Development Diary: Why we hate being called an MMORPG

Monday, November 29, 2010

Path of Exile is not an MMORPG, it’s an online action RPG. Yes, it’s played online on persistent servers with thousands of other players, but we don’t want to call it an MMORPG because the term has become so closely associated with World of Warcraft clones - games that allow hundreds of players to play together in each area. Path of Exile limits the number of players in any one world area (a few dozen in towns, much less in actual wilderness areas) by instancing them. It also features visceral action combat that would not be easily to simulate on servers designed to handle many hundreds of players in one area. Although this is stated many times on our website and in interviews, many people still believe we’re making a traditional MMORPG.

There’s only room for a few MMORPGs

For the last six years, there has been one dominant MMORPG in the western market. Many studios have attempted to muscle into the same space, sometimes spending upwards of $30 million in the process. Almost all of these games launched with a bang but failed to keep most of their users after a month or two. Many of them are struggling to remain profitable day to day, or to repay their development costs. The reason for this is clear – their launch content is unable to match the level of development and polish that World of Warcraft has had by now and players are unwilling to pay for more than one subscription at once.

There are major technical and business challenges that must be solved when developing an MMORPG that we literally can not afford to deal with. As a small indie studio, we need to deliver a compelling online RPG experience for a tiny fraction of the budget of a typical MMORPG.

Where possible, we try to make design decisions that mitigate these costs and risks while making the most action-packed game possible. This article contains some generalizations about how we see the world, the challenges of making Path of Exile and our responses.

Massively Multiplayer, but you can’t play with everyone?

One of the main selling points of an MMORPG (and something that seemed impossibly cool back in 1999) was that many hundreds of players could gather in the same area and all see each other’s actions in real time. Despite how technically difficult this is to implement, most online games want the “Massively Multiplayer” title so much that they invest in the necessary infrastructure to make it happen.

MMORPGs are very sensitive to population levels. If the game has too few players, it begins to feel like a ghost town. The MMO experience relies on the player being able to encounter other players as they explore the world (both to quest with, and for PVP shenanigans). If player levels stay low for a while, server merges are required - confusing the players and damaging the game economy.

If population levels are too high, the world has to be sharded into many concurrent versions of the same world, housed on different servers. Some MMORPGs have many such shards, which means that two arbitrary players who first meet outside the game are unlikely to both play on the same server. This is a major problem for many players - they can run around with hundreds of other people, but only the people who happened to pick their server. They have limited options to play with people from other servers, such as World of Warcraft battlegroups.

Because Path of Exile uses instancing for all areas (including towns, at a higher population cap than the wilderness), there is no technical restriction on the number of players that can be playing concurrently on a server cluster. If it starts to get overloaded, more servers can be easily added in parallel. We’re planning to only shard our servers for major geographical locations (America and Europe for example). This means that regardless of the number of players in America, they can play together if they choose to.

Another advantage of using instancing is that if the server population is relatively low (on Christmas day, for example), there are still the regular number of players per instance - just fewer instances running concurrently. Players don’t feel the ghost town effect so strongly when their instances are still full.

It’s quite telling that recent MMORPGs put a significant amount of their end-game content in instances. They’re trying to reap the benefits of this system but are still forced to shard so many copies of their world because of the large non-instanced portions of their games.

Action Combat

Unfortunately, when designing traditional MMORPG servers to handle this many players moving in the same area, many compromises have to be made. There’s no time for the server to do anything other than receive the movement input, check that it’s not inside a wall, and send it out to all of the hundreds of players watching. Checking for dynamic collisions with monsters or other players is often too computationally expensive with this number of connections - players can walk right through their enemies. Because an MMORPG cannot easily simulate tactical combat, they have to design systems that emphasise strategic decisions through skill selection.

Path of Exile doesn’t allow hundreds of players in an area. We don’t even want to. For this type of game, there are major gameplay benefits with restricting the non-town instances to a maximum of 4-8 players. Our instance servers are coded more like the servers for a First Person Shooter. They care a lot about fast movements and positioning of a low number of players at once. This allows us to have tactical combat mechanics such as monster-body-blocking, stun-lock, and dodging projectiles as they fly towards you. You can't dodge a fireball in World of Warcraft. The hit success is calculated as soon as it's cast, regardless of where the target is once it gets there.

Due to the low player limit, the world areas can be fun and challenging for both small party or solo play. There’s no need to wait for dead monsters to repopulate or queue for bosses.

Handmade vs Random

Traditional MMORPGs consist of giant handmade worlds, painstakingly crafted by talented teams of artists and designers at great expense. Once players are grinding end-game content, developers are pressured into adding new areas every few months so that their player-base doesn't get bored of repeating the same static areas.

Instancing world areas also gives us one more opportunity - we can randomly generate their layouts. This goes a long way towards facilitating replayability, because players are very sensitive to repeating the same area over and over. It also means that we can hide interesting events and subareas in the world.

Like some other Action RPGs, Path of Exile has multiple difficulty levels. Once the player has finished the game with their character they can then continue to play the game again in a mode where the monsters have become far more dangerous. Our random level generation allows us to also scale up the environments in these areas - mazes can be longer and caves can be deeper.

Random Loot

Most traditional MMORPGs have quite a large degree of item determinism. Monsters have loot tables and a small random chance to drop specific items. Whenever a new end-game area is added, multiple tiers of weapons and armour must also be designed and created, often at significant expense.

On the other hand, Action RPGs generally approach item drops in a completely different manner. All monsters have a small chance to drop all items of their level or less. The items have some number of magic properties that are randomly picked out of a pool appropriate for that item. This system is basically a slot machine. There are frequent chances to win based on either the item type or on the synergy and power of its properties. It eases the end-game content requirements substantially because playing normal areas can yield valuable item rewards also.

This non-deterministic item system is cheaper to develop. Rather than having to design hundreds of new items for each content patch, only a few new base types and new random properties effectively create millions of new combinations for players to find.

Business Risks

MMORPGs are not cheap to develop. There are plenty of examples of MMORPGs that have cost upwards of $30m USD, even before marketing costs.

Another significant issue with these projects is the risk. Game projects generally have quite high variance (only a small portion are profitable), but it’s worse with an MMORPG. Because most of the revenue comes from subscriptions (or microtransactions) after release, there are many opportunities for the game to become unpopular and fail. There are also substantial post-release development expenses that are difficult to fund if the game was already over-budget and performing poorly. Once a company doesn't have the funds to release promised content updates on schedule, users leave, and funds dry up further. This death spiral is the reason many MMORPGs ultimately fail.

The revenue models of modern MMORPGs are all over the place. Many of them require a boxed copy purchase, a monthly subscription and also have premium microtransaction content available for purchase. These schemes can be quite confusing (and expensive) for the customer.

ArenaNet’s Guild Wars had great success with just a boxed copy purchase and no monthly fees. Nexon’s MapleStory made a small fortune by supplying the game for free and just selling microtransaction perks. It’s worth noting that neither of these games are traditional Everquest-clone MMORPGs (Guild Wars uses instancing and MapleStory is a 2d side-scrolling online game).

Many players are unable to justify multiple $15/month subscriptions at once, and already have a World of Warcraft subscription that they are not giving up.

Microtransactions are great, but most companies quickly degenerate into selling powerful items that become mandatory for serious players. Most players in a free game aren’t willing to spend money on microtransactions, but are very important to keep around because they provide playmates for the paying players. The sale of mandatory powerful effects (currency, experience potions, weapons) quickly alienates the non-paying players and can cause a major backlash leading to player exodus.

With Path of Exile, we’ve targeted a specific niche group of gamers who enjoy online action RPGs with dark fantasy art styles. We are members of this gaming subset, and by specifically catering to it we will provide a great gaming experience on an affordable budget. Advertising costs are also substantially cheaper because we know exactly who we’re targeting. These factors mitigate a great deal of the risk in our project. We felt that it was much better to have a 75% shot at making a great online action RPG rather than a 5% shot at making a great MMORPG (for 20 times the budget).

In a similar vein, we’re giving Path of Exile away for free. We want as many people to try it as possible so that we can grow a large community of players who enjoy the game - it’s less risky that way. This is also why we’re offering only ethical microtransactions. We fear the risk of player backlash and it’s just not worth selling experience potions to try to make a quick buck. Sure, we’ll make less money per paying customer than our competitors, but we know the community will appreciate it". We intend to maintain Path of Exile for at least ten years, so it’s important that we have a good relationship with our players.

It’s common for developers to outsource portions of their development to various developing countries to take advantage of cheap labour. Communication problems over distances are common with this type of arrangement and can lead to wasted resources or assets of questionable quality. We’re developing Path of Exile in New Zealand, which means that, other than a couple of overseas contractors, most of our developers are English-speaking Westerners who are immersed in Western RPGs. It’s also good that New Zealand wages are among the lowest in first-world countries. Path of Exile’s low budget means that we don’t need to recover $30m via microtransactions like some competing free games do.

The Launch

After making such a gargantuan investment, developers of traditional MMORPGs want a massive take off with expensive and rigorous advertising campaigns, but, we don't blame them - it must be pretty scary trying to recoup 30m via the long tail (the period after launch where sales are tailing off).

Online games usually have lengthy lifespans and generally receive frequent content expansions (both free and paid for). Because of this, their content can be pseudo at launch.

A newly released online RPG might appeal to it's most hardcore of fans - although; it will generally lack the polish and breadth of content that keeps mainstream players around for more than the first few months. There's a group of roughly 200,000 MMO fans who drift around from new release to new release; playing out their free month on the initial content before getting bored and moving on.

One luxury that indie developers of non-massive online RPGs, such as us at Path of Exile have, is that we don't need to try and recover the development costs in the first few months of release. The reason that popular indie online RPGs seem to come out of nowhere, is that they were initially released to a limited niche audience and eventually gathered more & more mainstream appeal as they received updates in the first year or so.

We lack the advertising budget to have a giant mainstream launch for Path of Exile, unfortunately, so we are initially focusing on a specific core of online action RPG players. Once we've added some expansions and grown our company, we'll hopefully be able to put forward efforts into advertising, thusly luring in more die-hard RPG players.

It also helps that any additional content we add, can take advantage of the random level/item generation system that has been incorporated - doubling the amount of raw content will respectively double the amount of time players can be entertained and enjoy their gaming experience.

In summary, by choosing not to make Path of Exile a traditional MMORPG, we’ve effectively reduced both the technical and business risks for our small team. Instead, we focus solely on delivering a great action RPG. So far these decisions have paid off significantly - we are developing a product that we’re very pleased with and for a budget that we can afford.

If you’re keen to try Path of Exile when it enters beta next year, please sign up for a forum account now.

Community content is available under CC-BY-SA unless otherwise noted.