Designer Keith Burgun examines the concept of achievements, looking at how they're used, how they might be used in the future, and how they might even change totally -- for the better.
Despite having strong feelings on the topic, I put off writing about achievements for a very long time. This was because I thought that the problems that I saw with the existing model would have gone away on their own by now. It's clear though that for new digital game releases, they have clearly managed to lock themselves into the "set of qualities we should all expect in a modern video game."
I know that there's a ton of writing on this topic already out there, but I'd like to hit the question from a different angle. For instance, I'm aware of what Chris Hecker has talked about at length about regarding extrinsic motivators. While I think his views make sense, I'm actually not interested in arguing for or against extrinsic rewards in general. I'm arguing against achievements themselves, and how they, specifically, work on a mechanical level.
Lucas Blair wrote an extensive three-part piece on achievements here at Gamasutra. His article essentially took the stance of, "we're going to be doing achievements no matter what, so here are some best practices for using them." I don't agree with his underlying premise.
The one thing that remains constant is that things always change. I think that in time, we'll see achievements either go away or change dramatically. If this sounds crazy, keep in mind that there are a good number of successful games coming out today that don't have achievements at all, on iOS and Nintendo consoles.
I need to clarify and explain that I'm referring to achievements as they are usually implemented. I'm sure you can think of one or two games that seem to have a sensible, inoffensive and even interesting application of something that looks a bit like what one might call "achievements." It would be impossible to speak for every single case of achievements that ever existed. Instead, I'm speaking generally.
You may feel that achievements are great as they are. If this is the case, hear me out. Perhaps I can convince you otherwise, or at least, give you some advice on how to make them better.
Counter-Strike: Global Offensive is a fantastic example that highlights the problems that I see. Now, this is a 2012 release -- in fact, it was only released this past August. It is a brand-new game, by one of the most beloved and highly regarded triple-A video game development teams (Valve Software), and sure enough, it includes a ton of totally asinine achievements. I'll also include some achievements from the also brand-new games XCOM: Enemy Unknown and Resident Evil 6.
Before I carry on: don't worry. I don't expect achievements to actually disappear anytime soon. They're pretty well rooted into our culture now, and if they ever disappear completely, it will probably be decades rather than years from now. Indeed, you're required to have achievements on Microsoft and Sony's consoles (but notably, not anywhere else: neither iOS, Nintendo, Android, nor Steam make any such requirements). Regardless, I think it's in everyone's interest to understand the pitfalls of achievements; not all of which have been pointed out elsewhere.
I'll also pitch something that I think should replace achievements.
What's so bad about achievements? The mother-problem with any "achievement" system can be stated like this: at their best, they do nothing at all. At their worst, they influence player behavior.
What's wrong with influencing player behavior, you might ask? Influencing behavior is a bad thing because you (ostensibly) just spent roughly six to 12 months fine-tuning a set of game rules to do exactly that. Let's remember that a game is a set of rules that limit and motivate player behavior. You just spent a crazy amount of time tweaking, balancing, and turning knobs until player behavior was influenced exactly the way you wanted, all around one central goal and gameplay mechanism.
If you did not do this, well, that's a whole separate issue. In this case, you're simply not doing your job as a game designer, and no amount of metagame is going to distract people from the fact that your game isn't presenting players with interesting choices and dynamic, emergent and elastic strategic possibilities.
So let's assume that you have taken the time to create a balanced, dynamic, motivating set of rules for your game. Now you're just going to throw a bunch (most times, a ton) of other arbitrary motivators at the player? A great number of extra, optional goals that can be met even by accident? It's like spending years building a clock, and then just once you're done, pouring in a bag of random-sized gears and slathering over it with a dressing of industrial glue. In this way, achievements are yet another testament to the culture-wide lack of regard for the discipline of game design.
The largest category of achievements is of a type that I would describe as "unavoidable," "patronizing," "noisy," and sometimes even just "nonsensical." Here are a couple good examples from CS: GO:
"Body Bagger - Kill 25 Enemies"
"Shot with their Pants Down - Kill an enemy while they are reloading"
Here's a similar one from XCOM: Enemy Unknown.
"Bada Boom - Kill 50 aliens with explosive weapons"
And again, basically the same achievement for Resident Evil 6:
"Life Saver - Help or rescue your partner ten times"
Ah, the achievements you cannot avoid getting. You're going to kill 25 enemies. Chances are you're going to kill 25 enemies before you even think to check the list of achievements (if you ever do).
Therefore, a lot of players are simply playing the game, and suddenly some information pops up on the screen telling you that you have just "gotten an achievement." This totally meaningless information does not change the game in any way, except to temporarily distract you from the game.
These achievements also do one other thing, however, and that's patronize the player. Did you already design the game to have its own rewards/motivation system? If so, then what is the purpose of having the game to pat me on the back at arbitrary moments? 25 kills? Why is that significant? The rewards that the game gives me are those that I ostensibly have to earn. Not the case for these achievements. You may as well have a timer that doles out a random nonsensical compliment every 15 minutes, such as "you are attractive" or "you've got a great sense of humor."
Without going too far off topic, I want to quickly address this aspect. Those who are familiar with B.F. Skinner's work, particularly in operant conditioning, probably understand that doling out rewards at random intervals, like the current achievement-model tends to, is a well-understood way to squirt happy-chemicals into a user's bloodstream and thereby keep them playing long after they've stopped learning anything. Philosophically, I personally think that games have the capacity to do much more than just be unfulfilling exploitative operant conditioning chambers, but even if you don't, you should be aware that this common system of achievements is causing a similar effect.
Attempts to Script The Emergence
The one way that achievements are commonly talked about is with regards to them being an extrinsic reward -- a reward that's coming from outside the system. I join the aforementioned Chris Hecker and many others such as author Alfie Kohn in being skeptical of these kinds of motivators when applied to interesting tasks. Our view is that they take away from the feeling of accomplishment for a task that's already interesting and naturally rewarding.
I'd like to look at this problem in a slightly different way. First, let's take a look at a couple of CS: GO achievements which exemplify the issue I have in mind:
"Three the Hard Way - Kill three enemies with a single HE grenade"
"Aerial Necrobatics - kill an airborne enemy while you are also airborne"
Here's a good one from XCOM: Enemy Unknown:
"Xavier - Mind Control an Ethereal. Single player only."
Let's think about the concept of an explosive grenade in Counter-Strike for a moment. When you buy one, it's exciting, because of the possible destructive potential. If you happen to put one in just the right place, who knows how many people you might kill in one slickly placed move? You may just damage a few people, you may kill one, or you may even kill several. This elasticity makes grenades dynamic and dramatic, and you feel it.
When you throw a grenade, and it actually does kill someone -- or better yet, two, or even three people -- it's a huge rush. All of those times that you got a grenade and didn't use it, or used it but to no effect were all leading up to this moment. A feeling of having gotten better at using grenades -- a grokking of the system of grenades -- is thrilling. You were in a totally unique situation and you made a call that resulted in an almost magical success.
Just then, a little window pops up and tells you that you've gained some kind of achievement. Suddenly, part of that thrill of having done something dynamic and unique is taken away. On some level, you've merely checked off a box -- the same exact box that thousands of other players have also checked off.
The XCOM achievement is similar. What would otherwise feel like a clever "giving you a dose of your own medicine" turns into a "thing you were supposed to do."
I argue that the fact that the developers wrote this thing down for you to check off of a list has a subtle effect of making the event less special. Let me imagine. Let me discover. Let me experience a moment of having done something truly unique without telling me that I've met some developer expectation.
To those who might argue that achievements such as "Three the Hard Way" are needed to get people to even realize that you can kill multiple players with a grenade, you should know that the original version of Counter-Strike didn't have achievements, and HE grenades were very popular. Players don't have to be verbally told everything; some things are obvious and natural enough for players to discover.
Collect Them All and... What, Exactly?
As part of my philosophical view of what games are, I have a problem with collection for its own sake. I think that any system that is based on endless collection, or any system where there is collection without a larger purpose is exploitative and uninteresting (and therefore unfulfilling). It's exploitative because it's taking advantage of the biological human need to "gather", and not giving us back anything in exchange for our time. Most games challenge us, stimulate us, move us. Those that exploit us do nothing for us but the cheap.
In a game like Counter-Strike, why exactly do I want to collect all of the achievements? The game keeps track of what "percentage" of the achievements I've collected. Does something happen when I get 100 percent? What is the purpose of keeping track of this information? Is it supposed to impress my friends when they see 35 percent? When they see 95 percent?
And then what about when I do finally get 100 percent? Then what? It's just a dead system hanging off the side of the application? Does it make sense to have a game like Counter-Strike, one that can potentially be played forever, have some exhaustible collection system attached to it?
This is the worst offender of the whole achievements system. As I previously stated, a game already has its own motivators -- in fact, the purpose of a game designer is balancing motivators around a goal to create the intended gameplay experience.
But some achievements actually influence players to act in ways that they would not normally act. I remember this kind of thing happening a lot in Team Fortress 2. Often there would be a medic doing something really stupid instead of healing teammates. Angrily, I'd ask, "What the hell are you doing, dude? Heal us."
"I'm going for an achievement", he'd reply.
This is really not that rare an occurrence, particularly when a game is new. We now have a situation where players are actively not playing correctly and disturbing or ruining the game experience for other players because of achievements.
A common mistake would be to blame this on that player. Let's put it this way: if you're blaming a player for wanting to make use of the system of achievements, then you're proving my point even further that they need to go.
Here's an example of such a behavior-influencing achievement in CS: GO:
"Second to None - successfully defuse a bomb with less than one second remaining"
It's not too hard to imagine that many a game have already been lost by a player miscalculating when he should start the defuse and having it take too long, or by waiting a few seconds before defusing only to be shot right at the end of the defuse. This sucks for the other players on the team. Remember, the goals of a game should be agreed upon by all participating parties.
My Suggested Replacement: Variants!
Is there anything salvageable to this whole mess? Yes, there is. Some of the achievements -- those most-offensive ones that influence behavior, specifically -- have the potential to be interesting variants. While I don't expect achievements to vanish or dramatically change overnight, variants provide an alternative route that should be explored either in their place, or in addition to achievements.
What's the big difference between variants and achievements? A variant would be a new goal that you actively choose before the game begins, and only that single chosen "goal" is active during this session. One of the fundamental aspects of "a game" is that the rules and goals are agreed upon before the game begins. It doesn't make any sense to allow players to choose what their goals are on the fly, in the middle of the game. This will just allow them to choose whichever goal is most doable based on "how things are going". Worse, if you allow all the goals to be active at once, goals are going to be met by accident.
In Nethack, variants are referred to as "conducts." From the Nethack Wikipedia page,
These are voluntary restrictions on actions taken, such as using no wishes, following a vegetarian or even vegan diet, or even killing no monsters.
In Counter-Strike, being a multiplayer game, variants would have to affect all players. It would be strange if the Terrorist team won, but one of the terrorists lost because he had activated some special variant that said he wasn't allowed to take grenade damage, or something. Technically, there's nothing wrong with this, as long as all players agree to it beforehand, but it's messy and strange.
Instead, better Counter-Strike variants are already seen on public servers. Things like "No AWP/Auto", or "Infinite money", or "Betting" would all count as variants. These pose a new challenge to players -- "can you win this match when the AWP is disabled?" There are other more otherworldly server-variants that add RPG elements, zombies, and other rules.
Look at this achievement from XCOM, and tell me that it isn't a full-fledged variant waiting to happen:
"Lone Wolf - Clear a UFO crash site with one soldier on Classic or Impossible difficulty."
Why We Use Achievements
As a developer myself, I think that there's this feeling like "the audience expects achievements, so let's humor them." I suspect that players probably feel a similar way; something like "oh, well, the developers like to put in achievements for some reason, so let's humor them." In other words, few people actually like achievements, but everyone believes that everyone else likes them, so they continue to exist.
I also think that it's continued to exist because, if we're being honest, a lot of video games these days are not terribly interesting on their own. The thinking is that developers can use the cheap distraction / lame collection-game that achievements provide to create interest in an otherwise uninteresting system. Their primary function, much of the time, is to stretch out what little interest there is over a larger amount of time by compelling the player to "collect". They stand out the most when they're in a game that doesn't need that – a game like Counter-Strike: Global Offensive.
It's important not to fall into the trap of thinking that just because we've had achievements for over half a decade that we will always have them. Now, I'll definitely acknowledge that there is indeed a chance that we will always have them, at least in some form, but it's worth noting that Nintendo has made a point of not using such a system, and that hasn't seemed to affect their commercial or critical success. As I've pointed out, there are a number of flaws with the achievements model, and as time goes on, what I am certain of is that they will either change drastically or disappear.
If you're a fan of achievements, I would simply ask that you try to look at them with a fresh perspective and ask what it is they really do for your software, and whether or not the points I've raised creates issues for it.
So look -- people expect "metagame," and I understand that. But if you have great metagame in the form of variants, great networking (such as cutting-edge, smart online leaderboards), as well as additional gameplay content, the number of people who flip out because you don't have "achievements" will be negligible. At some point, people will stop expecting them, as quickly as they learned to expect them in the first place.