How can we create fun multiplayer games without the dependencies and annoyances of multiplayer games?
Before starting a new project, video game companies need to make a very hard choice: whether a new game will be multiplayer-based or not.
If multiplayer dependant, then the game needs to be designed in a certain way, and massive amounts of marketing and great timing need to be part of the equation. If multiplayer is not going to be a requirement, then the game cannot have the same design as a multiplayer game, as a single human player should be able to play and have fun at any given time, not dependent on other humans being awake and willing to join in the game.
The great advantage to designing, marketing and successfully releasing a multiplayer game is that, if popular, there are a massive amount of potentially paying humans that are engaging with your product. An extreme example of this is Fortnite, which is exclusively multiplayer (and which has made over a billion dollars in revenue so far). On the opposite side of the spectrum is Zelda: Breath of the Wild, which is a single player ‘Puzzle Adventure’.
There are great limitations to both types, as Breath of the Wild lacks depth in the characters and opponents one encounters, whilst Fortnite becomes abrasively difficult and off-putting for new players (as current players welcome one with too much enthusiasm in the form of immediate death upon arrival).
This conundrum is not new. As soon as a game opens up to the multiplayer dimension, the playing experience can be tainted with either the meanness and cruelty of masked anonymous humans, or simply the natural combination of random people with different levels of experience resulting in mismatching expertise, speed, knowledge and capabilities. Neither of these are fun at all. And most times, a heavily mismatched group will naturally lead to the manifestation of meanness.
So how can we create fun multiplayer games without the dependencies and annoyances of multiplayer games?
At Decisive AI we set ourselves out to answer this question and solve the paradox. The answer we came up with is to use smart, fair and good NPCs, or intelligent artificial players (IAPs), as we like to call them. But designing these is way too complex for humans to do, that is why NPCs are such poor opponents/allies, and often cheating entities too.
So we figured that the best way to create IAPs that play well and fairly was by using artificial intelligence. Specifically, we train neural networks in different ways using mainly reinforcement learning techniques. It is clear that a machine can learn how to play a game well, without having any advantage over how a human plays, and that this knowledge could be embedded in the design of a IAP which in turn forms part of a multiplayer game.
The availability of a IAP is not limited by the need to sleep or work - it is always willing to play. The manners of a IAP are impeccable, never rude or discriminating or sexist or bully-ish. The IAP can match the level of a human player by adapting with different versions of itself without getting offended, thus making the gaming experience fun and customized. The IAP does not have more resources or visibility than a human player, so the game is fair. But most importantly the IAP can play a game very well, far from hand-made half-baked patterns of hard-coded behaviour, fully satisfying the multiplayer experience without ever needing to go online.
And all of this is not only true for existing games, but even more so for games that will be conceived with this solution in mind. The obstacles that multiplayer games pose to both video game producers and consumers alike do not need to limit the kind of game that is produced or played anymore.
At Decisive AI we are making this solution possible with existing technologies from different areas of expertise, all the way from machine learning to software architecture and design to video games. And by partnering with video game companies, we can’t wait to enable a fun and healthy multiplayer experience for all.