Tobold's Blog
Friday, August 11, 2017
Is multiplayer an error?

I've never played Friday the 13th: The Game, but according to its Wikipedia page the game has 1.8 million players in spite of having only mediocre reviews. The only reason I know about the game at all is because this week's news about it: In this asymmetrical multiplayer game, one evil Jason against a group of camp counselors, the devs had to implement a patch to remove friendly fire, because the camp counselors kept killing each other deliberately instead of actually playing the game against the Jason player. Sigh!

Online multiplayer games are now about 2 decades old. Developers like them, because they have positive network effects, that is players serve as content which attracts other players; and you don't have to program a complicated AI if you just let players fight each other. But nobody managed to really solve the problem of gamers being such assholes in an anonymous multiplayer environment. Pretty much every multiplayer game has a long list of complaints, where players are unhappy about the behavior of other players. The removal of features which lets players communicate and interact with each other is more common than adding such features, because each such feature brings its own set of problems. The most successful multiplayer games are those in which players have the least opportunity to freely interact with each other, where there is no chat, and no way to hinder another player's progress.

So I have to ask whether this trend towards multiplayer games wasn't a mistake. Games which have positive network effects also have negative network effects, so that player number frequently fall precipitously a few months after release. When players leave it gets harder to find the necessary people to start a match, and nobody likes waiting for that. Some people like a game, but are then driven away by the nasty behavior of other players. And the multiplayer aspect frequently requires the game company to run game servers, so these games frequently die completely when the servers are shut down, as opposed to single-player games you can find on still decades later.

Whether players are annoyed about others deliberately griefing them, or just unhappy because their team members aren't competent enough, I am not certain that the positive effect of a game being multiplayer compensates for all the negative effects. I know already quite a number of games which I don't play because they are multiplayer, or would like more if they were single-player. The dream of being able to play with people from all over the world quickly turns into the nightmare of finding out that people from all over the world aren't actually very nice. And even game companies would probably recover the added cost of programming an AI over time due to having significantly less customer service costs for a single-player game.

Multiplayer has been tried, and failed. Maybe it is time to try something else, like developing better AI.

Multiplayer is one of the most important features of a game for me. If I can't play it with friends, it has 0 replayability, and I play it once, or twice. My top 15 most played has 2 single player games in it.
In the top ten games from Steam : (source : )
805,462 820,461 Dota 2
485,150 520,209 Counter-Strike: Global Offensive
108,840 126,675 H1Z1: King of the Kill
64,399 69,014 Grand Theft Auto V
62,102 75,434 Path of Exile
60,027 63,223 Team Fortress 2
55,837 55,885 Warframe
53,244 58,429 Football Manager 2017
41,452 45,063 Tom Clancy's Rainbow Six Siege

9 of them are Multi, and 7 are only multiplayer. So the saying that "Multiplayer has failed" is a bit of an overstretch.

Developing a good AI would be interesting, but the issue is I (and some others) find more interesting to outwit a fellow human, than some AI that as been designed by a developer to be just winnable by myself.
"Maybe it is time to try something else, like developing better AI."

We'll get right on that. I mean, how hard can it be, right?
I play mostly single-player games, and I want good strategy game AI for its own sake.

Multi-player is what it is. I just imagine I'm back in the schoolyard at the age of ten or so.
11 of my top 15 games are single player (actually a few of the strategy games offer multiplayer, but I've never tried it), but I still have to agree with Bhagpuss. Thousands of developers are constantly trying to develop better AI (and have been for decades), but I can't think of a single game with AI that I was really impressed with. The best you hope for is that it won't be game-breakingly bad.

I certainly hope they continue to improve AI, but I'm not holding my breath. Until then, it takes multiplayer to offer decent and varying strategy in games that you can't just figure out and exploit.
I disagree about devs having put much effort into AI. It often appears to have been an afterthought. Most of a dev team is coding graphics, and we have dedicated graphics cards to play games. Where are our AI cards?

And it is not an impossible task. The AIs made to play chess or go are now well at grandmaster level, better than most humans.
Chess (in particular, Go has only recently had a big improvement) have very simple unchanging fields of play - and have been played in the same form for centuries, with lots of people trying to program them.

That said, it should be easy enough to make a decent AI for a chess-like game (it would not be up to chess AI standards, but the players won't be up to chess human standards either.)

Card games are often not so bad these days. The luck involved means that a reasonable AI will sometimes get the right cards and provide a challenge. But when it steps outside the borders of cards only (e.g. Hearthstone or the excellent Elder Scrolls Legends I've been playing lately) the AIs often get terrible again (Duelyst may be an exception - I didn't really like it but the AI seemed solid from what I saw, despite having moving creatures. By contrast, Combat Monsters is fun to me, but the AI is terrible.)

Civ-like games are hard, though the AI could be better than it is - I would settle for HOMM3 level at this stage. I think part of it is that they design the game and then write an AI for it. AIs would be better if they made rules that an AI can work better with. Separate strategic and tactical levels like HOMM makes a great start.
Anyway, maybe things are looking up in some cases:

This AI bot just beat some of the best Dota 2 players in the world:
@Gerry Quinn

My understanding from reading about that Dota AI bot is that it is just extremely good at executing the very simple (but mechanically difficult) tasks in a 1v1. For instance, in denying last hits the bot can time its auto-attack to land exactly 0.01 seconds after a minion attack brings it low enough. In the strategy part of the game, bots continue to be terrible.

This is important because of what we're talking about, a varying AI that is fun to play against. Playing against this AI would just be extremely frustrating.
@Gerry Quinn

While the DOTA 2 bot is more advanced, these types of Bots have been around since the 97-98 time frame and were very popular in the first-person-shooters of the day as opponents to practice against. The major problem with the bots of this time, was that players were eventually able to adapt these bots to intercept and alter the net code, allowing the bots to join a game and appear to be a human player. Basically, these players were invincible, and it took a long time for developers to implement changes in the net code by which to identify these bots. If the code for the DOTA 2 bot ever gets out into the wild, in the right hands it could spell doom for the game based on what I saw in that video.
It's really the single player game that's the aberration. Consider that most sports, card and board games require human vs human competition and historically it's the single player game that's the oddity.

AI provides an alternative but the popularity of multiplayer exists because humans crave interaction with other humans even it's merely the knowledge that they bested another human.

We are social creatures and our default is to play with other humans. Now that said, when playing a game, we also like the rules of the game to be followed. If I'm playing Monopoly, I'm going to be pissed if the banker is stealing money from the bank. That's really no different than someone exploiting 'friendly fire' to break the rules and purpose of the game. In my mind, the mistake here lies with the game design and not that it's multiplayer.

As I said, the desire to NOT play with others is the aberration here. It may be that crappy game design and toxic players lends that aberration a hand but it's still not the default mode for us as a species.
What is an aberration is being forced to play with random people. You wouldn't invite anal spamming idiots, racial slur users and 1G please punks to your home for a tabletop RPG game.

But devs want these morons and slackers to enjoy the game and force us to play with them. THIS must end, not multiplayer.
We've always played single-player games, ever since the first cavemen skipped a stone on a lake. We play games with animals. We enjoy puzzles, whether they come from puzzle books or we thought them up ourselves. Card players invented a thousand varieties of solitaire.

Now we have magic boxes that can play some games that previously we needed other people to play, so of course we use them for that as well. It's not remotely an abberation.

Sure, we may argue that sometimes we substitute machines for people more than is ideal, but the opposite may be argued too.
I've found scientific study:
Post a Comment

Links to this post:

Create a Link

<< Home
Newer›  ‹Older

  Powered by Blogger   Free Page Rank Tool