Boss battles are among the oldest and most iconic of videogame traditions, so I was pretty excited when I was lucky enough to find a few among the level assignments of my latest project at work. These past few weeks I have taken myself on a crash course through the processes and demands of designing good boss battles, driven by the unforgettable nostalgia-coated highlights and disappointing lowlights of my own gaming memoirs. Late into the midnight oil it occurred to me this was just the kind of juicy clean-cut topic that I could happily return to my blog with, so here we are blowing off the cobwebs. And I mean literally, if my home keyboard was a limb it would have atrophied and fallen off by now.

Action and Reaction: Rock, Paper and Scissors
It seems like common sense, but anything the boss can do, the player has to be able to somehow anticipate, and subsequently defend against or dodge. This means two things. Firstly that any attack the boss can perform should come with a visual or audio cue for the player to see or hear (just like the enemies of my combat article), and that the bosses move list is ultimately determined by the move list of the player. If the player can jump, the boss could perform a shockwave which the player has to jump over. If the player can aim and shoot easily, the boss could fire projectiles that must be shot out of the air before they reach their target. If the player can already perform a block or dodge, well, this makes creating the boss move list considerably easier.. but at the very least the player can usually move around, which means the boss can perform attacks which affect a particular area of the environment, and require the player to run around/away. If you analyze the players move list first, you can then vary the boss' move list accordingly so that different attacks require the player to choose a different defensive attack from their repertoire, thus creating a nice rock/paper/scissors scenario. (Again taking a leaf out of my combat article.) The rock/paper/scissors scenario keeps the player thinking, and attentive to the boss' actions; they have to engage with the battle and learn/read the bosses visual and audio cues to survive and win, which is a pretty good starting point.
Find a Pattern: Boss Behavior
What goes on in a bosses brain? Most will remember how many old school bosses would simply run through a particular attack pattern which we would then memorize by spending a few hearts or continues, and use to predict the bosses actions in advance from then on. This is the easiest set up for a designer, but it does quickly reduce the boss from a believable organic nemesis to an unthinking clambering machine, stubbornly plodding through it's moves one at a time without any consideration of the context. This is alas one of my big boss bugbears. All bosses inevitably run on cold hard code and predetermined 1s and 0s the same as any other aspect of a videogame, but when possible we should do our best as designers to hide those limitations, particularly for a “boss” who is a living breathing adversary, not a automated environmental hazard.
The alternative is for the boss to act based on the immediate situation, and because the only agent of chaos constantly changing the situation is the player themselves; it's the player that makes the best controller for the boss. If the boss can read the players position and actions and factor them into it's decisions, a much more believable game of cat and mouse is created. At a most basic level, the boss could vary between ranged and close-up attacks depending on the distance from the player. In more advanced scenarios, the boss might read the tactics the player is employing and adapt to counter them by shifting what moves it prioritizes. It's up to play testing and the designers assumptions to prepare the boss for possible player strategies that might scupper the intended gameplay, whether the boss is using a pattern or a library of catered reactions.

Shields Up: Defence Mechanisms
As much as the player needs opportunities to defend and dodge against each of the enemies attacks, they need controlled opportunities to return the favor and deal some damage. The easiest (and thus not always the best) solutions are to keep the boss out of reach or in possession of a standard immunity to the players attacks, either for certain pre-set sections of the battle, or until the player succeeds in completing a pre-determined task or puzzle to lower that immunity/get in reach. It's my opinion that a boss battle is essentially a level, and thus should rely on the player applying the existing core mechanics to the situation in order to win. This is particularly true if you take the background puzzle/task route, but be careful because there is not many things in this world more frustrating than trying to achieve something while somebody consistently spams you with pokes and projectiles trying to get your attention. Either way, make the bosses weakness the players strength, and there will be a much greater sense of outsmarting and overcoming the boss. If the core mechanic is a grapple gun, grapple the bosses armor off, pull something down on it's head, zipline onto it's back, tangle it's legs, affix it to the walls.. use the tools that are available.

Hit the weak spot for massive damage!
Another classic defense mechanism for bosses is the “weakspot”. If this spot is only available at controlled intervals or after puzzles, it is basically the same as the out of reach/immunity approach, but some bosses wear their weakspots loud and proud the whole battle. The fact it is slightly trickier to hit the bullseye than the dartboard is sometimes just enough increased challenge to stop the player spamming out damage and draining the boss bar too quickly, particularly if that spot is hidden on the bosses back or lesser spotted areas. This system often links well with another design tactic in boss battles, that of the boss opening itself up to damage through some of its own attacks. Just as each boss attack must be designed so it can be avoided, certain moves may leave the bosses defenses down or “unintentionally” reveal that tell-tale weakspot. Ideally if this is the method you choose, more than one of the bosses moves should be open to reprisals, or the player will simply find themselves waiting for that move to come up, unable to take a pro-active approach to the battle and quickly recognizing the boss for the computerized flow chart sham that it is.
Fool Me Once: Shame on you
I absolutely loved Arkham Asylum, but it's a tragic irony that the thing that really got my hype bubbling in the first place was the idea of taking on Batmans gritty re-imagined rogues gallery. It turned out the showdowns with Gothams foulest were by far the weakest most disappointing aspect of the game; utterly tepid boss battle design that made every classical mistake. A sort of “How not to do boss battles 101”. I have talked previously about the “rule of 3” in games, but while repetition in level design is a misdemeanor, applying the rule in a boss battle is a design felony. What kind of high functioning evil-doer with half a brain would suffer a hefty beatdown thanks to the player performing a particular tactic or task or exploit, and then patiently allow them to do the same trick three more times without altering their own approach? I'm looking at you, Roid Rage Joker. Avoid this archaic mistep at all costs; it does save a lot of think-tank time and allow the player to get into a comfortable flow, but if it was any other level in the game, would you design a third of it and then just use that third two more times? Of course not. Let's try and drum this rule of thumb out of the medium. Also on my boss bugbear list (along with set attack patterns and the rule of 3) are simply including “artificial” windows where the player can attack, such as the boss stopping to chuckle maniacally or twiddle with a malfunctioning tool, or simply taunt the player out of sheer hubris. This is the evil twin of the “bosses own attack opens them to attack” approach, and is just a little too obvious. After the first chuckle cost them a few broken ribs, I think they'd start to take the battle more seriously.

Parry Time Excellent: Player Versus Boss
A good stress test for a boss design is what happens if the player throws all caution to the wind and simply bombards the boss with attacks. A lot of designs are not prepared for the player to ignore all obvious danger cues and atmosphere and simply slam the situation with brute force rather than strategy. This is why it's so much easier to control the time and place a bosses defenses are down, as it means consistent attacks are just the player pissing against the wind. This said, many bosses don't bother with immunities, or shields, or weakspots, or keeping their distance, they take the player on mano-o-mano with nothing but a heck load of health bars and a high resistance to incoming damage. These are the more challenging to implement due to the players ability to go “all-out” at the boss (particularly once you consider the boss having stun states), but in combination with aforementioned “catered reaction” method, it arguably makes for the more sophisticated endgame. What's important is to ensure the boss has something up it's sleeve to counter the players onslaughts. A clean solution is usually to add a “keep away” move to their library that kicks in after a certain amount of close-up damage. This can range from an auto-parry/implacable block that quickly stops the player in his tracks, a heavy attack wearing snazzy superarmour (during the attack the boss can still take damage but cannot be stopped or interrupted) to force the player to a distance, or simply make a swift movement (a dash or teleport) to somewhere the player isn't. Decent players will soon count the maximum amount of attacks they can get away with before this keep-away move kicks in, and not push their luck any further than that.

Lair Sweet Lair: Property Rights and Wrongs
Like I said before, boss battles are level designs, and the environment where a boss lives is thus an extension of the battle and boss itself. The environment can provide places for the boss to hide or retreat (particularly key to out-of-reach scenarios), it can provide the player cover or shelter that might be required as a defensive action against certain boss attacks, the boss may even use the environment in it's attack pattern; temporarily or permanently modifying the area to increase the threat and challenge, Robotnik style. Or perhaps it houses the “background” puzzle required to gain advantage over the boss, because as we know, a surprising number of self-respecting bosses decide to battle the player in the only place in a mile radius where their only weakness is immediately available to the player. See Resident Evil 5's ridiculously old school boss designs, such as the Uroboros, an almost indestructible creature only vulnerable to fire, that chooses to fight you in a furnace, and then beside a Flame Thrower refill station. As ridiculous as this practice is, and as preferable as it is to use existing mechanics than introduce environmental ones, this method does have it's nostaglic charms, and there are far worse mistakes to choose from.
Regardless of your approach (but I'm specifically looking at you, out-of-reach fans) ensure your boss isn't camera shy. There is little worse than a boss whose location you cannot readily keep track of being especially if they have ranged attacks AND the blind spots to use them. You can temporarily hide the boss to create tension, but as a rule think you should try to keep the slippery sods on screen, provide a quick way for the player to optionally check their location (lock on for example) or really polish up the timing of those audio cues. As a final minor boss bugbear, try not to include any moves which require artificial or infinite replication of the environment. For example “spawning” dropping rocks/debris from above or “ripping out” chunks of the ground to throw. Unless of course, it's a cartoony sort of a game where that sort of thing flies without comment.
Finale: Epic Staging and Sympathy
The setting of a boss battle of course has a secondary purpose; it is a key aspect of atmosphere and presentation. Outside the nuts and bolts of the design, a boss battle is a climactic showdown, a crescendo, the culmination of a rivalry, a final ultimate challenge; although it shouldn't bend or break the rest of the experience, staging and spectacle is important. If possible, push the environmental choice to a place fitting and sympathetic to the event at hand. More than other levels, a boss battle is a part of the story played in-game, it will be remembered and revered only if all elements come together as a cohesive whole and compliment each other.
Do!
- Include audio and/or visual cues for all boss attacks
- Vary the ways the boss' attacks have to be avoided to keep the player on their toes
- Use the existing mechanics in creative ways to help take down the boss
- Include ways for the boss to consider and react to player position/action
- Control the player chances to deal damage or include a way for the boss to stop/avoid/counter constant player attack
- Try and keep the boss on the screen
- Where possible choose an exciting/fitting backdrop for the battle
Don't!
- Trap the boss in a rigid attack pattern
- Include “artificial” windows where the boss stops to laugh/taunt etc.
- Use the rule of 3; no “do this twice more” design
- Have the boss fall for the same trick over and over where it's unrealistic
- Have the player wait for a particular attack to open an opportunity to respond
- Use the environment in unrealistic/artificial ways
Thanks for reading! It was fun to get something new up here again.
-Steve



Another good post Steve :)
ReplyDeleteBoss battles really arent my thing and frustrate me more than its worth to keep playing. If a boss is infinitely better than I am then I tend to feel like I'm on the loosing team and there's no hope. I have come out of some boss battles alive, but only because the game kept me playing and not the boss battles. Games such as Doom and Resident Evil Nemesis.
However I've still not finished Dead Space 1 because of the battles battles being so tediously boring but also stressful. I only finished Dead Space 2 because I knew people were watching me and I didnt want to fail. I dont find any joy in killing an arch nemesis, although there are some battles I have enjoyed which I dont think you necessarily touched upon in the post.
The type of battle I like is when I am equal to the boss. The boss will be able to use the exact powerups, pickups, weapons that I have, although they will have a few little bonus' just to keep you guessing. Maybe you have 100 health and they'll have 200. Maybe you can jump 1 unit and they can jump 2. Games such as Killzone 2 do this brilliantly with Raddick (I think his name was!) But he was impossibile to kill, I spent hours and hours on this one battle..but in the back of my mind he was only human (kind of) so I knew I couldnt let him defeat me.
In Crash Bandicoot when you have the 9 platforms and certain ones start flashing and then drop down whilst theres an angry villian chasing you. He can make the same mistakes as you and fall..but you feel equal in a way.
Theyre the ones that I find most enjoyable and when I come across one that is frustrating and takes up 50% of the entire game, I tend to just turn it off.
Thanks for posting another article man, you really need to find a platform where you can get more bloggers attention! I think I told you one once but AltDevBlogADay is a good one too!
Good stuff Steve
Pete
Too long didn't read ;P
ReplyDelete