Hello, my name is Benson Russell, and I'm an addi-- wait... wrong meeting! I'm a designer here at Naughty Dog in beautiful Santa Monica, and I was the primary combat designer for Uncharted 2: Among Thieves. I helped to implement the majority of the combat encounters across the game, as well as work with other designers and our programmers in developing our AI and weapons systems.
That being so, I wanted to write an article regarding how we approached combat encounter design for Uncharted 2. It turns out there was a lot of information to share, so I was given the opportunity to create a short series of articles on the subject. This is the first in that series, and I hope you find the information... informative.
Where We Came From
After completing Uncharted: Drake's Fortune (U1), we wanted to take a step back, look at what we accomplished, and then decide how we wanted to evolve it for our next game.
To set the scene, the combat in U1 didn't get a chance to fully come to light until the last year of its development cycle; hence we weren't able to make it as deep as we had wished. There were several key reasons as to why this had happened.
Firstly, U1 was our first foray onto the PlayStation 3 hardware. We decided to write a new engine from scratch rather than adapt our previous technology. This resulted in a lot of our efforts being aimed at just trying to make a solid-playing game and finish it on time.
Secondly, the style of combat wasn't determined until the last year of the project. We were still figuring out the direction we wanted to go (for example, we had a lock-on auto aiming style system up to this point). Lastly, our AI systems didn't fully come online until there were only about eight or nine months left on the project.
So given the time constraints, we really had to focus in on our combat to make sure the game would hit Naughty Dog's standards of polish by the time we shipped! We experimented with weapon tunings and AI strategies to see what was the most fun, as well as what was going to be doable in the remaining time. We came away with three core precepts to making fun encounters:
- Establish a front between the player and the enemies.
- Force the player to move around the encounter space.
- Spawn new waves of enemies to change the direction of the front.
Using these principles we were able to create a solid, polished and fun combat experience, but as the player progressed through the game, the combat started to feel repetitive and shallow. Here are some of the key lessons and issues we took away from making U1 and wanted to try to address.
Better Production Pipeline for Design Iteration
We crafted a lot of the encounter spaces first while we were waiting for various systems to come on-line. Hence we ended up having to shoehorn combat encounters into these spaces because of how far along they were.
As we started to iterate on these encounters, we were very limited in the kinds of changes we could make to the space, which in turn limited how well we could use our gameplay mechanics (such as shooting while traversing). We addressed this by having designers create a simplified blockmesh layout of an area to get the gameplay right. Then we would hand it off to our talented artists to make it look fantastic.
The Endless Assault of Enemies
We had several complaints about our enemies with regards to the number of waves players would have to fight. This came about due to technical limitations with our engine at the time. We were generally restrained to only having six to eight enemies active at any given time, so we decided to go with more intimate combat encounters with fewer enemies. When we needed to extend an encounter, we had to spawn additional reinforcement waves -- sometimes to the point of nausea.
Why Won't You Die!?
Another side effect of our limitation on enemy numbers was that we had to let them take more hits so they would last longer in combat. Normally this wouldn't be an issue, but due to a combination of several choices we made, it ended up causing frustration for some players.
First, the majority of our enemies weren't particularly well armored, wearing just shorts and T-shirts. This created an expectation that they should only take one or two shots to go down. Second, we didn't offer enough feedback with the hit reactions of the enemies, so it was unclear when a bullet actually scored a hit.
Last, our weapon feedback was very minimal due to development time constraints. We weren't able to put tracers on the player's weapons, our impact effects weren't as pronounced as they could have been, and weapon accuracy wasn't communicated well. This lead to players not realizing the true accuracy of the weapon they were using, and not realizing they were missing with a lot of their shots.
As an interesting side note to how critical these points were, we corrected these issues for Uncharted 2 and the average player thought the enemies took less damage than they did in U1 -- when in fact they actually took a little more.
The last key lesson we learned was that we needed to do a much better job of training the player on our mechanics, as well as crafting scenarios that can allow the player to creatively use those mechanics.
Where We Wanted To Go
So when it came to making Uncharted 2: Among Thieves (U2) we wanted to evolve the combat and help bring it into its own. We knew we had a solid foundation to work from, so it was a matter of taking the lessons we learned and playing to the strengths of our core mechanics better. Here's a high-level synopsis of the key goals we wanted to tackle.
Many of the goals we set for ourselves helped to reinforce this one idea, to give more tactical choice to the player during combat. In U1 the encounters that we enjoyed the most were in places where the player had more choice in how they wanted to tackle the encounter. We wanted to find ways to emphasize that sense of tactical freedom even more.
A key goal in support of Tactical Thinking was to exploit more gunplay while the player is traversing around the environment (which we call Traversal Gunplay). This includes shooting while hanging, while walking on a balance beam, or while jumping.
We did touch on this in U1, but there were limited times when it was really usable, not to mention Drake's move set didn't support shooting from all of his modes of travel. Also, our AI did not have the ability to traverse around the environment to follow the player.
Sorting these issues out became our first goals for U2. For the player we added the ability to shoot from just about any of Drake's traversal mechanics, allowing the player to engage in combat regardless of position or action.
For the AI we created special objects we could place in the world that would allow an AI to play an animation to be able to get from one place to another. We called these Traversal Action Packs (TAPs for short). Things like climbing or jumping down a vertical surface, using a ladder, jumping across a gap, using a balance beam, etc.
Using our enhanced Traversal Gunplay mechanics we were able to expand the play-space not just horizontally, but also vertically. Allowing the player to find places they can climb to gain an advantage proved to be a lot of fun and very engaging. Combat no longer had to be point and shoot, but also could become a sort of puzzle where the observant player could think of creative ways to use the environment to their advantage.
Another tactical element we didn't get the chance to explore in U1 was stealth gameplay. In U1 the player was almost always on the offensive with no real way to sneak around. For U2 we wanted to open up this tactical option to the player.
The trick was making sure we kept the pace up as traditional stealth gameplay tends to mean creeping along really slowly and looking at stealth meters. This wasn't the tempo that we were going for with Uncharted. "Action Stealth" was our motto, meaning our stealth had to be fast-paced and simple. We purposely simplified the design to be as player favoring as possible to prevent players from getting frustrated with stealth and abandoning it.
Our key reasons for including stealth gameplay were to utilize the tone and mood of stealth to support our narrative, and to give the player more tactical options on how they wanted to start a combat encounter. This would give the player the ability to get a leg up on a combat encounter by quietly eliminating a few enemies, and then getting into a good position to start the combat (usually with a bang).
In U1, once combat was started the AI would always know where their target was (usually the player). This didn't result in a lot of opportunities for tactical thinking, so for U2 we wanted to make the AI less perfect in this regard. They could now track you based on information gathered from their senses and their fellow AI instead of outright omniscience.
Sensory input includes their vision and hearing (weapon and explosion sound locations only) in addition to the location data they get from their fellow AI that have spotted their target. If they lose sight of their target, after a while they go into a hunting mode starting with the target's last known position.
From there they'll begin searching other positions radiating outward. They also have to acquire their targets in order to hone in on them -- meaning if they're coming into an area blind but they know someone's there (such as after hearing an explosion), they'll aim at the location of the disturbance rather than right at the source that made it.
This allows the player a chance to be crafty and outsmart the AI. For example, if the player's been spotted on one end of a long piece of low cover, he can move in cover to the opposite end and see the AI is still targeting where the player was originally. These might seem like minor points, but it's the little details like this that help to bring a bit more human quality when engaging them.
To give our combat more depth, as well as to keep it from getting stale over the course of the game, we wanted to introduce a greater variety of AI for U2. We touched on this in U1 with the shotgun and grenade launcher class of pirates, but for U2 we wanted to expand it further. In addition to shotgun, rocket launcher, and sniper classes, we added a fully armored class, a heavy class that wields a mini gun, riot shield class, and we separated our basic grunt into light and medium classes (the light and dark uniformed soldiers).
By mixing and matching various combinations of these classes and weapons, it allowed us to create combat encounters that would force the player to make on-the-fly tactical decisions. This also gave us the means to create a stronger overall difficulty ramp across the entire game. By training and introducing tougher and tougher enemies and then tougher combinations of enemies based on the encounter space, we could keep the combat exciting, giving the player something different to experience with each space.
Active Cinematic Experience
Lastly we wanted to tightly integrate our combat encounters with our storytelling. Creating the "Active Cinematic Experience" was a central goal in the development of Uncharted 2: put simply, we wanted to tell as much of our story as possible in-game rather than resorting to cinematics. To facilitate this, the tone and mood of the combat had to strongly support where the player is in the story arc.
Creating a mood and feeling through our encounters and our environment was a key point we wanted to drive home. Using lighting, time of day, color palette, weather effects, tight vs. open space, and ambient sound in conjunction with the style of a combat encounter helped to strengthen the emotional attachment of the player to the characters and what they were going through. A great example of all of these things in action is the progression Drake takes through the war torn streets of Nepal:
- Near the beginning of the area Drake meets up with Chloe. It's bright and sunny, and this helps give the feeling of kicking off a great adventure. A few challenging combat encounters occur, but they're nothing too overbearing that they can't handle together working cooperatively.
This helps to support the feeling of getting to know and care about Chloe. At one point when they reach the top of the hotel, storm clouds can be seen off in the distance foreshadowing darker times ahead.
- This eventually leads to a very high-energy combat sequence with the helicopter chasing them through the collapsing building and across the rooftops.
We build up tension with an overpowering force (the helicopter) that is relentless in hunting Drake down, forcing him to find shelter. We build upon this to increase the payoff moment when Drake can finally confront the helicopter and take it down.
- This high-energy climax then allows us to bring the pace back down as they meet up with Jeff and Elena to make their way to the temple. It's now becoming a little bit overcast, tension is building up between Chloe and Drake with the introduction of Elena and Jeff, and the spaces are becoming more cramped and closed in (going from the open rooftops to alleyways).
There's light combat here to just keep up the sense of being hunted, but allowing emotional space to play up the drama brewing between the characters.
- The combat encounter uphill against the machinegun turret provides a short spike of intensity, and being an uphill fight helps continue along the lines of being pinned down and closed in upon.
- Once Drake and Chloe get to the map room in the temple, they have to fight their way out as Lazarevic's men have entered the scene. What's happened to Jeff and Elena is unknown as the soldiers would have had to make their way past Jeff and Elena at the temple entrance. Deep underground, the only way out is through the incoming soldiers, giving the sense of being overrun and oppressed.
- The combat is designed to feel like having to dig your way out of the temple as the soldiers take up strong positions in your path. We use this to build on the tension that Drake really wants to get back to Jeff and Elena.
- Once they exit the temple, the rain has arrived. It's now very grey and overcast, adding to the desperation of seeing Elena and Jeff pinned down in the temple square. Snipers peering through the windows from the high ground, along with a fairly overwhelming force help support this. This also foreshadows the coming conflict between Chloe and Drake regarding what to do with the injured cameraman.
- The fight of escorting Jeff through the alleyways continues to build upon the tension between Chloe, Drake, and Elena. Drake is also very limited in his ability to move while escorting Jeff, giving a sense of helplessness and reinforcing the tension. The combat puts pressure on the player around every corner, building to an overwhelming crescendo of rocket launchers and trucks with turrets.
- The climax of this whole sequence is when Chloe is forced to turn on Drake and Elena to save herself. This is also setting up the first face to face meeting with Lazarevic. This creates a feeling of complete desperation -- that this could be the end.
These are only a few examples from the game of how we tried to use combat to help support the arc of the story. It's not about just putting in combat for its own sake, but to also try and tailor combat scenarios to fit in with the mood and moment of the game.
It's a balancing act of making sure we have elements that let the player use what they've learned, introducing new elements to challenge the player, introducing new weapons and abilities to add to the player's toolset, and making sure it's all done in a way that strongly supports the story.
Alrighty! That concludes this first section in this series. The remaining sections will cover:
- Our development process at Naughty Dog in how we create an encounter from top to bottom
- The technical side of our AI and how they work, along with the tools we use to control and craft our experiences
- Techniques and tips that we've used
- The lessons we've learned
So until next time...