On Retention Rates Of Competitive Games

This post was difficult to write because when I tried to use various examples from many different games, the essay quickly lost its focus and became bogged down with explanatory context in case you weren’t familiar with them. It takes a great deal of writing space to satisfactorily explain even a single game so this quickly became unwieldy. So, I’ve since decided that I’m just going to mostly stick to a few, well-known examples of popular games and hope you’re somewhat familiar with them.

These games are (mostly) StarCraft II, League of Legends (or any MOBA really), and Overwatch. These are all competitive, multiplayer games that include matchmaking, a ranking system, and are/were somewhat popular.

I’m going to define success as sustained or increased player population growth over time. Of course there’s a variety of reasons for why each game may not achieve success. For example, maybe the game gets old quick, has low production quality, or there’s game breaking bugs that the developers never fix. There’s a multitude of reasons why any game lives or dies. In this post I’m going to try to ignore any obvious, low-hanging fruits that have already been discussed ad-nauseam elsewhere and instead bring focus to more obscure, although arguably just as critical, variables affecting player retention. The list will not be exhaustive and will mostly be psychological aspects that some developers ignore, much to their peril.

I’m interested in why certain games just seem to slowly die off despite having no glaring issues and are otherwise well-crafted and fun games. In this post I will present certain psychological hypotheses that I believe hinder the retention of many competitive multiplayer games.

Let’s begin with StarCraft II. The game is extremely stressful to play. In fact if you google “ladder anxiety” the first result is a guide for how to deal with the anxiety associated with just queuing up for a match in the video game, StarCraft II. This was a real phenomenon that used to be discussed a ton on the forums. Eventually it got so bad that StarCraft II mostly became a spectator sport, where everyone would rather watch pros play on Twitch rather than play the game themselves.

But why was it so bad? Certainly there doesn’t seem to be nearly as much ladder anxiety for other contemporary competitive games. What gives? Here are my main guesses:

  • StarCraft II was 1v1. You had no team to fall back on. It was just you against your opponent. (Other results for ladder anxiety nowadays mostly involve Hearthstone, despite its casualness, slow game speed, and randomness (RNG) factor. Ranked 1v1 games seem to come with ladder anxiety as a package deal.) Also, insert obligatory evopsych remark regarding humans being awful at fighting any animal 1v1 throughout history and evolving to thrive in and prefer group engagements for survival.
  • There was no RNG, other than maybe spawn positions. If you lost, it was entirely your fault. You couldn’t blame it on bad luck.
  • The game was mentally taxing. It was real-time, not turn-based. The only limit was just how fast you could play. Professionals often had hundreds of APM (actions per minute). From the second the game started to the very end you would have to play your very best at all times with no breaks.

I’m honestly surprised the game lasted as long as it did. If not for the thriving spectator culture, the fact that it was created by and backed by Blizzard, and the existence of team-based and custom games, I’m sure it would have met its demise much sooner.

Needless to say, most people like to (or at least they tell themselves) play games to relieve, not induce, stress. When a sizable portion of your player base has to expend considerable amounts of mental effort and willpower just to queue for a match, in the long run there will be fewer and fewer matches and players each day.

Contrast this with other popular competitive, multiplayer games out right now. MOBAs (multiplayer online battle arenas) seem to be a pretty safe formula. Ever since their inception via DotA, there have been numerous spin-offs that are more-or-less all successful: League of Legends, DotA 2, Heroes of Newerth, SMITE, Heroes of the Storm, etc. It’s a hell of a formula considering that most games fail but these games in this category are all quite similar and experience great success and continued growth. Basically, it seems pretty hard to fuck up making a MOBA.

So what sorts of properties do these MOBAs have? Well for one you have four other teammates. This provides a diffusion of responsibility and makes individual contribution difficult to measure. Depending on your skill level you can play more or less important roles and you can always deflect blame to anyone else on your team for the loss. One needs not feel as culpable for the loss and it’s difficult to ascertain your actual skill level. Ignorance certainly is bliss when it comes to personal and teammate skill level in competitive matches. When your teammate performance is transparent, those doing well will be quicker to flame lousy teammates, which leads to increased tension and less desire to play the game.

Most successful games nowadays are team-based, with the sweet spot seeming to be around 4-6 players per team. Any less and there’s too much pressure on you to perform, and any more and it feels like the entire outcome of the game is due to luck since you cannot realistically carry that many players. In other words, your contribution is too diluted. Free for all (FFA) is okay, but it often leads to turtling, unlucky situations (getting ganged up on), and has a lack of camaraderie. On the other hand, there’s not as much pressure compared to a 1v1 since players can fight each other and you can be more selective in your engagements. People also don’t seem to like FFA as much, as evidenced by their lack of market share in popular games. Even Quake Champions is currently trying to promote their team-based mode over the class Quake FFA, most likely due to the above factors.

More importantly, whereas in a game like StarCraft II you may soon feel accurately placed in the matchmaker and give up ranked, in a team-based game you are never truly certain of your own skill level. As a result you will continue to re-queue because it’s plausible that you’re just repeatedly getting poor teammates and actually belong at a higher rank. Maintaining the illusion that you’re not accurately placed is what keeps many people requeuing.

Continuing with our contrastment, MOBAs also require much less mental effort. The first ten minutes or so is usually the laning phase where the players perform a mini-game to get last hits on creeps to collect a bounty. This is reasonably low effort and can be auto-piloted quite effectively. After that you usually group with your team for the rest of the game and repeatedly engage in team fights. These team fights require all your mental effort but they’re short-lived activities that are interspersed by simpler processes like roaming, farming, walking to lane, waiting to respawn, etc.

My point is that the games follow a relatively simple flow chart and require little expenditure of mental energy. The graph of effort would be mostly baseline with the occasional spikes during ganks and team fights. This helps avoid fatigue associated with utilization of maximum mental effort. Games that consist of short-lived matches that are heavily concentrated in effort can still be fun, but it’s rare for players to play them for extended durations.

Consider also that people desire fair matches. If people cannot play your game for a lasting time frame, the matchmaker has less of a player pool to select from and has to compromise by having less equal matchups. Or it can have longer queue times. Both of these result in player dissatisfaction and can make them exit the session. This can also become a negative feedback loop where even more players quit in discontent, worsening the situation even more. Before long the population reaches a critical low and the game dies.

Another important factor is a high skill floor. Take a game like Chess-you can learn the entirety of the rules of the game in under an hour, yet the game’s skill cap is so high that people can spend their entire lives improving and learn something new from every match. This is good because it doesn’t lower the skill ceiling, just the barrier to entry.

Games like Overwatch and League of Legends are likewise easy to learn and hard to master. You can just hop into a game and learn a single hero whose abilities take mere minutes to internalize. Games with a steep learning curve can quickly have population problems because most people find it difficult to invest heavy resources into learning a game when they’re not sure if they’ll later find it fun. When over half of your players never even kill the first boss in your game before quitting, you might want to spend additional resources improving the beginning experience.

This is similar to a trend in fiction-rather than having lengthy expository inundations at the start of fantasy novels explaining all the various minutiae of this universe’s properties, it’ll start in media res. Otherwise people just get bored and put the book down. You have to first prove to the player that the game is fun. Once they’re invested in it you can trickle in additional rules and concepts.

High skill floors not only protects against an immediate exodus of players, but also increases the range of players in the demographic for the game. Overwatch is a fantastic example of this-the majority of FPSes (first person shooters) all require their players to have adept aim. Overwatch instead includes classes that span the entire spectrum with regard to aiming ability required. Some have complete auto aim, others ease of aiming due to their short range and high rate of fire, and some are of course the archetypal snipers and pistol players. As a result, players with different aptitudes (positioning, team coordination, healing, etc.) can all thrive. This grows the possible player base by a significant amount because it can now appeal to many disparate demographics.

Speaking of Overwatch, they do a lot of things right. Let’s take the scoreboard-a lot of players complained that you couldn’t see everyone’s stats like in other FPSes. For example, in a lot of games you can hit TAB and see a breakdown of everyone’s kills, deaths, damage, etc., so why not Overwatch too? It certainly would have been easier to just include a table dump of everyone’s stats compared to their medal and scorecard system they currently have, and many players also request the simpler scoreboard, so what gives?

If I were to guess, I’d say it’s because a scoreboard makes players feel pressured, and makes you think everyone is judging you. With a public scoreboard everyone can see how well or poorly you are performing. If you are playing badly you feel self conscious which leads you play worse, and if you do well you get upset with your team’s performance because you can see how much better you are doing than them. You might ask them to do better which they obviously would already be doing if they could, and it just leads to arguments and flaming.

Similarly, it becomes harder to play with friends. When all stats are public and you see just how poorly you are doing, it makes you feel bad and not want to drag the team down. When this information is hidden, you can play easy knowing that no one can see just how poorly you are doing. More importantly, you may also be ignorant of how poorly you are doing, which doesn’t make you feel like a burden to your friend group and want to leave them.

The medal system is also ideal because an elitist only wants to know he’s doing better than everyone on his team, and the presence of the gold medals do just that. He can’t know how much better he is doing, but the knowledge that he’s at least doing the best is good enough for him. He can’t go as far as to flame his team though, because someone could be right behind him in terms of kills and damage, because that information is hidden.

The scorecards at the end of the game are a bit random too. Even if you did mostly poorly, you might get a card in some obscure category, like “Teleporter Uptime”. There’s then upvoting of each cards, which has literally no effect at all other than making other players feel appreciated. All anonymous, competitive, team-based multiplayer games are going to have a high base rate of toxicity, so all these little things help with curtailing the amounts of rage and frustration in games.

I could go on and on about things Overwatch does right. It really goes to show how the structure of games can heavily influence overall toxicity levels in the community. Competitive anonymous, multiplayer environments always leads to high levels of flaming, but my Overwatch experience generally defies this. Out of all the all the competitive games I’ve played, Overwatch has some of the lowest levels of toxicity I’ve seen, and this is with all chat, voice communication, etc. It’s really quite remarkable how cordial most people are, especially when contrasted with other games like League of Legends, which may very well host the most toxic environment online within its ranked system, despite harsh punishments for even slight infractions. The system put in place by the developers has a huge impact on overall hostility levels in the community.

Speaking of toxicity, Blizzard has had a recent trend in some of their games of removing all chat. That is, you can’t type to the enemy team. I’m beginning to think that Blizzard has some psychologists on payroll. Obviously the technology is there and is trivial to implement, but they choose not to implement it. Why? Well, in 99% of matches, it’s completely unnecessary. Very rare in competitive games is the enemy going to be complimenting you or being cordial to you. People also really don’t like being shit-talked, so the cost analysis becomes easy. That is, while removing all chat has negative utility in the rare case someone wants to positively interact with the other team, it has overwhelming positive utility in the common case of preventing flaming and mocking of the enemy team. The removal of all chat also makes the game feel less competitive and real, in that the opponents seem more like bots than humans, but Blizzard seems to think this downside does not outweigh the benefits, and I think I agree with them.

Another aspect that is important, and this is not just for competitive games, but games in general, is progression. The need isn’t as strong in competitive games because you can always be working towards that next rank up. You see this evidenced by the “gamification” of everything. Every game and activity is getting experience bars, unlockable cosmetics, achievements, levels, etc. Recent competitive games all involve a plethora of skins, loot boxes, levels, etc. Everyone is doing this now because it works.

For example, take a Dungeon Crawl like Binding of Isaac. It involves entering rooms, killing all the enemies, and moving on. Occasionally you’ll find an item to make your character stronger, and there’s bosses at the end of each floor, but if you die, you start over from square one.

Needless to say, if you actually lost everything and made no progress with each failed run, no one would play, or they’d get bored very quickly once they won a few times. Progression is added to the game through the form of achievements, unlockable items that will now spawn in future runs, unlocking new characters to do runs with, various goals (such as killing the final boss 10 times), etc. In almost every run, even if you fail, you at least accomplish something, and if you won, you accomplished even more. People are generally quite loss averse, and will avoid activities where they keep losing without gaining anything.

Fairness and difficulty is very important. People are very antagonistic towards “unfair” scenarios, and rightly so. I think this is a big reason why a lot of open-world PvP (player versus player) games fail. Individual skill can slide the scales a bit, but in most engagements the group with more players will win. As a result, the winning strategy is to typically “zerg” your opponent with larger numbers. Often times these games will have serious death penalties such as the loss of items or experience. Needless to say, combining loss aversion with unfair matchups depresses enjoyment and retention of said games.

Most games now counterbalance this by restricting engagements to equal numbers on both sides in the form of arenas and battlegrounds, with skill rating matching done to ensure relatively equal matchups. This is the case with the aforementioned League of Legends, Overwatch, etc. While everyone’s skill rating is slightly more fluid due to the team-based aspects of these games, the aforementioned psychological benefits outweigh the costs of slightly unfair matchups.

The above factors all contribute to startup costs. That is, the ease with which a player launches the game and queues for a match. Each negative factor plays a part in the final decision of “do I want to play this game?” Put another way, elements such as having a good sized team, fair matches, reasonable amounts of mental effort, etc. all lower the resistance towards the execution of that double click to launch the game. Players may not be consciously aware of all of this, but all the preceding ingredients translate to levels of fun and enjoyment. The aggregation of these collective feelings are realized in total player retention numbers over time.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s