen no indication that Blizzard intends to dramatically reverse course on it recent design paradigms in that way.

The NYTimes AIG Tax Story

Since my day job is being a corporate tax lawyer, though I’d comment on this interesting story: (via TaxProfBlog). I don’t know any details of the case and I’m not going into an elaborate analysis, just explaining a few basic points as I’m wont to do when a tax story goes mainstream.


Loss — You’re all familiar with  the concept of income being taxed; losses are basically negative income. If you lose money the resulting negative tax offsets (in form of deductions) your income tax for the year. So in tax speak, having a “loss” is a good thing: it counteracts tax on “gains.”

Net Operating Loss (NOL) — If your tax losses exceed your tax gains in a given year, you don’t actually pay total negative tax (get money from the gov’t). You just get to store the excess loss to use against income in later years. This makes sense: if you lose money in one year and make money in the next, there’s no reason for your tax loss to vanish because of the arbtirary demarcation of calendar years, you should get to offset them.

Section 382 — Now for a simplification of a really arcane part of the corporate tax code. If a corporation has NOL’s sitting in it, and it undergoes a “change in control” (very loosely speaking, more than 50% of the shares change hands), those NOL’s get limited to a certain cap. The cap is based on the value of the corp; if it’s 0, NOL’s at the time of the control change will become permanently unusable. This is a very bad result for the corp. Those NOL’s represent tax relief arising from losses the company incurred, and losing them results in worst-of-both-worlds treatment: the corp would have gotten no tax benefit from years it lost money, and would pay tax fully in upcoming years where it makes money.

Why do we have this rule? Because otherwise corporations with NOL’s would be too easy to manipulate. The corp’s owners (who have no use for the NOL as they’re losing money every year and have no income to offset) could sell the corp to some other business who’s making money, who would use the NOL’s to offset their income. NOL’s in shell corporations would be an easily transferable asset, a sort of prepaid tax offset tradable on the open market. So Congress enacted 382 to prevent use of NOL’s after a corp changes hands.

Now it’s true what Treasury says—that policy is not applicable here (nobody was passing AIG around as a shell containing only tax deductions). But understanding the basic background above, at a first cut you can see that something is amiss. Normally a company that loses a lot of money gets NOL’s stored up—and if it turns things around and starts makgin profits, the NOL’s protect its income until it gets back to 0. But AIG didn’t actually start making profits on its own—it was bailed out with a large infusion of government cash in a sort of unique situation. They did issue stock to the government (79.9% of the company) in exchange for the cash, but still, this was nothing like an ordinary investor coming along and providing capital.

Point is, AIG had huge tax losses, which did represent real economic loss of the business. But instead of using those losses to cover income back to the zero point, they were sort of suddenly teleported to a new position where they were much less in the red, and began operating again from there. It does somewhat resemble a bankruptcy in that way. So just on basic tax principles, it is not clear why their NOL’s should persist through that event. What makes this confusing is that sec. 382 happens to pick this up even though it was meant to deal with a different problem altogether. AIG seems to have successfully argued to Treasury (awkwardly, now its 80% owner) that 382 wasn’t meant to cover this scenario, and while they may be right, the resulting situation is not totally coherent. They are continuing to use tax losses even though the underlying economic loss was counteracted by the bailout.

I’ll stop here, but hopefully that clarifies somewhat what’s actually going on in this story, and you learned a little about corporate tax on the way.

Review of Super Mario 3D Land

After a little break, I just went back and finished all the content in Super Mario 3D Land (all levels beaten, all coins, all flags, secret bonus level finished). Figured I’d write a few things.

Overall the game is great. It’s basically an attempt to translate the feel of SMB3 into 3D. It feels as though they went back to 2D mario games, asked “how can we make a 3D game out of this”, and then did so in a direction totally different from the one they chose for Mario 64 and all its descendants. It’s a return to the basic Mario concept of progressing through a (usually) linear level, only now with some width to the game space that allows for new kinds of level design. Reinforcing the feel of old-fashioned Mario is the fact that you have no control over the camera—it remains at a pre-scripted angle at each point in the level, allowing segments where you move in the familiar left-to-right direction as well as segments where you progress towards/away from the camera or other odd angles.

I won’t turn this into a review of the 3DS, but let me say that it was the first 3DS game I played, and is a solid introduction to the system (honestly, this is the sort of thing the 3DS really needed at launch). The 3D is used well for this game—precise jumping in a 3D environment is definitely aided by having actual depth perception. A few other 3DS features are used in silly or pointless ways, such as StreetPass and gyro controls, but they don’t at all get in the way of the gameplay. Well, aside from one perennial DS issue: why do I have to use the touchscreen to release my “spare” item when the 3DS has 6 buttons and Mario games only have 3 controls (jump/dash/crouch)?

One difficulty: SMB3-style jumping is definitely weird at first with an analog stick. Particular in sections that are functionally 2D left-to-right platforming, you feel really silly when you miss a platform towards/away from the screen simply because you weren’t holding the stick exactly to the right. I got used to it pretty quickly though, and it’s hard to complain since we’ve been doing jumps with an analog stick ever since mario 64.

Highlighting one awesome mechanic: the red/blue flip plates. Basically, it’s a set of two platforms, and at any time, one is solid and can be stood upon while the other is empty. What’s new is, rather than alternating according to a regular rhythm, they flip every time you jump. Even for a very seasoned Mario player, they are disorienting the first time—I guarantee you will have instances where you jump comfortably at a solid platform and realize once you’re in the air that you’re about to land on nothing. It’s a novel Mario idea, and not anything gimmicky that relies on the 3D—simply a new kind of platform showing that they’re still innovating on the basic ideas of the game.

Worst new idea in the game: when you’ve failed a level 5 times, you’re given a special Invincibility+Leaf powerup with no duration that helps you beat the level. If you somehow fail 5 more times after that, you’re given a P-wing which teleports you to the end. This was nicely satirized in Penny-Arcade here: . I won’t go off too much on the concept of helping beginners through a game—done correctly it can be a very good addition—but want to note a few reasons why this is clumsy:

—5 deaths is very low limit. Attempts at a hard Mario level can take a few seconds each, and probably rarely last a minute. So the amount of time this game allows you to attempt a level before deciding to need to be “skipped” past it is much less than 5 minutes. I’ll let you make your own judgment about that.

—You can choose not to pick up the powerup each time, which does generally prevent this from ruining the gameplay for people who want to beat all the levels honestly. But personally I think that “helping” people in this way should be at most an option that can be enabled.

—In particular, this game has a special award on your file for beating the game without ever using the special powerup (think of it like a Feat of Strength). But its considered failed when you allow the thing to even appear, not when you actually use it. So (based on the forums) even the players who want to actually beat all the levels and ignore the special leaf tend not to get it. This seems like a minor point, but honestly these sorts of “beginner helper” features in games would be a lot less annoying if they could either be turned off, or if there were a correctly working acknowledgement of whether you used it or not.

—The P-Wing seems really over the top. So the “help” to new players consists of skipping them past the level entirely? That doesn’t even pretend to meet the usual justification of “letting them see the content.” The only content they even get to see is the flagpole. At some point I think that, if a player can’t beat a level with permanent invincibility and raccoon tail, there’s no meaningful way for them to experience it. Besides, what’s going to come after the level? More, harder levels that they still can’t beat. Suffice to say, letting them skip every level is not a much better situation than having them stuck at the one they’re on.

All in all, great game. Really a solid introduction to the 3DS while also calling back to the feel of SMB3 in a way that few Mario games have.

e: PA actually did a TV episode of the creation of that strip, where they discuss this game heavily (plus some great personal stories):

Bloodlust and Execute

A discussion popped up on twitter about a point of WoW math that’s surprisingly deceptive: is it better to use Bloodlust in the post-35% range of a fight because of slight DPS boost that some classes get in this range (“execute” mechanics)? This was widely assumed to be the case for a while, but a big discussion on EJ made people realize it wasn’t, if you think about it carefully. Someone worked up a very long proof here: . But let me see if I can give a short intuitive explanation.

To be clear, the point here is to show that there is no inherent synergy between these two effects. Lots of things impact correct choice of Bloodlust timing in reality, but the point is to explain why this commonly-cited factor is actually a mathematical fallacy. To do that we will use a simplified model where Bloodlust and execute are the only things affecting DPS.

Say a boss has 1000 HP, and in execute range (under 35 HP) DPS on him is increased by 25% (exaggerated from reality to make the effect clearer). Wrap your head around this bit of intuition: for all intents and purposes, the boss has 930 HP and the execute phenomenon can be otherwise ignored. Why? Well, the last 350 HP on his bar are essentially “worth” only 0.8 HP each due to the execute effects*. The raid has to do 650 damage to get into execute range, and then 280 more damage to remove the final 350 HP. This is going to be true no matter how the damage to the boss is distributed in time—when the raid has output 930 points of damage, the boss will die. All that the execute effect represents is a slight reduction in effective boss HP.


To make sure, we’ll work the example. Say the raid does 1 DPS unbuffed. Without Bloodlust, that boss takes 930 seconds to kill as we’ve described (650 pre-execute, 280 in execute range).

With Bloodlust at the beginning. During the 40 seconds of Lust, 40*1.3=52 damage is done. After 598 more seconds to do 598 more damage, the boss is at 350 HP and execute starts. At that point, it takes 280 more seconds to die, as before. Total: 40 + 598 + 280 = 918 total seconds to die.

With Bloodlust at the beginning of execute range (650 seconds already passed): During the 40 seconds of Bloodlust, the raid 52 damage, but the execute buff increases this to 52*1.25 = 65 damage. The boss has 285 HP left. Due to the execute bonus, this takes 285/1.25 = 228 more seconds to kill. Total: 650 + 40 + 228 = 918 seconds to die.


Final math note: note that Bloodlust saved 12 seconds in both cases. Well, Bloodlust increased DPS by 30% for 40 seconds. So in other words, Bloodlust lets the raid accomplish in 40 seconds what it would otherwise accomplish in 40*1.3=52 seconds. And this is actually another way of thinking about our whole result here: no matter when you use Lust, it lets you do 52 seconds’ worth of DPS in 40 seconds, saving 12 seconds in every case.


So in practice, when should you lust. Well, by and large it’s fight-specific factors that determine this—is there a burn phase, a phase with adds, a phase when everyone can DPS the boss uninterrupted, etc. But say a fight doesn’t have any particular strategic time to use it, when is it going to give the biggest benefit?

Well, there’s one thing we’ve been leaving out of the above: other DPS cooldowns. There is a benefit to using them during Bloodlust, since their effect will be amplified. So in the absence of other strategic considerations, the time to use Bloodlust is when it can be coordinated such that most or all of the DPS have their cooldowns ready. This is virtually always going to happen at one time every fight: the beginning. After that, different classes will have different cooldowns at different times, but the beginning is the one time when everyone simultaneously has everything possible ready to throw at the boss (including pre-pots to boot). That’s generally going to be your best time.


*If you’re asking “why did he say ‘0.8 HP’ and not ‘0.75 HP’, good question. It’s a mathematical subtlety I glossed over. Think about it this way: if the raid does 25% extra damage, then in order to do 100 damage to a target, you have to do 80 pre-buff damage, not 75. (75 damage would only cause the boss to lose 75*1.25 = 93.75 HP). So a 25% DPS buff effectively devalues target HP by 20%—it’s just a quirk of how the concept of “percent” works.

Mythology of Zelda

I’m a few hours into Skyward Sword (incidentally, the game appears to be rather long). It’s very well-done and I recommend it even as someone who hadn’t been enchanted by any titles in the series for a number of years. I’m sure I’ll say a lot more when I’m done but, was having some thoughts on the story that I wanted to write down.

Skyward Sword continues the basic story trend of most major Zelda titles—Link’s quest to save Zelda and save Hyrule. Putting aside handheld titles and offshoots like Four Swords, this is the Zelda story we’re all familiar with. The only major deviation is Majora’s Mask (and really Link’s Awakening—even though it was a handheld, it was only the fourth game and a major release, in addition to simply being one of the best Zelda games even now).

But I want to focus on the original Zelda, Link to the Past, Ocarina of Time, Wind Waker, Twilight Princess, and Skyward Sword. What’s really unique about these games is the relationship their stories bear to each other. They’re not sequels (there’s some contention here I know, but I find the interpretation that all of these events happened in sequence to be rather strained to say the least). But they’re all not retellings of the same story, the story varies quite substantially. So they’re essentially variations on a theme.

Even Hyrule itself follows this pattern: it tends to have a Death Mountain, a Lake Hylia, a Kakariko Village, a Hyrule Castle, a Lost Woods, and other things. Some or all of these things are in each game, some other things are unique to each vision of Hyrule, and the configuration changes completely. Wind Waker had it underwater, and Skyward Sword has it under an impenetrable cloud cover. But it’s always the familiar Hyrule.

Similarly for the game’s story. Link seeks the Master Sword and explores a series of ancient dungeons on a quest to find the Triforce before Ganon, along the way rescuing (but also being helped by) Zelda. Again, enough of those elements are always there to make the story recognizable, although the presentation changes drastically. In what’s probably the most well-known an memorable Zelda games to current readers, Ocarina of Time, Link doesn’t know Zelda at the beginning of the game—he knows nothing of the world outside of the forest, in fact. In Skyward Sword, Link and Zelda are childhood friends and have a very close relationship that’s fleshed out in the game’s extended introduction, a presentation we’ve never seen before. But it’s done in a way that slots right into the story template described above. Incidentally, anyone intrigued by the Zelda character in the past will love the strong characterization she gets at the beginning of Skyward Sword.

It’s as though each Zelda game is written by a new person who’s only been told the most basic framework and told to elaborate a whole story built on it. The result is a sort of mythological feel that gives life to the “Legend” in the game’s title. Each time you play a new game, you revisit the legend in a way that’s wholly new, but still powerfully familiar. Near the beginning of Skyward Sword, when Fi tells Link that “oral tradition” is an unreliable way of communicating a story, it feels like a wink from the writers. The story of Zelda is at this point an oral tradition—it’s a story we know and look forward to sitting down to hear once again, knowing it will be new and different every time.

Twitter reply

Let me back up a moment first. I don’t think WoW cosmetic pets are that bad—mostly because their influence on the game is rather limited (they have no impact on any game systems except for acheivements). WoW has been good about keeping their RMT awards almost firmly on that side of the line so far (the “almost” being the ever-awkward issue of paid race changes).

What did I say was “absurd” a few minutes ago? That was specifically regarding people who have a personal goal of maximizing their pet number. Remember in Perc’s/my essay about how games are defined by setting goals? For some people, part of the game of WoW is defined by getting your pet count up as high as you can. Each time you get a new one by doing something in the game, it’s a little victory. But when you pay $10 to see your bar go from 156 to 157? I that point I would ask whether you think your game has gotten a little degenerate, to say the least.

There is a cultural component to it all, and the discussion is just going to crop up more and more with D3 having an RMT auction house. I’m not trying to conflate the cosmetic pets here with that, but, it’s an area where I always think people should be inclined to be very wary.

Why I Didn’t Like Oblivion, and a Question about Skyrim

I haven’t been playing any games much since quitting WoW, and with people talking about Skyrim coming up, the notion of a great single-player RPG has definitely perked my ears a bit. The problem is that Morrowind and Oblivion both were games that I could never get into, for pretty specific reasons. I’m curious if enough info has been revealed about Skyrim’s systems to know whether it would have the same problem or not.

Going to describe Oblivion for now—from memory, tell me if anything is wrong. Even aside from my questions on Skyrim, Oblivion is an interesting study in character advancement systems. It was much more organic and RP-focused than your usual “gain XP, get a level, get bonuses,” and was popular for this reason, but what I want to focus on is how it broke down badly in terms of the gameplay incentives it created.

Characters in Oblivion had stats (Str/Int/Stam etc.), a level (we’ll get to the significance of this in a moment), and skills. Skills encompassed everything you might want to do, from skill with different classes of weapons, to different schools of magic, to utilities like persuasion. The basic concept of the game was that the only way to improve a skill was by practicing it—swing a sword at things a lot, and you’ll see your skill with swords increase by 1, etc. Each skill did have an associated stat that is relied on—you’d be better with swords if you had high Strength. 

I want to pause here to note that that skill system could been a workable game all on its own. No further customization, no advancement other than the list of 21 skills which you could train independently as much you liked by doing things in the world. But they tried to layer on a system of levels and stat growth to give more complexity, which is what causes the problems I’ll outline here.

Character customization consisted of choosing 7 of the 21 skills to be your “major skills.” There were meant to be the focus of your character’s abilities, and choosing an array of 7 did allow for quite a lot of diverse setups. Your major skills started at a higher point, were easier to increase, and most importantly: for every 10 skill ranks you gained across your major skills put together, your character gained a level. What happened when you gain a level? You could choose to increase any 3 of your stats. How many points you got to increase a stat by though, depended on how many skill ranks you’d gained since last level, in skills associated with that stat (major or minor). That sentence is confusing, so a quick example:

Say Blade (swords) and Alchemy are two of my major skills, and Blunt weapons and Athletics are not. Blade and Blunt are Strength-based, Alchemy is Int based, and Athletics is Speed-based. Say I’ve just gained a level, so I have 0 of the 10 major skill ranks needed for my next level. If I immediately hit a bunch of things with swords and make a bunch of potions, gaining 5 ranks in each of Blade and Alchemy, I’ll gain a level. Since my last level, I’ve gained 5 in Strength-based skills, 5 in Int-based skills, and 0 in Speed-based skills. If I choose Str/Int/Spd as my stats to increase for this levelup, I’ll get 3 Str, 3 Int, 1 Spd. But say that before I finished gaining the level, I’d also stopped to hit a bunch of things with Blunt weapons and run around a lot, also gaining 5 in Blunt and Athletics (minor skills remember, so don’t cause me to level up). Now when I level I’ve gained 10 in Str skills, 5 in Int skills, and 5 in Spd skills. If I inrease those three stats, I’d now get 5 points of Str, 3 of Int, and 3 of Spd—a much better result. If I really wanted to increase Agi for some reason, I’d only get 1 of that, since I haven’t trained any Agi skills.

I know this was a lot of description, but here’s the point. It was heavily suboptimal to level an Oblivion character by playing the system naturally—picking the skills you wanted to use most as major skills and then using mostly those. Because every 10 skill gains, you’d be forced into a level up—but you really want to gain a lot more than that each level or else your stat growth would suck. And this mattered a lot because, Oblivion had level-based scaling of enemies. So a level 7 character played naively would have weak stats, and a level 7 character played by someone who carefully planned their skill usage each level (grinding minor skills to max out stat growth for 3 stats each time) would be very different in power level. But the whole game was balanced around throwing the same monsters at those two people, because they had the same level. The first player might actually start having a lot of trouble once he’s gained too many levels “poorly.” And on the flip side, the min/maxy player has an awkward, grindy game where he has to ration use of his “favored” skills to make sure he doesn’t level too rapidly, while spending time training his minor skills to actually make his character stronger.


To sum up:

—gaining a level in Oblivion was actually bad. Every monster got stronger. You only got stronger to the extent you managed your skills/stats properly.

—To maximize stat growth, you needed to gain 30 skill ranks in a level (10 for each of 3 stats, for a +5 bonus). To even get close, +3 or +4 to each, is 15-20 skills. But you’d level as soon as you hit 10 in your major skills.

—So the most important part of your character, oddly enough, was your minor skills. The more you used them, the stronger you’d be as you leveled.

—In fact, the best way to go about things was to make the skills you planned to use most your minor skills, so you could skill them up as much as you liked without forcing level upgrades, and only make things that could be controlled easily (e.g. Alchemy) into major skills.

This post was initially a question on Skyrim, but it turned into a case study of a skill system gone wrong. Hopefully that was interesting though. Getting back to my original thought—this is why I couldn’t play Oblivion. Has enough info been revealed about Skyrim to know whether it will have these problems?


What does it mean when people say they’re “offended” by things? I’m just going to note here that I have no idea, and I basically ascribe no meaning to it at all.

I actually don’t know whether I’m totally clueless in this respect or whether it really is some kind of meme that gets perpetuated despite having no underlying significance. I’m someone who doesn’t always have totally normal emotional reactions to things, and there are sometimes feelings that people commonly talk about that I have no experience with whatsoever. This is one of those cases, and I only know what people will offensive essentially by pattern-matching what I’ve seen people bothered by in the past, not by any legitimate evaluation of content. To the best of my ability though, here is the most practical translation that I keep in mind for what it usually means.

Phrase: “I’m offended by that!”

Meaning: “I want you to stop saying what you’re saying, but I can’t articulate any relevant reason why.”

This seems to get me through most everyday usage. Curious what people think.


For a little added background, it’s likely that my thoughts here are colored by my experience frequently talking about religion as a major topic of interest. One of the things that most sets apart religion from other phenomena is the sheer strength of the social norm immunizing it from criticism. In order to talk about religion in anything but a positive and reverent tone, you have to learn right away to shrug off the “offense” (and its close cousin, “respect”) that are the first lines of defense thrown up against any attempt at discussion.

Shard of Woe nerf

The latest 4.3 patch data contains this toolip change:

Spell Cost Reduction: Reduces the base mana cost of your Holy and Nature spells by 405 and all other spells by 205.

That’s the mana reduction aura on Shard of Woe. I doubt there’s much debate over whether this trinket needed nerfing. But during a recent discussion I did some quick and dirty math on it, so I’ll just copy that here so people have context. 

An EJ poster had said: “This is a discussion to maybe have if you’re still running Sinestra for mage recruits when you’re working on Deathwing. It’s pretty hard to take seriously, right now, the argument that it’s a travesty for an ilvl 379 Sinestra drop to be part of a BiS setup when raiding a zone that drops 378/391 loot.

I replied:

It’s not merely that it’s BIS now, which as you say would be pretty uninteresting (aside: although the 359 Volcano being BIS for some people now is a bit odd). It’s that the numbers on it are commensurate with something that should be many tiers higher.

The most typical healer trinket is always the Int passive + mana proc. At 359, stock example was nonheroic Fall of Mortality (Core of Ripeness, Tyrande’s, and Tsunami were all basically the same though).

In current gear a 359 Fall gives me about 450 MP5 between the Int and Spirit (and 321 Int worth of spellpower).
372 Fall, about 510 MP5.
372 Jar (a mana-only trinket), 600 MP5.
378 Fiery, 540 MP5 if I sync it with Innervate (450 MP5 otherwise).
378 Jaws, about 550 (this one’s trickier to evaluate).
391 Fiery if it exists, about 610 timed with Innervate.

Shard is conservatively about 1200 MP5 for a Druid (along with throughput comparable to Fall). No amount of ordinary item scaling is going to catch up. It’s been a complete anomaly since the day it was first datamined in the beta. The only reason this isn’t distinctly worse than the DST situation is that healers might tend to be a bit softer on min/maxing. Aside from that though, people will basically want to farm Sinestra for the entirety of this expansion, because it’s not just “BIS in T12”; it’s two trinkets in one slot.”

Based on this, cutting the Shard proc in half is roughly in the right ballpark. And that’s what they’ve done.


I’d been looking for a place to put thoughts that were more elaborate than what I can easily do in a few tweets, and people said tumblr was good for this.  I have no particular plan for a blog with a defined topic or regular posting schedule, but I want to have a ready-made place to post things as they come up.

You may have seen me do a little posting at Math of the Lich King.  That was all WoW-math posts and the occasional copy/paste of Druid feedback I sent in to Blizzard. Figured I should take the chance to move someplace more convenient for general topics.

Site name is taken from my favorite book Gödel, Escher, Bach: An Eternal Golden Braid.