I have a confession — I used to hate achievements. It started with Xbox Live, and I was pretty sure that I was the only one. It was hard for me to articulate why exactly I disliked them so much — so many of them were just, well, bad.

Let's toss our nostalgia aside for a moment and recognize that game design has come a long way over the past couple of decades. I loved the original Zelda, but no one is going to convince me that the process of uncovering secrets by methodically burning every single bush across dozens of screens was good game design. We've learned a lot about what works and what doesn't. We've refined and polished concepts and gameplay mechanics.

Why, then, did we throw everything we'd learned out the window when it came time to design achievements? I think that's what bothered me about early Xbox Live achievements more than anything — as goals, they violated too many game design guidelines. Too many of them wouldn't make any sense as in-game goals for progressing through content.

In June of 2007, I was tasked with creating site-wide achievements in user-submitted Flash games on Kongregate. I was forced to ask myself, "If I hate common achievement design so much, how would I do it?"

I've spent the past 2 1/2 years designing, testing, naming, and releasing over 800 achievements across hundreds of Flash games. I've obsessed over earning distribution data, game comments, forum posts, direct user feedback, and changes in game ratings far more than I've let on to Kongregate's often critical community (can't give 'em the satisfaction!). I've made (and continue to make) plenty of mistakes, but I've also learned a lot of lessons.

It's also worth noting that I have been obsessed with World of Warcraft's achievement system over the past year, the design/implementation of which has really blown me away in terms of thoughtfulness and overall user experience, even if it's had a few blunders — mostly related to highly random, time-sensitive achievements tied to a meta-achievement with an enticing in-game reward.

I'd like to take some time to share the lessons I’ve learned with all the game designers out there who are now adding achievements into their games — whether they be on Xbox 360, PS3, Steam, a Flash game, or an MMO.

These points are given with the assumption that the player is at least somewhat engaged in your achievement system. If he's not, well, how you create achievements in the first place is a moot point anyway!

1. Why is this hard, and is that reason fun? Ridiculously hard achievements are okay, if they're hard for the right reasons. In middle school, I was completely glued to GoldenEye 007 until I unlocked every damn cheat code, with the Silver PP7 in particular taking up 4 hours of a Saturday on its own. When I finally succeeded, it was one of the most rewarding experiences of my life (I haven't exactly climbed Mt. Everest here).

Here are some good reasons for achievements to be hard: high skill component, perfection of a core game mechanic, and handicapping the player if the new core mechanic of this handicap is still compelling (such as removing high-powered weapons when the low-powered ones are just as much fun).

Here are some bad reasons for achievements to be hard: low skill component (luck-based), perfection of non-central game mechanic that isn't overly fun on its own, handicapping the player by removing a fun element of core gameplay, excessively grindy, severe punishments for small mistakes after a long period of time, aggressive real-life demands (completing a grueling task within 24 real-time hours, for example), difficult logistics of even attempting a task (such as finding an active multiplayer game), and high reliance on specific behavior of opponents (especially in a multiplayer setting).

Something I quickly learned when designing achievements on Kongregate is that the more difficult the achievement is to earn, the more difficult it is for me to design it well. Because, well, look at the bad reasons above — it's like a damn minefield!

Return to these two questions whenever designing difficult achievements: "Why is this hard? Is that reason fun?"

2. Grindy achievements are fine only if they overlap with something that the player would naturally do anyway. "Grindy" achievements are defined as achievements in which the task is to perform the same action repeatedly, with little or no variance between each repetition. "1000 kills," "$100,000 earned," etc. They're extremely prevalent in achievement design because they're incredibly easy to come up with (none of that troublesome creativity), and they seem like a pretty harmless way to reward the player for playing a game a lot.

The problem with grindy achievements is that game designers will often underestimate how much brain-liquefying drudgery players will subject themselves to just to earn these achievements — it's often a negative experience for the player, and it makes the achievement system as a whole less fun.

Assign these achievements to stuff that the player is doing anyway, or at least something that the player wouldn't mind doing a bit more of. Grindy achievements are usually fine in multiplayer games because there's essentially no limit to how much a player will enjoy playing a fun game online, as any Counter-Strike addict can attest. In MMOs, grinding is generally an accepted aspect of the actual game design (though even WoW removes its most insane grinding task, aptly titled "Insane in the Membrane," from its core achievement system). Additionally, MMOs get a bit of a free pass because their success is highly reliant on making the player feel immersed in the game world, which is aided by the quantity of time spent playing. Games that don't rely on long-term player immersion don't really benefit from subjecting the player to copious grinding.

When I purchased Ecco the Dolphin over Xbox Live Arcade, I was somewhat baffled to see an achievement for completing the game 3 times. I'm not sure what the point of this is. The game doesn't change, and that's not really natural behavior — I was hoping to run through the game, relive a bit of nostalgia, then move on to something else. It never crossed my mind that I would repeat this process twice more for an achievement, after the game had already gotten my money.

In general, before implementing grindy achievements, ask if completing that task without the assigned achievement would be considered normal behavior. If it's not, the player has to change his behavior. When achievements cause players to change their behavior, the designers should ask themselves whether this change is leading to a better experience for the player. With the case of excessively grindy tasks, the answer is almost always "no."

3. Creating multiplayer achievements is like playing with fire; they influence player behavior in sometimes dangerous ways. I once set out to earn the "Battlemaster" title in WoW, an honor bestowed upon those who complete a series of PvP-related achievements. One in particular was giving me a bit of trouble — I had to kill 5 enemy players carrying the single neutral flag in Eye of the Storm, a battleground scenario similar to capture the flag. This task had to be accomplished in a single game, and, after a few games of enemy players not even touching the flag, I got frustrated.

So I got an idea. I recruited a few of my friends, we ran in together, and I grabbed the flag myself. Rather than capturing it for my team, I sought out lone enemy players, dropped the flag in front of them, and then backed away, hoping they'd take the bait. When they did, my friends and I would pounce, killing the flag carrier (thus satisfying the achievement requirement), grabbing the flag back, and then repeating the process.

After about 10 minutes, my team caught on to the fact that my little posse wasn't exactly working toward a team victory. "WTF R U DOING WITH THE FLaGG?!?!" an ally demanded. "Achievement, duh," I replied. The conversation grew more explicit from there.

But I'll admit it — I was being a jerk. I had to. If the achievement wasn't demanding it, it was at the very least encouraging it. And I had to have that achievement.

I was reminded of Halo 2's multiplayer achievements. Some were so completely ridiculous to achieve that the only rational way to attempt them would be to get everyone to cooperate. I heard rumors about people being banned for this, and achievement addicts fleeing to the Italian rooms to continue their collusion. I don't know if that part is true, but what I do know is that in half the games I joined, some guy would propose the idea of intended foes working together for the purpose of earning achievements.

What game designers in general often seem to ignore is that when players are presented a goal, their first inclination is to devise the most efficient (not necessarily the most fun) means of reaching that goal. This is true of any game, with or without achievements, single-player or multiplayer. Show the player the end point, and that player will take the quickest and easiest route, regardless of whatever path the game intended for him to take.

When designing multiplayer achievements, do not trust that they are tasks that the player will just naturally bump into. Put yourself in the player's shoes. Pretend that earning this achievement is the most important thing in your life, and no ethically questionable actions are beneath you. How will you earn it? What's the most efficient method? Once you know this, ask the most important question of all: "Will this method negatively influence the gameplay experience of others?"

The other problem with multiplayer achievements occurs when extremely specific behavior is required of your opponent. The player does not control his opponent's behavior, and when a crucial element of earning an achievement is beyond the player's control, it can get extremely frustrating. "Kill 10 enemies in X area" is okay because there will almost always be enemies in a specific area of a map, and these achievements can possibly have a positive effect on the gameplay experience if it's a naturally good spot for people to clump together and fight it out.

In general, multiplayer achievements are dangerous. Unless you're absolutely sure they're impossible to exploit and won't negatively influence how people play a game, I would highly recommend keeping them simple and in line with the general style by which a game is played.

4. Always evaluate the weight and fun of an achievement by the most efficient method by which it is earned. Again, players do what's efficient, not what's fun. It’s up to the designer, not the player, to ensure that efficient actions overlap with fun ones. If there's an achievement for completing levels 1 through 3 without dying more than 10 times, then the player is likely going to reset the whole game every time he dies on the first level (assuming no checkpoints), even if this process interrupts game flow and makes the play session less fun.

If there's an achievement for killing X enemies in under Y minutes, the player isn't just going to play through the game normally and hope he gets a good streak — he's going to round up the biggest group he can find, avoid touching them, then unleash his fury. If there's an achievement for killing 10 enemies without being hit, the player is just going to go to the easiest room on the easiest difficulty setting, find the easiest enemies and get to it.

And sometimes this is okay. Other times it's not. But the job of an achievement is to help the player have fun, to give the player an accomplishment, and to allow others to recognize what the player has achieved. Regardless of how the player earns an achievement, the weight of that achievement and the perceived value of it from other players is always going to be determined by the easiest method by which it can be earned.

If there's a room full of enemies on level 10, and it's a super awesome expression of skill to kill 10 of those enemies without being touched on the hardest difficulty setting, then make the achievement "kill 10 enemies on level 10 on the 'hard' difficulty setting without being hit." If the achievement is just to kill 10 enemies without being hit, even if the player achieves it by a means that's more impressive than playing the first level on the easiest difficulty setting, there's no way for other players to know; thus, the achievement isn't doing its job of helping that player show off how awesome he is.

Once the easiest path to accomplishing an achievement has been planned out, again ask if it's actually fun, or just mildly tedious. Players will pick easy/tedious routes over more difficult, fun ones. Make the achievement requirements specific to cut off potential exploits if you need to — don't just rely on players naturally having the experience you want them to. Force them.

5. "Holy crap, did you see that?!" achievements should either be easy or fully within the player's control to accomplish. Sometimes something crazy will happen in a game, and it'll be so over-the-top and ridiculous that anyone experiencing the moment will start laughing and pointing at the screen. It'll be fun. They'll want to share the moment. Achievement designers will want to enhance the experience by attaching an achievement to it.

The problem here is that once an achievement has been attached to it, it's no longer a fun little event that occurs naturally. It's now a very specific event that the player must painstakingly attempt to create. Doing a crazy stunt in a Tony Hawk game is fine because the player has full control over the situation, and it can be difficult for the right reasons (mastery of the game's core mechanics).

But when an event is highly reliant on enemies performing a specific way, or it's so crazy that it's basically left to chance (such as having perfect accuracy on a moving target with an inaccurate weapon at an incredibly long range), then it's just not fun anymore. The player is not going to play the game naturally and hope that it happens. He's going to create the ideal scenario for it to occur, then he's going to play through this scenario again and again and again, hoping that this crazy moment will happen and the achievement will be awarded. That's not fun, and it's not the spirit in which the achievement was designed.

Perhaps the worst-designed achievement in WoW has to be "We Had It All Along *cough*", an achievement awarded for having a PvP victory by exactly 10 points out of 1600. The achievement is agonizingly part of the aforementioned PvP meta-achievement (achievements naturally warrant far more scrutiny if they’re part of a larger meta-achievement). The idea behind this achievement had to be pretty simple — winning by exactly 10 points is a crazy moment in the game, and the designers wanted to reinforce this moment with an achievement. But once this achievement is added, it's no longer a crazy moment with an achievement attached; it's a moment of "thank God that finally happened by some miracle so I'll never have to worry about it again." The confounding problem here is that players who are adamant about earning the achievement are encouraged by the game's mechanics to simply stop contributing to their team's success if it appears as though their team might win by 30 points rather than 10.

By contrast, the WoW achievement "Mine Sweeper" is an example of this design done correctly — it's awarded for bouncing across 10 explosive mines without touching the ground. Sure, the in-game objective is to actually avoid the mines, but hitting 10 mines is not too unlikely by pure chance. For those who don't notice the achievement, it does successfully reinforce a fun little moment. And for those seeking out the achievement, it doesn't take long to earn.

6. Acknowledge the highest difficulty on which something has been accomplished. I've never understood achievements that require completing a task on an easier difficulty setting. If the player completes the game on "hard," he should not be required to go through and complete it again on "easy" just to earn the achievement. Have both achievements awarded for completing the game on a harder difficulty setting.

7. Achievements will thrust their subject matter into the spotlight; make sure it's worthy. This can be good or bad. In most cases on Kongregate, adding achievements to games will cause the user rating to drop. There are many theories about why this is — my best guess is that there's a difference in psychology between people who play a game just to have fun (how weird!) and people who play a game to earn achievements. For the latter category, the whole game can be viewed as merely an obstacle.

One thing is for certain, though. Nothing kills a game's rating on Kongregate more than poorly designed achievements.

I can only theorize about an analogous scenario with games on Xbox 360. Game reviewers probably won't mark a game down for having bad achievements, but I could easily imagine players changing their opinion about a game based on its achievement design.

Over the past couple of years at Kongregate, I've learned to dissect games: picking apart the fun elements and setting aside the stuff that's maybe not so great. Some of the worst achievements on Kongregate were created out of the mentality that if something is there, and it's significant, there should be an achievement for doing it. To hell with whether or not it's actually fun!

The power of achievements in driving user discovery didn't fully hit me until a game was uploaded called Dolphin Olympics 2. It's a fairly simple game where you control a dolphin jumping out of the water and doing a few simple tricks. It started out with somewhat mediocre rating on Kongregate — just barely high enough to be recognized with achievements at all.

What many players failed to realize, however, is that the game is actually pretty insane. Once you get past a certain elevation, you break free of Earth's atmosphere, spinning around planets, sliding along strips of stars, and eventually flying past the Restaurant at the End of the Universe. Once achievements were added to accomplish these things, the user rating skyrocketed.

A great power behind achievements is that they can force people to examine something they might have otherwise overlooked. If there's an awesome optional side quest in your game that's easy to miss, it might take a lot of work to change the game's flow to direct players toward that optional content, but it's far simpler to just add an achievement for doing it — the players will figure it out from there; don't worry!

Conversely, if there's a part of the game that didn't work out as well as intended, and maybe it doesn't quite shine as brightly as the rest of the game (the vehicle driving section of a first-person shooter, so to speak), then there's really no inherent value to forcing people to play it any longer than they want to just to earn an achievement.

And just because the player can do something crazy and miserable in a game doesn't mean there has to be an achievement for it. Think of the people who wouldn't otherwise do it, not the one(s) who did.

8. Pretend that the player dies if he reaches a point where an achievement is no longer earnable, and then evaluate how frustrating that experience would be and whether it makes any sense. Since death and reaching a point where an achievement is not earnable are essentially the same thing for a player who’s only playing something to earn an achievement, this can often serve as a useful check into how the player will react.

It’s really easy to imagine for "don't get hit" achievements. It's basically saying that the player has 1 point of health. And sometimes that's okay. Plenty of games cause the player to die and restart a specific section if he takes any damage. So those achievements will pass the test, unless the game is designed around the expectation that the player will take steady damage over time.

It starts to get a little trickier with timed achievements. Many older games will punish the player with death if he takes too long, but Shenmue is perhaps the only game to give the player a game over screen after taking too long and playing for an extremely significant amount of time (some might argue that the mechanic works in a hyper-realistic game like Shenmue, but it’s still incredibly rare).

Achievements that require completing a game in less than 8 hours are unlikely to pass this test, as it would be unbelievably frustrating to get a permanent game over screen halfway through a game after playing it for 8 hours. Instead, break it up into chunks, and require that individual levels be completed under a set time, then award an achievement at the end for doing this with each level. Same display of skill, far less frustration.

If there's a relatively bizarre condition for an achievement, such as not killing a specific enemy, not using a specific weapon, etc., then again, recognize that the player will likely restart if he's going for the achievement and these conditions are violated, so pretend that killing the wrong enemy or using the wrong weapon grant instant death.

No achievement is more bizarre (and, in my opinion, poorly designed) than achievements for dying a certain number of times. This is not an achievement — this is just a public record of the player's failure (except it's not even really that, since the player will likely go out of his way to intentionally die repeatedly on an early level just to earn it). This type of achievement reveals itself to be even more ludicrous when applied to this section's check — the player dies and has to restart if he doesn't die and have to restart?

9. Do not allow the player's save file to easily and accidentally reach a point where an achievement is no longer earnable. It is unbelievably frustrating to pass up a narrow achievement window, save your game, and then be unable to earn it without starting all over. It forces the player into a mentality of nervously using online guides to see when achievements are coming up, making sure they're all charted out, and then carefully earning each one before progressing naturally with the game. It completely sucks away the carefree enjoyment of just playing through a game's story, and there's really no reason for it. Let the player go back and earn missed achievements, if possible. If not possible, evaluate whether to have that achievement at all, or at least how it can be made difficult to miss.

For the vast majority of games on Kongregate, it is impossible to miss a window for an achievement and be screwed out of earning it with a specific save file. (The few exceptions are found in games that are relatively short and heavily time-based.)

Also, fun fact: Every single achievement in WoW is earnable by a single player character of any race or class, with no achievement ever becoming unearnable or requiring the creation of a second character.

10. Insultingly easy or excessive quantities of extremely difficult achievements can negatively influence players' perception of the system as a whole. Prince of Persia crossed the threshold of awarding an achievement for simply watching the intro movie, but this territory was flirted with before, and the practice continues today. Ridiculously trivial achievements are not truly "achievements." They're just records of something that happened — and not even interesting events, at that. They're the Twitter feed of achievements.

Just as individual achievements only have as much weight/meaning as the easiest means by which they are accomplished, the numeric quantity of achievements earned within an entire system only has as much weight as the easiest means by which a player can reach that quantity (or point value, if harder achievements are worth more points).

Thus, ridiculously easy achievements inflate the system as a whole, devaluing others in the system.

I saw this as a potential problem when I started designing achievements for Kongregate, and I had a simple philosophy to fix it: Achievements are achievements, dammit — they're serious accomplishments, not to be taken lightly! I wanted each and every one to carry some serious weight behind it.

A lot of games do this, too. There will be an easy achievement here and there, but 90% of them will require some serious dedication. The problem with this philosophy is that players won't feel like the system is really for them; rather, it's for some nebulous other person — someone who's better than they are, with more time to commit to that specific game.

To attach players to a system and make them feel invested in it, the achievement designer must strike a careful balance between giving the player a good chunk of achievements that are realistic to achieve, and making sure that these achievements still have enough difficulty to carry some psychological weight and value once they are earned.

I would like to think that the achievement system on Kongregate is truly for everyone. There are some easy achievements, but very few (if any) insultingly trivial ones. Most players can fill their profile with achievements, but only the more dedicated players can really rack up significant points in the system with the harder ones, which award more points. It draws people in and always gives them a little extra something to push for, which is what achievements should be all about, anywhere.

11. Don't use achievements to force the player down a specific path in an open-ended game. BioShock, why did you do this? An achievement for completing the good ending, but not the evil one? Not only does this violate the aforementioned guideline of not allowing the player to screw himself out of being able to earn an achievement later, but it's actually detrimental to the game as a whole because it completely removes the sense of choice from the player's end. Choosing between good and evil might have been an interesting decision to weigh if the achievement didn't essentially make it for the player.

Likewise, if there's an achievement for upgrading a weapon down a specific path, or getting a certain set of abilities, etc., all this is doing is removing a sense of choice from the player. You might as well just have a little "THIS IS THE ACHIEVEMENT OPTION" arrow pointing to the player's "correct" decisions whenever a specific path is required to earn an achievement.

So what should BioShock have done? Had a second achievement for completing the evil ending? I would argue "no," simply because the game itself differed far too little between good and evil — there wasn't enough alternate content to justify playing through the game again for most players.

When evaluating whether to make extra achievements for going down multiple paths, ask whether the alternate-path content is varied enough to justify playing through it even without the achievement. For example, it makes sense to have achievements for playing through all the starting areas in Dragon Age Origins because they're all vastly different and worth experiencing on their own, but it wouldn't make any sense at all to require the player to complete the entire game from each and every starting point, since the game's entry points converge relatively early on.

12. Don't make timed achievements for puzzle games in which the solution can be easily memorized. These might as well just be called "memorize what you just did and do it again" achievements. They're not fun, and there's no point to them. On Kongregate, I've taken a stand against any achievement design that's realistically impossible the first time a game is played, then ridiculously easy the second time. Just award the player for completing the puzzle and move on. (Randomized puzzles are exceptions here, since those time requirements can be balanced around the assumption that the player will not have a solution memorized.)

Again, ask why the achievement is hard. If it's balanced around an expected time the first run through, then it might be hard for the right reasons during that first play. The problem, though, is that the weight/value of the achievement is still pretty low, since it can be achieved easily on the second play through. If the puzzle time is balanced around the expected time for the second run through, then it's either going to be ridiculously easy (and needlessly tedious), or it's going to be difficult for the wrong reasons — expecting the player to quickly execute something that has already been solved, when the game's difficulty is intended to be derived from reaching the solution, not the speed of the execution.

13. "Secret" achievements (in which the earning criteria are not displayed) are dumb. The description for them might as well just be, "Go look this up online." There's absolutely no point to them, other than just giving the player an errand. The only people who are surprised to earn secret achievements are the people who wouldn't have bothered to check what the achievement is in the first place.

14. "The player doesn't have to earn the achievement" is not an excuse. If I'm a chef at a restaurant and I serve your sirloin steak with a side of dog food and gravy, sure, you can choose not to eat it, but it's still going to affect your opinion of me as a chef and of the restaurant as a dining establishment.

Most people will not earn most of the achievements that are presented to them. But how they regard the achievements that they don't earn, and what rationale they give for not earning those achievements, is important to how they view the system as a whole. If an achievement is ridiculously difficult (and difficult for the right reasons), then the player might give up on earning that achievement, but still have respect for the task and for the people who do manage to earn it.

For example, if there's an achievement in Guitar Hero for getting 5 stars on "Through the Fire and Flames," the player will acknowledge that he's not going to earn that achievement because the song is ridiculously hard, and he's just not good enough at the game. For the people who do earn the achievement, the player has great respect. It’s good design.

If there's an achievement for playing through the game using an Xbox controller, that's just not a fun experience that most people want to have, even if it’s not overly difficult. Most people are not even going to attempt it — and they'll cite the reason as being because the game is not fun to play with an Xbox controller, not because it is an amazingly awesome display of skill. The player will react to those who have the achievement more with "I can't believe you actually played Guitar Hero with an Xbox controller," rather than "Holy crap, that is amazing."

When players come across poorly designed achievements, the bottom line is that it lessens their psychological investment into the system as a whole. It is impossible for this to happen without also lessening the sense of accomplishment the player feels when he does earn an achievement in this system.

To quote the great Jurassic Park philosopher Ian Malcolm, "Your scientists were so preoccupied with whether or not they could, they didn't stop to think if they should." This was shortly before the entire park and multibillion-dollar facilities were torn apart by dinosaurs. Think about it.