A version of this article first appeared in the August 2001 issue of Game Developer magazine.

There’s an epidemic sweeping across the games industry.  It’s a sweeping onslaught of gaming tedium that makes an average day of C-Span seem like New Year’s 2000.  I am referring to our boss monster encounters.  The elements of the game that should act as the climaxes of our gaming experience are devolving into boredom and frustration, instead of providing the pure gaming bliss that they should.

Now, I’m not naïve.  I know that boss encounters are usually lackluster because they require special code and art, which in turn translates to money, time, and bugs.  Also, they are often moved to the schedule’s end, which makes them ripe for gross oversimplification (if they don’t get cut altogether).  Complex bosses also cannot be reused easily – and resources you can only use once per game are extremely expensive.  Still, we can do better.

In this age of beautiful graphics and sound and well-crafted stories, most boss encounters are still “whip out your biggest gun, go mano a mano, and hope you don’t die.”  Which usually devolves into a health meter that moves down too slowly and way too many quickloads.  We’ve seen minor progress, but this usually “Shoot him when he taunts”, “shoot him in the stomach”, or if we designers are really clever, “Shoot him in the stomach when he taunts.”

Been there.  Done that.

Boss monster encounters should be the emotional highlight of your game, the apex of adrenaline, the part of the game that guys describe around the water cooler.  And yet, I often hear about people reaching for cheats as soon as they hit the first boss, because us game designers are still making the most basic mistakes.  Such as not putting a save point before the boss encounter.  Or not giving any visual feedback of how much damage has been done.  Or giving your boss critter insta-kill attacks.  Or, God help you, incorporating jumping puzzles into the epic fight.

To all of the game designers out there who are contemplating the boss encounters for their game, I would ask them to do three simple things.

Know the limitations of your game.  Quite simply, identify what in your engine is fun, and what isn’t.  Then don’t base any boss encounters around the ‘not fun’ parts.  The easiest example I can think of is swimming.  In many 3D console games, swimming is difficult and unwieldy, due to the limitations of the controller.  If this is true for your game, then don’t put in underwater bosses!  Remember, just because you can do something with your engine, doesn’t mean you should.

Multiplayer games should be aware that latency puts severe limitations on their boss encounters.  An attack that can be easily sidestepped in a single player game may become impossible to dodge once latency is added to the equation.  Timing-related vulnerabilities (i.e. shoot him when he taunts) offer similar problems.

Server-client games like Diablo or EverQuest should never let a boss hover around a teleport-in location, especially if there are long load times.  Few things make one want to chuck a CD in the trash quite like dying before your avatar is even visible.

Challenge the notion of a boss monster.  Bosses are merely the emotional apex of part of your game.  They can be reinvented.  Getting out of the space station before it blows or trapping Bossy in an Airlock and releasing him into space also have the potential to offer a similar high.  Serious Sam did a great job at this by making you kill a thousand little critters instead of one big one.  Conker’s also did a great job, by having boss encounters that incorporate the environment.

Also, consider that one does not have to be limited to increasing the challenge level.  One can also increase the power of the player.  After all, the real point of a boss experience is to provide an emotional high point for the character.  Consider that an even more potent psychological drive than one of fear is one of self-actualization – the notion that, for a little while, your player can go toe to toe with the big boys.

Have mortals playtest your boss encounter.  I buy any shooter I can, and typically rank high in any deathmatch game I play.  So when I face a boss monster that crushes me the first 19 times I try, my first thought is ‘is it POSSIBLE for Joe Wal-Mart to get through this?’  And if Joe Wal-Mart gets crushed repeatedly by the first boss, will he finish the game?  Or buy the sequel?

If you must have extremely tricky boss monsters, then be sure that you give obvious clues as far as how to beat the monster.  Another alternative is putting in some kind of workaround.  You don’t have to just open a door.  You can also give hints at the 5 minute mark.  Have superweapons appear at the 15 minute mark.  Reward the player who can’t win, but can survive.

Boss monster encounters in most games right now are below the quality of the rest of the game – but they don’t have to be.  Even with the harsh constraints put upon us designers, it is possible to make encounters that elevate the game to new levels, instead of making players reach for the cheat codes.