100 Followers Special! A Time I Designed a Game and the Lessons That Taught Me About WoW

In this post, we’re going to have some fun at the expense of poor old me.

Since I started writing this post, I’ve hit the 100 followers mark, so this will also serve as a celebration of sorts! I’m so happy to have such a thoughtful, engaged audience who bring a lot of different takes to the table and remain engaged even as our engagement levels with World of Warcraft fluctuate. So thank you all!

This post is going to largely be closer to a more conversational tone. The idea here is to share why I don’t always buy fully into my own ideas or suggestions for WoW – while I think I have something to offer, I don’t pretend to know all the nuance of game design, and this post will illustrate that all too well :).

So here we go!

Back when I first started playing WoW in 2005, the systems behind the game were something of a mystery, being cracked slowly by dedicated players. Sites that collected and catalogued the data being discussed slowly grew into the collective consciousness of the community, like Elitist Jerks, Thottbot and Allakhazam, worldofwar.net, and this eventually gave way for the database sites to consolidate to WoWhead, news and datamining via MMO-Champion, casual discussion via WoW Insider (then WoW on Joystiq, and now Blizzard Watch), and theorycrafting via Elitist Jerks, until today, where we basically have class discords doing the theorycrafting, Wowhead and MMO-Champion both basically doing the same things, Blizzard Watch covering the game a little bit less, and Elitist Jerks is gone and some of its pivotal members are now a part of the game team (most notably Ion Hazzikostas, who is still the EJ guild raid leader, and Hamlet, who is now on the class design team).

In 2005, the scene, while large in some ways, was more interesting, as there were a lot of pieces of the game we fundamentally did not know about until addons and analysis began to expose them.

Into this environment, a young and fresh-faced 19-year old me, finally committing to playing after 7 months on the fence post-launch, enters the scene. Well, sort of. I was mainly lurking and reading all the various posts. I had finally settled on WoW after a brief flirtation with the original Guild Wars, and the idea stirred in my head – “what if you could get the GW1 server setup (one “server” with instances of pretty much every zone to load balance) but with a WoW-styled gameplay and progression?”

Thus, an idea was born – Fantasy Planet Online.

A friend and I talked about it at length, to the point of building an email account for it and putting up a listing on some community game dev site that I don’t remember the name of, even getting a sound person to do a single 43 second sting for the game!

Like many things this friend and I discussed, it went to the wayside, as playing WoW was easier than trying to envision making something more ambitious than it.

But later in 2005, some Molten Core under my belt and spare time at work available to read forums and theorize about how things worked in WoW, I came across the first discussion I ever saw of item level.

Back then, the idea of item level was hidden by the game UI, although some addons could expose the data. It was simply a component of the database records for the game, but slowly, high-end players began to try and figure out how it worked. The concept began to be reverse-engineered, with theories of how both item level and item quality affected the scaling. It was relatively simple math compared to today, because the values were so low, but there were enough variables to confuse. The community had found a relatively stable formula that seemed to work, zeroing in on a base formula that spit out an item value and then used that value to spend to buy stats. At this point in time, gear was more interesting, so since not every piece followed our current formula of (base stat + stamina + up to two secondaries), it required some careful review to try and peg a “cost” to certain stats. It was also a confusing time, as the number of additional stats on gear was huge. It was an era of spell damage vs spell and healing power, Spirit regen vs. MP/5, Paladin sets with 3 base stats plus stamina and spirit on them, and a variety of other delightful oddities – probably the main reason that gearing in Vanilla was so flexible.

The idea lit the fire in me for this game again, and at the time, I was a supervisor at work, so no one really questioned me making spreadsheets – plus, I worked a late shift, so I was usually the only authority available, which made my pursuit of mathematical formulas easier to do.

From that, I began to concoct “new” gameplay ideas – which were really just thinly-veiled rip-offs of the analysis I was reading about WoW. In the end, I created an 8-page design document, which really only touched on a few things: class design, race ideas, and itemization.

I’m going to walk through these, and then by the end, I’ll tie it into the types of armchair game design those of us that write about WoW and other games often do, and how they might even inform how the current WoW team operate!

Sounds good? Here we go!

Class Design – I Thought Up The FFXIV Job System but on Crack, 8 Years Before A Realm Reborn


My big takeaway from those early months playing WoW is that class kind of didn’t matter as much as a role (which, to some extent, was accurate in Vanilla, as you often got shoehorned to a role based on class!). So my big, leading idea was that you’d start the game with a role instead, mostly – you picked melee, magic, or ranged, with melee having tank abilities, magic having heals, and ranged being the only pure DPS. As you leveled, you’d eventually hit decision points that would present you with a branching path – that first path offering the choice of role more clearly, allowing you to define as a tank, healer, or damage dealer, and then closer to the end of your leveling journey, you’d branch again, choosing further specialization according to your preference. It was a weird design, but ironically, it kind of almost fits the job system in FFXIV, although Jobs just expand the base class with more stuff, where my idea was you could pick 3 different types of stuff to add on the first time, and then another choice from 3 new and different things. It mainly served to address things I wished WoW had at the time, while also appealing to the dual-classing you could do in GW1.

For example, my two big hangups were, “What if I wanted to be a holy avenger priest and smite foes?” and, “What if I wanted to be an archer without a pet?” which are two things that WoW in modern times has kind of fixed (Discipline being close to the first idea but still healing and Marksmanship getting Lone Wolf). I had a slew of other ideas that came from discussions with my friend who had originally thrown in when we talked about the idea – stuff like melee classes that could use magic in conjunction with their melee attacks, Necromancers (we were all big Diablo II marks), and shamans, but with like, stone armor instead of mail or plate.

One of my favorite ideas that I still wish could be a thing to this day was the Arcanist (it was a later idea so it wasn’t in these documents, sadly). The core design concept I had was that it would be rad as hell if you could have a caster who had to really wind up spells to land them, but could deal catastrophic damage for sticking it out. The big cornerstone spell was a 10-second long spell cast (?!) that would hit for a ton of damage if only you could ever succeed at casting it. While I want to emphasize that this core concept is kind of bad on the surface, if you asked me in 2019 to make it, it would be built around channeling and DoTs, using certain abilities to drop the cast time on your next cast of this spell and stacking until it had a “normal” cast time, rather than sitting with your thumb up your ass for 10 seconds, or getting 60%+ haste. I always liked the idea of strong magic spells and the gameplay implications of positioning correctly and knowing the mechanics of a fight enough to be able to hardcast a 10 second spell, which should tell you a fair bit about what I find fun(read as: I’m a bit weird). Ultimately, I never made a game, but if I were making an MMO right now, I would sure as hell come back to that idea!

Basically, the idea kind of wound up being WoW specs, but sort of different, with 27 ending classes which matched the number of WoW specs in Vanilla.

Race Design: Or How I Really Really Liked the Chapter Black Arc in Yu Yu Hakusho and Tried to Make It


The approach to racial design was ripped right out of classic WoW, more or less. The idea of base stats by race wasn’t even new to WoW, but was rather an adaptation of the way such things often worked in pen and paper RPGs, much less MMOs. The concept of racial abilities also came from these same roots, and the end result is a design that feels very samey – you could have min-maxed your race/class choice to maximize stats or racial abilities, and the act of designing them to maintain balance was pretty tricky – and without an actual game to play, impossible to directly validate. Looking at it 14 years later, I can definitely say it looks pretty imbalanced, and it was definitely going to pigeonhole some roles into certain races (there’s no way a serious tank in this model wouldn’t be a Dwarf or Mazoku)!

However, I would still be pretty okay with this design, all told (with appropriate balance). While I imagine the racial abilities would be slightly imbalanced in perpetuity (as is the case in WoW), I think there’s still something fun here. The aesthetic of each of the races offers something, although I think distinguishing them would have taken some more effort. For example, one of the races listed, Mazoku, was literally me ripping off the story arc of an anime, Yu Yu Hakusho.

(Spoilers for it coming, in case a 1992 anime that was highly popular is on your radar but you haven’t watched it yet)

Highly abridged, Yusuke Urameshi is the main character, a punk teen getting into fights, who dies early by rescuing a small child from an uncoming car, is resurrected by completing trials and forced to serve as a Spirit Detective, investigating demons, which leads to standard action anime fights (it is really cool and one of my favorite shows so I am totally underselling it for brevity) but eventually, Yusuke meets a foe named Sensui, who was formerly the spirit detective, who kills him again. All seems lost until it turns out that Yusuke is the descendent of a Demon King, which makes him part-human and part-demon, which the show terms Mazoku.

Now, while none of those dramatics factored in, I thought it would be cool to have such a race, which basically amounted to different stat mix, racials, and additional customization options for a human akin to what a Demon Hunter offers the two elf races they can be – body tats, gnarlier customizations, and nothing more. Some of the other options offered more interesting ideas, but basically, I thought that a full gamut of fantasy races with a few new twists (usually borrowed from elsewhere in pop-culture) would do the trick. Maybe it would have – who knows! Of course, at the time, min-maxing hadn’t yet spread to the point where we were all that bothered by racials, although by the end of Vanilla, we definitely knew. At the time I was drawing up all of this though, it was a little more ambiguous and open to interpretation, and so the design I outlined drew from a more innocent time, before entire guilds race changed for a 1% edge in the world-first race.

Itemization: What If Item Level Inflated Even Worse?


So back to the story I told higher up in this post, where the best approximation of itemization math we had in WoW was that there was likely a value assigned based on level, a multiplier for item quality, and then a percentage of that base score based on a slot valuation. My take away for my “original” concept was to extrapolate a few additional points from this:

-What if we added a modifier by equippable level, meaning that gear would inflate exponentially as character level and item level both climbed in tandem, and as quality further climbed?

-What if we had around 17 different stats you could spend on?

-What if we fundamentally misunderstood the valuation of trinket effects and itemized them completely incorrectly so that they were usually garbage? (okay, this wasn’t a paradigm of mine, but I accidentally fell into it!)

-What was a good hook for the 4-armor types WoW offered? I know, a fifth! What should be it be? Stone! Why? Because nature classes!

Otherwise, the idea fit pretty well into what the community had published as its knowledge of WoW itemization. Stats were assigned a value cost, applied against the value pool assigned by the formula to determine the item budget. My urge to use round numbers in many values led to items either being under value (only had 6 points left to spend, can’t buy any stats, oh well) or over value (it only takes 2 more points to get that extra strength on this gear, better do it!). This led to me designing a bunch of items for funsies where there was no consistency in the valuation between items of identical slot, armor type, item level, quality, and character level modifier, which is, as I understand it now, IS NOT A GOOD THING FOR DESIGN.

Imagine the whiplash of having two upgrades drop from a boss with relatively equal stat valuation, but one just randomly manages to have stats out of line with the valuation. While that isn’t necessarily a huge issue, it still could lead to various imbalances, especially if one class can equip more of these overbudget pieces than another, or if certain classes end up stuck with their best upgrade being under budget. It was a bit chaotic and weird, and such design would have only gotten worse as leveling progressed.

The other thing I realize looking back on this is how fucking weird early WoW itemization really was. My priest was decked out in a mix of Healing Power, Spell Damage and Healing, and gained regen from both Spirit and Mana per 5 seconds, although Mana per 5 seconds was usually at the time regarded more highly, since Spirit was subject to strict rules around when regen could occur.

It’s a digression, but in case you didn’t theorycraft at the time, here is how mana worked in Classic – you could gain mana by equipping more Intellect, which is fine, and you could gain more regen by equipping Spirit. Regen was a formula that gave a mana per 5 second value, so if you had 200 regen at the time, that meant in a 5 second window, you’d gain 200 mana from passive regen. However, there was also a casting regen value, which cut regen to 0% of your base regen (unless talented), and that reduction lasted for 5 seconds after casting.

A lot of healing in Vanilla was built upon rotations of active healers, utilizing downtime between spells to prolong the exhaustion of mana. For example, on Ragnaros, my guild used 3 priests (me included!) using Greater Heal rank 2 on the main tank one after the other, using a macro with the spell and a chat announcement to our healer channel telling the next person to cast in 3 seconds. The way this mathed out, we’d be able to recover a good portion of the mana spent on the downranked Greater Heal by the next time we had to cast, meaning our mana was going out in a slow trickle, and the tank had a constant stream of healing.

This made MP/5 valuable, because it had a constant value – 12 MP/5 would always regen 12 mana every 5 seconds, regardless of casting status.

Upon that foundation of understanding, I built my concept, which ended up basically being “WoW, but weird and worse.”

Other Miscellany – Experience Tables


I made an experience requirement table for the game. The scaling was percentage based, so it scaled up more and more, reaching a pretty high point. It would have probably been the most usable design element, because you can always fit a curve of values into the rewards for quests and kills, although if I were designing an MMO in 2019, it wouldn’t have a level (which might be a topic for another time). And that concludes the documents I dug up for this post!

So Why Share All Of This – The Nature of Being a Game Fan

Ultimately, I think the thing I most want to communicate with this is why I would share in the first place – which is simple enough to answer. I think, in times like these, we often get critical of the creators of the things that we love, and may not fully understand the reasoning behind their decision-making.

Which is not to suggest that they’re doing well or that any lack of enjoyment is squarely our fault, mind you – rather, just to say that the job of designing all of this is often harder than it appears!

But I also noticed in this trip down memory lane that there was a difference in how I thought about game design as I got older. When I was a kid, growing up with games, I often designed them as a fun spare time activity, keeping notebooks around to jot down ideas. When I think about the time I spent doing that as a child, I often focused on gameplay – how things should feel, what would be fun, etc. I noticed in the documents I just shared here that there is something different about my approach in early adulthood, compared to the ideas I had as a kid.

When I was a kid, it was all about gameplay – what was fun, what would be exciting to play. As an adult, many of the designs I’ve toyed with are systems-focused. Look at those documents again – there’s no lore, no world, no characters, no story, nothing of the sort – it is instead stat sheets, an experience chart, and itemization formulas. Sure, if I am being charitable to myself at 19, it was the case that I was indeed just grafting things onto stuff I was already familiar with. For all intents and purposes, the “game” I was designing could just take place in WoW! What I was doing wasn’t so much a fully original design, but more of a “what-if” – what if I could get what I thought was fun class design and gameplay grafted into WoW?

But the larger point this led me to is something that, I think, speaks to a deeper observation – much as how my own design scribbles moved towards systems-first design and away from core gameplay and fun concepts, I think WoW has observed a similar shift.

Look at the design of the original game compared to now – a large delta exists there, but let’s go with it for a moment. The launch of WoW didn’t offer a bunch of content where the sole purpose was to consume all of it. Instead, the game didn’t really have much in the way of systems – it had crafting as a basic UI-driven activity, questing that was more robust than many MMOs on the market at the time, leveling and progression systems with talents and spell ranks, and some basic frameworks for how you’d travel around the world.

Compare to now – there is the Heart of Azeroth, a system that demands immediate attention when you reach Battle for Azeroth content, Azerite armor, which flashes at you and demands attention, 3 leveling zones which are intended to be fully explored within the initial leveling experience, and a set of endgame systems all designed to be rapidly consumed, in some cases, using existing content as the wrapper by which it is delivered. The game has dropped a lot of the old wrapper it used to conceal this in, and has become a highly-visible set of interlocking systems.

To be clear, this alone is not inherently a bad thing. At many of the better times in games, I think we can see the systems driving us forward – but there is usually still some surviving facade, the sense of world that binds everything together cohesively. There is something of a problem when those systems drop the facade to reveal the inner-workings so readily to players.

Early WoW is full of spreadsheeting players trying to figure out how things worked, poking and prodding the gears of the machine to figure out the responses. However, for all of the harvesting of data that occurred back then, the game itself was still largely built to maintain the artifice, and I think in that, a part of the magic.

Current WoW is largely built by a new generation of designers and developers that came up through that era, and in many cases, were the spreadsheeting pioneers themselves. Their modality of thought tends towards systems-oriented design, and it reflects in the end product we all play. That, again, isn’t bad, and when done well and built with a strong surrounding game world and lore, it does extremely well. Legion was built in much the same way – interlocking systems designed to all push a common progression element, and it did not receive nearly the aggression from the playerbase. I think that a large portion of that is due to the ways in which the systems, on-display as they were, managed to also integrate with the lore and world. The class hall missions made more sense in-world than the Garrison or faction missions. The artifacts had way more lore surrounding them than the Heart of Azeroth. These things can feel insignificant, even to a player like myself who largely focuses on gameplay, but they matter.

The current team is very systems-driven, which I think can be leveraged as a strength, done properly. I think the larger issue is that I struggle to identify the ways in which the larger lore of the game and the world around us is being merged into these systems. Even in 8.2, I might really like Azerite as a system, but what is the lore? Why is it that me scooping up Azerite and leeching its power is any better than the randoms that mine it in the myriad world quests where my goal is to stop them? If I’m not ultimately putting the Heart back into Azeroth, it seems kind of like I’m just stealing the power myself.

The artifact at least sort of made sense in that way – it was specifically empowering me, by draining the power of these little trinkets and baubles, and at the end of Legion, we burned out our artifacts by draining the demonic energy from Sargeras’ blade. The gameplay of that system was both more transparent (clear Artifact Knowledge ranks, single-screen UI with clear progression) and more well-defined.

I guess, at the end of this long ramble, what I really want is the systems-focused types to be paired more tightly with the writers and the lore, able to work on how these systems present in the game and how to ensure that presentation maximizes the fun of the game.

Because no matter how good your idea might actually be, if I can see the spreadsheets behind it easily – it might not be great.

Worse, it could be my (non-existent) game!

2 thoughts on “100 Followers Special! A Time I Designed a Game and the Lessons That Taught Me About WoW

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.