When we at Riot Games set about building PC online multiplayer-centric title League of Legends, we did so with the lofty goal of creating a high-quality competitive experience with a rich, evolving metagame and low barrier to entry. We believe these features are essential for a free-to-play game aimed at core gamers to be successful.

Multiplayer Online Battle Arena (MOBA) games like League of Legends have their roots in the modding community. And while players are more than willing to accept a single map from a modder, they tend to be more demanding when it comes to virtual real estate in a full-fledged product. As a result, before League of Legends officially released, we set about working on our next battlefield.

It is our belief that in order to maximize the success of a multiplayer map, the new maps would need to be available to everyone. Therefore, they're not something that can be explicitly sold to players in our virtual store. They do, however, still play an extremely important role in a game that is operated as a live service.

First, and most importantly, they provide new, fun, competitive options for players by providing an alternative to the core experience.

A new battleground can also revitalize the interest of players who had previously moved on, motivate players to purchase new items better suited to this new battlefield, draw in new players with the promise of unique gameplay, and just generally increase the fervor of competition.

So, while content that isn't offered in the store doesn't directly drive sales, new maps are certainly important and can play a role in the game's growth.

A champion doing battle on Summoner's Rift, the first map released for League of Legends.

Releasing a new map, however, can be a more daunting task than it sounds; particularly if you want it to be balanced and ready for competitive play. From the time a new battlefield is just a twinkle in a designer's eye, to the time players are poised in a pitched, online battle with one another, there's a lot of work that needs to get done. So what goes into making a balanced and innovative MOBA map? Well, to start with, let's talk about what makes a MOBA game tick.

What is a MOBA?

MOBA games harken all the way back to the days of the original StarCraft, when the custom map Aeon of Strife was created. This map streamlined the traditional real time strategy paradigm by placing the opposing bases under the control of the AI -- each base spawned minions to attack the other down predictable, preset paths called "lanes".

Allied with each of these two AI-controlled bases were two opposing teams of players, with each player in command of a single, powerful hero unit (or "champion"). Without the intercession of the players, the two bases would fight to a standstill, neither making any headway towards the other.

Aeon of Strife, however, was subject to the limitations of the StarCraft map editor, and as a result, didn't possess enough intricacy to hold the lasting attention of a hardcore playerbase.

As a result, the MOBA genre didn't truly begin to blossom until the birth of Defense of the Ancients, a community mod built using the more powerful Warcraft III world editor, which allowed for the creation of much more robust champion units. Players could now earn money and experience for their champions by killing enemies, gaining levels and acquiring equipment to help them wage war. It was here that the standard paradigms that would define the genre would begin to take shape.

A game of Aeon of Strife, the forefather of the MOBA genre, in progress.

What defines the traditional MOBA map?

The map layout for Defense of the Ancients was a bit more advanced than that of Aeon of Strife, and quickly became iconic of the MOBA genre. This map, which was mirrored diagonally from right to left, contained three lanes dubbed "top", "mid", and "bottom" (or "bot") relative to their location. Along this central axis, a river connected the three lanes, serving as geographical delineator of the midpoint of the map, and facilitating combat between lanes by allowing a centralized access point.

Since each team now consisted of five champions, one player on each team would have to fight solo. While this made him more vulnerable, players in duo lanes had to split experience, allowing the solo to potentially grow in power faster and become more threatening later in the game.

Additionally, between each lane there were neutral areas that became known as the "jungle," which provided ambush routes for players to sneak up on enemy heroes, as well as areas where players could go to fight neutral monsters for additional gold and experience.

To complicate the fighting, powerful defensive turrets were placed along the length of each lane at even intervals, their threatening presence providing shelter to an allied champion under duress.

At the base of each lane were barracks structures that we later renamed "inhibitors". Destroying an enemy's inhibitor caused each of your minion waves in that lane to gain the assistance of an extremely powerful super minion to help further the attack.

Finally, at the heart of each base sat the enemy's command structure, or "nexus", protected by two extremely powerful turrets. Destroying the enemy's nexus meant final victory for your team.

Making drastic changes to an established genre, however, is a tricky business, and League of Legends was already shaking things up with some significant changes to the traditional MOBA gameplay.

As a result, we decided that our first map, Summoner's Rift, would be heavily based on the one used in Defense of the Ancients to allow veteran players of the genre to (at least initially) remain in their comfort zone. For our second map, however, we specifically wanted to challenge these establish trends and do something a little bit more edgy. We called this new map the Twisted Treeline.

In DotA, evenly placed turrets defend each lane, complicating attacks on the enemy base.

What were the goals of Twisted Treeline?

When our design team set about creating the Twisted Treeline, they did so with certain gameplay objectives. For starters, we wanted a smaller team size. Since Summoner's Rift called for a team of five players, this could complicate scheduling for a casual player who wanted to play with a full team.

With this in mind, we designed Twisted Treeline around teams of three players. Additionally, games on Summoner's Rift could run rather long, averaging around 35 to 45 minutes, which could be a prohibitively long time for a session based game. We wanted something that came in a 15 to 20 minute package.

In addition to these practical concerns, games on a map like Summoner's Rift were characterized by a long "laning" phase during the early game, in which players would focus primarily on leveling up, and movement outside of someone's assigned lane is minimal. We wanted to truncate the laning phase to bring additional emphasis to roaming the map and ambushing (or "ganking") lone opponents, as well as creating opportunities for exciting, early game team fights.

So how did we change the map to help reach these gameplay goals? For starters, we wanted to preserve the strategic depth of having a single solo player, so one of the lanes had to go. We also wanted additional ambush points and larger neutral areas, so the shape of the map had to change.

What we ended up with was a map that was mirrored horizontally with two asymmetrical lanes. The first ran along the bottom half of the map; a straight shot from one base to the other. The second, ran along the middle, but took a more meandering route. Each lane had a large jungle area above it, leaving one situated between the two lanes, and the second running along the uppermost part of the map.

Moreover, control of each jungle area provided a different advantage. While control of the central jungle provided the already significant tactical advantage of ambush routes against his opponents in both lanes, the upper jungle was populated by neutral monsters that provided beneficial enchantments (or "buffs") to the champion who dispatched them.

These buffs could be used to gain the advantage over a team that was playing extremely defensively, or to retake the initiative when losing to an overly aggressive opponent. With these goals in mind, let's see take a look at how things changed over the course of Twisted Treeline's development.

Guinsoo acquires a powerful buff from a monster in the upper jungle.

Early October: Prerelease

The base layout on Twisted Treeline has seen a lot of upheaval since its initial construction, even prior to its release to open beta.

In its original construction, both of Twisted Treeline's lanes followed the same layout as Summoner's Rift: three turrets per lane and two turrets protecting the nexus.

Almost immediately, however, internal playtests revealed this to be a logistical problem.

Not only did the lanes feel extremely cramped -- with your enemy's tower sitting just over a screen away from your own -- but game length was coming in at around half an hour: nearly the average length of a game of Summoner's Rift.

So the decision came down: bulldoze the two outer turrets and increase the distance between the now outer turrets and the base turrets.

This not only helped to reduce game length, but also allowed for a larger no man's land in which players were vulnerable to ambush through the jungle, increasing the importance of control of these neutral areas of the map.

Reducing the number of turrets on Twisted Treeline provided a more manageable game length and increased the strategic significance of the middle jungle.

Late October: Open Beta

Around late October, just prior to the release of League of Legends, we decided that Twisted Treeline was ready for open beta. Almost as soon as we pushed it out the door, high durability (or "tank") champions emerged as extremely dominant. With fewer champions per team, damage heavy champions simply didn't bring enough firepower to the field to be able to quickly and effectively focus down a three tank team.

To combat this, we implemented changes to many of the neutral buffs provided by the monsters in the upper jungle. Most pointedly, we changed the buff provided by the Rabid Wolf, a powerful monster that was still manageably dispatched by single player, and the team buff provided by Ebonmaw, the large Dragon that required the attention of the entire team.

The Rabid Wolf now provided a bonus to attack speed and the rate at which abilities became available, and Ebonmaw now provided a flat 1 percent increase in damage per champion level instead of his previous bonus damage against turrets. These buffs allowed damage powerhouses to even the odds against tank heavy teams by controlling the upper half of the map.

The neutral buffs provided by the monsters in the upper jungle help even the odds for this team of fragile, high damage characters.

Also, even with the removal of a turret from each lane, bases were still absorbing a bit too much damage. With Twisted Treeline being substantially smaller than Summoner's Rift, and with fewer champions in play, taking down an inhibitor before the enemy team could rally a defense was extremely difficult. To combat this, inhibitor health was reduced by 20 percent, and respawn time received an across-the-board increase of two seconds.

Additionally, with two turrets guarding the nexus, the base was still able to readily dispense with incoming minion waves on its own with a single inhibitor down. While this was a desirable mechanic on Summoner's Rift, where there are three lanes, and therefore three inhibitors to lose, on Twisted Treeline we really wanted players to feel the impact of losing one of these important base structures.

To increase the pressure caused by the loss of a single inhibitor, we removed a nexus turret and increased the health of super minions by an additional 200 HP. This forced players to deal with the afflicted lane personally, allowing for a three champion team to more easily take map control, and ultimately make the final push.

With only one turret defending the nexus, this team is forced to personally deal with the enemy super minions in order to avoid heavy damage to their base.

Mid November: Gold Statistics Rebalance

As time went on, it became clear that the action on Twisted Treeline wasn't building at the pace we wanted it to.

Games were still coming in well beyond the average target length, and, with fewer minions and towers on the field, player income wasn't scaling properly to accelerate the game phases.

As a result, players had to rely heavily on champion kills to supplement their income.

In addition to favoring certain play styles, this also presented a problem for champion balance, as champions that were extremely effective at killing minions had trouble keeping up with those that typically rack up more frequent player kills.

To bring these champions back into line, the gold per minion kill was increased by a small amount and the gold that players automatically gained over time was reduced slightly.

Additionally, the frequency of minion waves increased, but the number of minions in each was reduced by one. This helped reemphasize the importance of lane control, and helped rein in the long roaming phase that was developing in the middle of the game.

Increased gold per minion kill allows this champion to keep up with other champions that might have scored more player kills.

December: Neutral Spawn Time

Even after all the changes that had been made, matches on Twisted Treeline all still seemed to have one thing in common: an early game fight at the Lizard Elder monster situated in the middle of the map.

This monster was designed to provide the team who killed it with vision of the middle jungle, an important intelligence advantage when watching for an ambush. As a result control of the Lizard Elder, particularly early game, became extremely important, and players always seemed to begin the game by fighting over this scouting advantage.

While this was an interesting phenomenon, a sound victory at Lizard had the potential to become the deciding factor in the game. In order to prevent the game from being so drastically influenced by a level 1 fight, as well as to enable players to feel that they had strategic options early game, design decided to delay the spawn time of the Lizard Elder.

The Lizard's spawn was moved to 2:30, giving each team some time to level up before the race for fight for control of the middle jungle began, making for a more developed and intricate fight for this important resource.

Two level 1 teams start the match off by fighting over the Lizard Elder located at the center of the map.

The Current Picture

The Twisted Treeline originally envisioned needed a lot of work to bring it in line with our initial design objectives. In order to balance it to be on equal footing across the various champion types, to allow for multiple opening strategies, and to keep the game length under control, many different aspects of the map required extensive tweaking from the original design. Ultimately, the Twisted Treeline of today is a very different place than the one that debuted on October 21, 2009.

And there's a lot more to the map that we haven't discussed here, such as the implications on our automated matchmaking system, our dual currency system and our metagame advancement, but you'll have to play to see them for yourself.