I’ve been really getting into a game called League of Legends lately. It’s the latest incarnation of a relatively new genre of games popularly called “MOBA” or Multiplayer Online Battle Arena. The high-level synopsis is that it’s a real time strategy game (think Warcraft 3) on Ritalin.
Gameplay
Rather than building up slowly from the start, the way you do in a typical RTS, you play a Summoner and start off each game with just one hero character which you choose out of dozens of available ones. You can choose a different hero each time if you like from ones you have unlocked or a set of 10 freely available ones that rotates each week. Each hero has a unique set of abilities and statistics, and there are a few different roles heroes can play. Some are “tanks” who run into battle and start bashing everything in sight, some are “supporters” who hang back a bit and beef up themselves and other allied heroes in the area, some are “carries” who have special abilities that help a team push down towers and bases at the end, etc.
Games are typically played 5 on 5. Both teams have a base on opposite corners of a square map guarded with lots of beefy defensive turrets (the purple and blue objects on the map below) and there are 3 paths between them, each of which has more turrets. Each team spawns minions (the green and red dots) periodically at the same time that move down each of the 3 lanes attacking whatever enemies they find and, if left unmolested, meeting in the middle.
For anyone who has played Warcraft 3 this may sound familiar. It’s based on the popular custom scenario called Defense of the Ancients (or DotA). It’s built by a company called Riot Games, which as far as I can tell has raised somewhere north of $15m and hired 40 people, including some who built DotA, to put it out.
Monetization
Possibly what fascinates me most is how they monetize: exactly the same way my Facebook games do, by letting people play for free and selling things in-game. They primary sale item seems to be the heroes. There are dozens of them, and you could unlock them yourself by playing a lot and gaining experience, but it would take you forever. Even unlocking one of the top tier heroes could take you weeks of casual play. Instead you can buy them at prices ranging from about $2 to about $8.
We’ve thought a lot about doing some sort of more traditional RTS game on Facebook. The problem is that Facebook games monetize so well because they are persistent. When you buy an android in Starfleet Commander, he provides you with a permanent resource boost. That’s why we went with the perpetual RTS model that we use now, where it’s like a very slow version of Starcraft that lasts forever.
Normal RTS games, however, last generally around a half hour to an hour. Thus it would feel like a waste to a customer to buy something that will only help your team out for that short a period of time. The tough design decision then is finding a way to give people permanent incentives (and therefore advantages) to purchase items while keeping the game somewhat competitive for those who do not.
League of Legends solves this to a large extent with the Summoner/Hero paradigm. Summoners gain persistent advantages over time, both as a reward for playing and from purchasing Riot Points for cash. Heroes start off every battle anew, though they gain some of those persistent advantages.
Summoners level up on a persistent basis. Playing games nets you experience that increases your level. With each level you get a mastery point and a rune slot. The mastery point lets you improve one of the two Summoner spells you choose for your hero at the beginning of every match. The rune slots allow you to purchase runes (using another type of points earned by playing) that increase your hero’s stats as well.
Heroes also level up in-game, but start each game again at level 1. So if a level 20 Summoner playing a hero called Ashe faces a level 1 Summoner also using Ashe, their heroes both start off with the same base stats. The only differences are that the level 20 guy will have more and better runes and better masteries, giving him a pretty serious advantage.
Matches
Because of the level advantage provided by runes and masteries, a strong matchmaking algorithm is necessary. I feel this is the biggest flaw in the game so far. The matchmaking is very uneven.
From what I can gather they use an ELO-style rating system (similar to what chess uses) that is unseen by players to determine (probably to a fairly accurate degree) each team’s chances of winning. They try to match up players such that they will win half of their matches and lose half.
The problem with this system is that a game is fun when teams are evenly matched and the 50/50 split is a result of that. The problem is that there are other ways to get a 50/50 split than evenly matched teams.
Suppose, for instance, you took a group of 5 Level 20 Summoners. One game you matched them up against 5 Level 40 Summoners, giving them effectively no chance to win. The next game you matched them up against 5 Level 1 Summoners, giving them effectively no chance to lose. Your group of 5 guys just won one game and lost one, but neither were fun at all.
While the net effect of their ranking system isn’t quite that drastic, it’s not far off either when you’re new to the game. Once you have some experience, though it seems to improve a decent amount.
I found out, through our Starfleet Forums, that one of the game’s designers happens to play our game and had a chance to talk to him about it. It sounds like they know it’s a problem and are working on improving it. I’m sure they will. It’s a tough problem to get people in even matches, especially when time is a factor (nobody wants to wait more than a couple minutes for a game to start) and one that I imagine they will be able to improve upon greatly both with effort and as their user base grows.
Interestingly, matchup quality will give them somewhat of a network effect when competing against other games in the genre, much the way game selection did for online poker. The more players you have, the more you will be able to get, say, a group of 10 level 1s all playing together in 2 minutes.
Either way I’m excited to see where they go with this. Riot Games is really breaking ground with their business model, and doing it while making a big-budget, high quality downloadable game. They’ll be one of the startups on my watch list for the next couple years.
