#multiplayer game design

[askagamedev]

Why do Fighting Games include invincibility frames for moves that are not blocks/dodges/etc? What is the benefit to having attacks have fractions of a second where the opponent cannot land a hit?

Invincible attacks are a form of counter play - opponents must keep aware when pressing the offense because the defender can use an invincible attack like an Uno Reverse card to stop the offense in its tracks... but only if that invincible attack hits.

Attacks with invincibility are a much bigger commitment than a block or a dodge, because invincible attacks usually have a significant recovery and vulnerable time. If I guess wrong and my invincible attack misses or is blocked, I'm in serious trouble - I'm probably going to get hit by a very damaging punish combo. If I guess right, however, I take the momentum of the match (i.e. "taking my turn"), deal significant damage to my opponent and I either get a knockdown or can start my own combo. Whether my character has access to an invincible attack can affect how my opponent chooses to play against me. Invincible moves are a high risk - high reward choice, which is what makes them interesting in gameplay.

[Join us on Discord] and/or [Support us on Patreon]

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Short questions: Ask a Game Dev on BlueSky

Long questions: Ask a Game Dev on Tumblr

Frequent Questions: The FAQ

[askagamedev]

What is the benefit of mechanics that give the winning side advantages that help them win more, such as COD poweups?

COD's Killstreak rewards are exciting, memorable, and (generally) difficult to obtain. They're similar to, for example, randomly-determined critical hits that deal additional damage or dizzying your opponent in fighting games. Statistically speaking, the killstreak rewards are designed so they won't greatly affect overall game results, but they're really cool to activate and the players engaging with them have a lot of fun. It provides a feeling of power and provides a momentary break from the tension of the normal game experience and allows players to relax for a bit before getting back into the game.

Having rare moments of feeling totally awesome is better for player experience than having a steady overall incremental advantage, even if these two are statistically equivalent in terms of overall match results. Players like these memorable moments - they stand out and engage the player when they happen, and they stay with the player long after. The cooler and more intense the moment, the more memorable it is to the player who's doing it, but it's also not as big a deal for the players getting hit by the killstreak reward. Getting killed ten times by the enemy team is really not that different from getting killed eleven times by the enemy team. Getting killed by a drone rocket or dogs also isn't that different from getting shot in the head. Thanks to skill-based matchmaking and ranking systems, the players get dispersed into their different skill buckets where matches tend to be more balanced and less lopsided, which also makes killstreaks less common.

Overall, these kind of generally hard-to-obtain win-more mechanics exist to make players  who get them feel good and have a memorable game experience, while the players on the receiving end don’t really feel that much worse. They aren’t really random so they can be both planned for and against. They’re a nice bit of polish that can differentiate a particular game from its competitors.

[Join us on Discord] and/or [Support us on Patreon]

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Long questions: Ask a Game Dev on Tumblr

Frequent Questions: The FAQ

[askagamedev]

As a game developer, what are your thoughts on "pay to win" mechanics (I.E. being able to buy more "power" in a game, particularly multiplayer ones, with money), both in business terms and in "fun" terms. Based on your "flow" graphs, it seems it would be a bad idea overall as the players facing pay to win players would fall into frustration area while those buying power would rise into the boredom area.

These sort of paying/nonpaying issues are typically ironed out by the multiplayer game's matchmaking system. Players often (mistakenly) think that matchmaking only considers player skill when distributing players into games. If player power (such as gear, level, paid or earned boosts, etc.) is a factor, I guarantee you that the matchmaking algorithm will take it into account because we don't actually want lopsided games - if we pit teams of equal skill against each other but one team has much higher player power level than the other, then the game will be a curbstomp. We actively want to avoid this - curbstomp games are both frustrating for those getting stomped and boring for those who are doing the stomping.

The actual goal of the matchmaking system is generally to even out player power and player skill on both teams. Our rough formula often looks something like this:

[Player Skill] + [Earned Player Power] + [Paid Player Power] = [Total Player Power]

Equally geared players who have paid boosts but are lacking in player skill, for example, would probably get crushed by a significantly more skilled or geared player without those boosts. A mediocre skilled but well-geared player can ride lots of paid boosts higher, but will still hit that ceiling of highly skilled players, some of whom might occasionally pay for stuff they like. As long as our metrics can measure player skill well enough (and we generally get it right after a few games to calibrate), the matchmaking algorithm usually does a pretty good job of setting up reasonable multiplayer games. This does mean that the upper ceiling on total player power does include paid player power after skill and earnable player power have hit their own upper limits, but this is also only at the very highest skill levels in general.

This is where the idea of "pay to win" tends to break down - there comes a point where players can't actually guarantee a win by paying more, they can only increase their player power. At best, their Total Player Power will increase due to the paid boosts, but they’ll get eventually sorted by the matchmaking system and reach equilibrium for their new Total Player Power level. In the vast majority of games where players can buy power, there's still an upper limit on the power a player can buy or earn. When they hit that limit, then the only recourse is raising their skill level. This is, for example, why nobody can win competitive Magic: The Gathering tournaments by buying the most expensive cards. Even if a player owns every card, they still need the skill to construct a good deck from that card pool and make the right in-game decisions to pilot that deck to victory. While some exceptions exist, this does tend to be the general rule when it comes to selling power. Since one major goal of matchmaking is to reduce the frequency of lopsided games in favor of closer games, our matchmaking and ranking systems are designed to take paid power into account. Our games might not always hit this goal, but it isn’t usually due to malice so much as flaws in our matchmaking and ranking systems.

[Join us on Discord] and/or [Support us on Patreon]

The FANTa Project is being rebooted. [What is the FANTa project?]

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Long questions: Ask a Game Dev on Tumblr

Frequent Questions: The FAQ

[askagamedev]

Why in some cases in multiplayer games queue in matchmaking takes so long?

Consider what it is matchmaking does - it spins up a new game and directs chosen players from a player pool into that new game. The matchmaker has to consider various factors to determine who among the players in the pool are the best choice for putting into the game. Each factor that the matchmaker uses to determine whether the players are eligible for the game reduces the size of the player pool to choose from. If the player pools don't have enough players in them, the matchmaker might not be able to find enough players to start a game and thus be forced to wait for new players to join the pool. The more factors the matchmaker has to optimize for, the longer the process usually takes.

The fastest way of matchmaking is just throwing every single player into the same pool, and then firing up a new game whenever there are enough players to start. The only thing the players have to wait on is there being enough players in the queue to start the game. Let's say that a typical game requires four players to spin up - a two on two game. This means that the highest number of players waiting in the queue will be three - every fourth player that enters the queue will fire up a new game. However, this completely ignores gameplay factors like map choice, team composition, player skill and experience, character power level, and so on. This can often result in lopsided games and curbstomp multiplayer sessions... but it will be fast.

Let's say that we now give players their choice to queue up between two maps - Alpha Map and Omega Map. Now we've split our pool in two - players who queue for Alpha and players who queue for Omega. Now the maximum number of waiting players is six - we can have three waiting on Alpha and three waiting on Omega. The next player that joins will have to choose either Alpha or Omega and the appropriate game will spin up. We have, however, doubled the maximum number of players waiting. If no one else queues for Omega, those three waiting Omega players will be waiting forever (or at least until they give up).

We can add the option to queue for either, in which case we can pull from three pools (and hopefully the Either pool is popular enough), but this also starts its own set of potential issues - what if Alpha is just way more popular than Omega? In this case, players who want to play Omega specifically will be in for longer wait times as well, since there will always be players looking for Alpha games, which will not only pull from the Alpha pool, but also the Either pool. This still keeps our maximum number of players waiting at six (three Alpha-only, three Omega-only) in the worst case, but usually fires games more quickly on average because both Alpha and Omega games can pull from the Either bucket. 

This kind of player pool distribution is then extrapolated for every factor we want the matchmaker to consider - player skill, character power level, ranking, map preference, multiplayer game type preference, etc. - and then all of these are balanced against queue time. Most matchmakers will start removing consideration factors as wait time increases because our general thought is "a player would rather play a game that isn't ideal than wait a long time for one that is". We try to ensure that the removed factors will still result in a play experience as close as we can to the desired one, but that isn't always the case. "Assemble four players to play" is generally a lot easier and quicker than "Assemble a tank, a healer, one ranged DPS, one melee DPS, and a support that all want to go to Wombat Dungeon II: Wombat Forever for time attack mode that are all diamond rank".

[Join us on Discord] and/or [Support us on Patreon]

The FANTa Project is being rebooted. [What is the FANTa project?]

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Long questions: Ask a Game Dev on Tumblr

Frequent Questions: The FAQ

[askagamedev]

How often do mmo games with an auction house system have dev-side bots autosell things, or was it just an example? Is it common to doctor the market in such a direct manner outside of bugs like item duping?

That particular feature isn't super widespread to my knowledge. When I saw it, the system only generated common high-volume auction house items like crafting materials and consumables if there weren't enough of those items up for sale already. So, for example, if there are fewer than 10 Wombat Leather for sale on the auction house for at least 8 hours, the system kicks in and generates enough new Wombat Leather sales attached to randomly generated names to make up the difference. Only commonly traded items (that is, items from a designer-defined list, no rare stuff) would be generated this way.

The issue with creating a system like this is that it spends resources treating a symptom and not the root cause of the main problem - that there aren't many players engaging in the gameplay to keep the auction house stocked. Most of the time this is due to a low overall server population, and that tends to bring a slew of other issues with it. We can spend our dev resources building systems like this, but a better overall solution to fix this (and all of the other problems at the same time) is usually server merges. That would allow the auction house to remain full while the devs who would have built this system get to work on something else instead.

[Join us on Discord] and/or [Support us on Patreon]

The FANTa Project is being rebooted. [What is the FANTa project?]

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Long questions: Ask a Game Dev on Tumblr

Frequent Questions: The FAQ

[askagamedev]

Does simply being active on an online game generate revenue for the developer? Am I contributing to Blizzard’s bottom line by playing Overwatch?

Yes you are. You might not be contributing directly, but you are adding monetary value to their game by playing it. Just by playing the game, you contribute in these ways:

#1: You might spend in the future

People who don’t play won’t spend, period. The more people play long term (weeks, months, years), the more likely they are to eventually spend some money. There is no greater correlation metric between money spent than with time spent playing. We call this time spent playing “Engagement”. By continuing to play, you’re making it more likely that you will relent down the line and spend something. It’s in our best interest to keep you around and playing to eventually reach that point. You might not get there yourself, but the longer players play, the more likely they will eventually give us a few bucks.

#2: You help provide entertainment for other players

Players don’t like playing games (especially multiplayer games) when they’re ghost towns. We need a critical mass of players to pull from in order to keep matches spinning up and running. For things like ranked play, the pool of qualified players is even smaller. Furthermore, many of those players who are playing with you have spent money at some point in the past. If they have a good time, they’ll probably spend some too. By playing with other players, selling or buying stuff on the auction house, interacting with other players, you provide other players reasons to continue playing. 

#3: You help provide useful data about player behavior

We’re always watching what you do. We keep track of what maps you like, what game modes you participate in, what characters or classes you like to play, where you spend most of your time in game, and so on and so forth. While one player’s behavior data isn’t super useful, having millions of players’ aggregated behavior data is really useful. We can use it to figure out elements of the game that aren’t working. That might mean figuring out whether Bastion’s turret form is really OP, whether Genji really does need a nerf, where the pain points are in a specific stage, or even additional things that we can sell to players that they want. 

#4: You are easier to reach with cross-promotional offers than non-players

Overwatch is fun and all, but what if Blizzard wanted to drum up interest in their BlizzCon Virtual Ticket? Well, they can show advertisements for it in the game and it reaches even non-paying players like you that play regularly. Some number of these players might decide that the ticket price is worth it and buy it. Or maybe Blizzard wants to get more players for Diablo Immortal, so they offer a special Overwatch skin or something to players if they download and register Diablo Immortal. Some number of those players will continue on to keep playing Diablo Immortal and possibly spending there too. You’re already invested here, whereas somebody else who doesn’t play at all like EvilDeathCrab would not be. It’s much easier to convince you to engage with something than it is to convince him.

This is, by no means, an exhaustive list. Overall, non-paying players like you provide a good amount of value to the player community and developers by playing. That value might not be direct monetary contributions, but it’s still valuable nonetheless because it indirectly contributes to monetary contributions. We want as many players as we can to fill the servers and engage with the content. Some of them will pay, some of them will not, but it’s an ecosystem where we try to provide enough value and fun to all segments of the player base.

[Join us on Discord]

The FANTa Project is currently on hiatus while I am crunching at work too busy.

[What is the FANTa project?] [Git the FANTa Project]

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Long questions: Ask a Game Dev on Tumblr

Frequent Questions: The FAQ

[askagamedev]

If multiplayer is where the big bucks are, why are so many beloved games focused mainly or exclusively on one-player? I imagine it's because those franchises (Sonic, Mario, Zelda, Mega Man) date to the 80s and 90s, and the market was a different beast back then, but what do you think are the most important differences that led to the change? Is it because the advent of online subscriptions enables monetization that wasn't possible before?

Back when those beloved franchises were first created and popularized, the only way to play multiplayer was when all players were physically within a few feet of each other. There was no other way to play with somebody else. Because the majority of video gaming back then was done alone, the best games tended to focus on that single-player experience. It should come as no surprise that the shift towards beloved multiplayer franchises (like Call of Duty, Halo, Team Fortress, Starcraft, Warcraft, Counter-Strike, Dota, etc.) came hand in hand with the spread and popularization of internet connectivity.

It’s important to remember that online gaming didn’t really start taking off for PC until the early 00s, when internet speeds better than dialup started becoming more popular. Consoles didn’t really follow suit until the X360 and PS3 era starting around 2005. Technology that we now think of as ubiquitous didn’t exist back then. As a point of reference, keep in mind that the very first iPhone didn’t launch until a year later in 2006. 

It really wasn’t the monetization and subscriptions that made it work. It was the fact that we could transfer data over the internet itself to players (and from players to each other) that made the shift multiplayer happen. Before that, the only way to update a game at all was to put it into a central location (i.e. a store or BBS) and hope that people would go there to get it. Playing games with friends was an event - you’d have to gather in a central location (like a video arcade) or at someone’s home. Gathering at someone’s house was also finicky, because most computers and consoles logistically only supported up to two players at once. It just so happens that playing games with other people is actually really fun, and people really like doing it - especially if you can have a lot of them. As technology that allowed us to communicate instantly over large distances became more popular and widespread, games evolved to take advantage of the new connectivity by providing players stuff they didn’t know they were missing before.

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Long questions: Ask a Game Dev on Tumblr

Frequent questions: The FAQ

[askagamedev]

How does a developer decide when they stop supporting a game in terms of actively releasing any more new content updates or patches? Is there a checklist you can go through and say, "We think this game is finished" and any further work on it has increasingly diminishing returns? And is that point much sooner for a single player game versus a game with multiplayer content?

Oh, that’s easy. In the AAA world, we developers finish working on a game when the publisher says so, because the publisher is the one paying our bills. Usually this is planned out reasonably far in advance so that the team can scope and assign tasks to create as much quality content as we can in the time we have allotted. On the business side, the developer and publisher collaborate to figure out how much overall engagement there is with the game, and come up with revenue estimates for the continued support (paid DLC, microtransactions, GotY editions, bundles, merchandising, subscriptions, etc.). Eventually the projected return on investment will hit that tipping point, and that’s when they pull the plug.

As for when that happens… it’s always a case by case basis. It really depends on the game, the business model for continued support, how many are playing, how much they’re playing, who’s playing, etc. It isn’t necessarily different for single player or multi player, per se. You can look to the continued success and support of great single player games like Fallout and Skyrim as great examples of single player games with long lifespans.

It’s just that multiplayer content tends to be much more efficient at converting developer resources into player engagement than single player content. Gamers are practically insatiable when it comes to consuming content - hardcore players can easily consume months of developer work in mere hours. However, in multiplayer, the act of playing with others becomes new content for the players - you can play the same map, or the same mode, and have wildly different (and still engaging) experiences because your team members or your opponent is different. Meanwhile, single player content tends to be very similar on subsequent playthroughs.

Remember, the goal isn’t to champion either multi player or single player content. The goal is to keep as many people playing the game as long as possible because people who keep playing are far and away the most statistically likely to spend additional money, and there’s no universally optimal way for all games to reach that point. If we can get great levels of engagement via efficiently created single player content, we’ll do it. If we can get great levels of engagement via efficiently created multiplayer content, we’ll do that. Maybe we’ll get to do both. Maybe there isn’t enough initial player engagement and we do neither. Above all, however, it always boils down to efficient allocation of our limited developer resources.