Fighting the Finale

After an up-and-down struggle with what Naughty Dog considers “easy” difficulty, I finally finished Uncharted: Drake’s Fortune (PS3) recently. For all the things the game does right – it’s a magnificently polished title – the end sequence is an absolute disaster of rigid game design, bringing to mind the question of how to build final-bosses that don’t invite turning the game off.

Maybe we can’t get rid of this style of finish, but the least we can do is consider our options and plan appropriately. The way I see it, there are typically three schools of thought for bosses:

Make it Easy

The theory here is that the player has already suffered enough. Do something friendly, give the player the ice cream, and let’s head to the credits. It’s a safe-bet strategy, which is good for projects with not a lot of time, but also risks leaving the player with a tepid “where’s the climax” feeling to their adventure.

Key Example: Beautiful Katamari (360)
For all of the difficult moments in Beautiful Katamari, the end level really is a satisfying way to finish. The player begins a re-enactment of the King’s childhood, being asked to periodically roll things up that he wants to think he liked at the time. But the player is essentially left to his own devices, free to build a katamari big enough to swallow the solar system without time constraints. It’s exactly the gameplay players want from just hearing about the concept, only without the shackles of progression that would have applied earlier.

Make it Difficult

The old-school, Nintendo style theory goes that the final boss is the player’s final hurdle. A giant, monstrous, very difficult hurdle in which the player must execute extremely well using the abilities they’ve learned. It’s a logical way to do it, but very reliant on the player actually being able to execute like you think they will. These cases also tend to suffer from “rails” syndrome, where there is little or no room for creativity in finishing the encounter.

Key Example: Metroid Prime (GCN)
The Metroid Prime series is notorious for the Make-it-Difficult style, so for argument’s sake let’s bring up the first game. In a 2 stage, no save fight, you have to attack using different guns based on presented colors. Then later, you need to switch between all of the visor types in order to even see it. There’s also a new beam type that is environmentally activated, and the only remaining unused gameplay elements (bombs) are needed to fight off the swarming metroids.

Quite frankly it felt tiring just writing all of that out. Now, the idea certainly makes a lot of sense. Here’s this mega-boss that forces you to express mastery over all of the equipment you’ve acquired over the last twenty hours. Visual cues tell you how to rotate through your gameplay options, and theoretically, the player should have all of the skills and training to win. But it’s also a very strict, on-rails encounter. The boss tells you, the player, what your only strategy is for the next minute. Deviate from the script, and all you can expect is punishment. In the old days, that would be “INSERT COIN” but now you can only look forward to the title screen.

Try the Unexpected

Not anything else, per se, but this line of thinking from a designer assumes the player is spent on the game mechanics and using them further would come across as boring or aggravating (and believe me, this happens more than the average team might like to admit.) Instead, the theory goes, do something else, probably something simple, just to reach the game’s conclusion without losing attention entirely. The trick is that if you do something too different, the player feels alienated by something they had no chance to prepare for.

Key Example: Ace Combat 5: The Unsung War (PS2)
Arguably Namco’s finest entry in the series ends with an amazing and unusual finale. Sure, the player still has to dogfight a little bit. But with a full orchestra and chorus blaring, they are also tasked with destroying a large, undefended object that is crashing towards the earth. Sounds simple…until the player realizes that every hit sheers off metal plating that, according to gravity, could fling itself at the player at speeds of roughly mach 3.

While the core gameplay (steer and shoot) is untouched, the nature of the “boss” and the sudden change in cause-and-effect (nothing else sheers into pieces like that!) more than qualifies AC5’s finish as a case of “something else”.

Now, from a practical standpoint we can’t just ignore any of the three. Have a game where you’ve built up lots of equipment and stats? Difficult might make sense sometimes, as a validation of the formula you’ve developed. Forcing the player to restart often? Easy would fit the bill better. And doing something else entirely makes sense in a lot of situations where running and shooting for the thousandth time just isn’t going to be fun anymore.

Just keep in mind whether the player will think your design choice is fun twenty hours down the line, because when that ending hits, fighting that boss will be the latest memory in whether your games gets replayed…or resold.


  1. Radiant

    The final boss is a touchy subject indeed. And as you had said, a lot of it should depend on what you’ve been doing for the rest of the game, and what type of finishing you want the player to have.

    Harry Potter and the Prisoner of Azkaban was a perfect example of how NOT to make a final battle. There was no boss, you had to fend off a bunch of enemies, and then that’s it. The End. Considering the game was so well designed and some of the boss fights were extravagant, the ending was very flat.

    Take another example, Shadow of the Colossus. It was consistent throughout. The only difference was the colossal size of the final boss, as well as how it was presented. Story elements drove the player to understand “this is it. this is the last fight. and it’s going to be epic.” When the player expects what is to come, it becomes a much better experience. At that point, you really don’t need to test the player’s abilities too much. You just run them through their usual paces with just a slight bit more challenge to keep the tension up, but that should be about it.

    I think last bosses should not be difficult, or too easy. It should be on par with the player’s other experiences. It just needs to be presented differently to make the player feel like it’s the most horrifying thing that game’s world has ever seen, but the player is experienced enough to deal with it without too much difficulty.

  2. Richard Knight

    You make a good point, and one I didn’t really elaborate much on, in that going too easy can end up flatlining the rhythm of the game. What seemed like the safe path for the developers to take may have actually cost them more than a ridiculous Harry vs. Foe battle.

    (sigh. I think I watched too much of those films on the latest plane ride.)

    The Shadow of the Colossus example is excellent – a fine example of incremental difficulty adjustment – but I think it also represents a luxury few game developers have, in that it’s a boss remade with the actual gameflow in mind.

    I would be *very* surprised if the design was the same as it was in preproduction, unless they had demo machines workingly early. (Maybe it was…did they use the ICO engine again?)

    The problem with incremental difficulty comes when a concept that is untested goes to production. When the game theory inevitably fails at some point, the end of the game can be treated like the tail of a lizard, and promptly be lopped off.

  3. Radiant

    If you think of it, each of the colossi were presented to you in a seemingly random fashion. There doesn’t seem to be any reason to present one colossus over another, and I do believe that was the original intent. At that point, the designers could order it in whatever order they felt was needed, and if a colossus had to be removed, it would be easy to do so. However, the last boss was definitely the last boss. There was no denying it. The way it was presented, it seemed they had full intention on how they wanted it done.
    If the game was designed with how the last boss was to be presented totally in mind, then it wouldn’t appear like the tail of a lizard (I like your witty analogy). The worst case would actually be a removal of the mid-section, and a potentially disruptive re-connect.

Comments are closed.