by Steve Bowler on March 4, 2011 · 4 comments

in design,mechanic

Last month, while I was crunching my butt off a friend alerted me that Tim Rogers wrote a lengthy retrospective on the Quick Time Event in Game Developer Magazine, and he was praising something I contributed directly to the design of: the Standoff mode in Stranglehold.

The Standoff in Stranglehold is one of the things about the game I’m most proud of (besides maybe some of my motion capture performances*).  From a design perspective, it was one of the most difficult aspects of the game we tackled.  For at least half of the game, we knew we had to take on the Woo “standoff” signature moment seen in all of his Hong Kong cop movies, but we had a lot of different (and often very bad) ideas about how to attempt it.  All we knew was:

  • It had to be cinematic in nature.
  • It should result in some sort of amazing gun battle that was different from the game’s typical gun battles.

We never seriously considered the whole “Shenmue” or “God of War” style QTE that Rogers discusses in depth in his article.  I think the whole team was pretty against that idea; that it really didn’t mesh well for Stranglehold’s style.  At one point, Ty Primosch, one of the cinematics guys prototyped out a cool looking Standoff scenario action scene (seen below), but we had to punt on it because we realized that the only way to make something like that work was to do it God of War style, and as I mentioned, we wanted to shy away from the typical QTE.

One of the ideas that I do still remember being kicked around (I don’t remember many of them) was the concept of maybe having two different kinds of timer wheels, and each wheel had different pie shapes on them, one for “action” and one that was “null.”  You would have one wheel that was for shooting, and one that was for dodging.  When your desired wheel’s pie wedge came across the active arrow over the wheel, you could push the appropriate button and you’d do that action.  It didn’t last long before it was axed.  I don’t even think we prototyped it.  We would keep bouncing ideas like this off of each other for weeks (was it months?)  until one day we all just forced ourselves to sit in the conference room and hash out why one worked or didn’t work.

The meeting went south pretty fast, and every idea that was thrown out was hacked and slashed to pieces and just didn’t cut it.  I sort of mentally checked out of that meeting, because for the record, I was not the biggest fan of the Standoff mode.  Simply put, I didn’t want it to be a gameplay mechanic.  In nearly every film where Woo was using his Standoff, it served to advance the (super ultra thin) plot.  Maybe two enemies would realize they’re not enemies.  Maybe one would try and shoot the other but realize he is out of ammo and the other one would smile and back away (giving away both characters’ motives).  Maybe both would realize they’re really looking into a mirror reflection of themselves and realizing they see their enemy within themselves.  They rarely wind up actually shooting at each other.  And yet there we were, trying to find a mechanic that would let us express two people shooting at each other at point blank range and making it feel fun and interesting.

So instead of only offering negative criticism about everyone else’s ideas at this meeting (because that’s all I had to offer at the time, along with everyone else at the meeting), I decided instead to try and look at what we were already doing with the game, and how we could apply it to the Standoff mode in some way.  We knew we wanted you to stand in place, for the most part, and decide how you would shoot at enemy AIs.  And for the hell of it, I just asked myself the question “why can’t we just use the sticks the way the player expects to use them?”

We were talking a lot about making the Standoff a puzzle, but all of the suggestions made them overtly a puzzle.  I guessed that we could make it a covert puzzle, by making it not look like a puzzle at all and disguising it with realtime controls the player already knew how to use.

The idea was that the player would use the left stick to dodge, and the right stick to aim the reticule.  This is ostensibly how the player uses the sticks in game.  Left stick moves, and right stick aims.  Kudos to Tim for picking up on this in his piece.  This was pretty much exactly how I pitched it to the Lead and Sr. Designers on the design team after the meeting was over.  Why change what the player already knows?  Let’s just change how we present that and change how the player can use it.

The other aspect of the Standoff “puzzle” was that it would be created, or driven by the player.  It’s just that the player wouldn’t realize it.  Originally I just threw out that the AI would fire bullets in a “pattern,” but what wound up happening is that we realized that all the AI would need was a set timer, which stated something like “I shoot bullets on the 2nd, 4th, 5th, and 8th seconds of my standoff.”  That’s all he does.  If the player doesn’t succeed in killing him before the 2nd second, he spawns a bullet that travels at wherever the player is currently standing, which is almost always at the “neutral” position.  This forces the player to dodge, or face the consequences of eating a bullet.

By being forced to dodge, the player is no longer just in an “aiming” mindset.  Now they have to rub their bellies while they pat their heads, so to speak.  The challenge of dodging left while trying to meticulously aim a slow reticule has the delicious side effect of dividing their attention span and making aiming even harder.  Now, if they take too long to kill the AI, he spawns another bullet, and the player is forced to dodge yet again.  This goes on until the player either succeeds in killing the AI before his timer runs out, or the timer expires, and the AI is left alive to shoot at the player when he exits the Standoff mode.


Selling it With Reference

It turns out that the idea, while well received, wasn’t a slam dunk.  To help convince the powers that be that we needed to prototype this, I wound up digging up some references in other games to similar dodge mechanics, specifically ones where you felt like you had a high level of control of your dodge, the dodging changed the core mechanic of the game everyone was used to, but it didn’t feel disconnected from the game.

The first example I gave was Steve from Tekken 4.  Steve is one of the most interesting Tekken characters because he doesn’t kick; he only punches (I take that back I think he has like three very special case kicks?).  The buttons which would normally control a character’s kicks become a sway and weave move on Steve.  Skilled players can time the sway and weave to dodge incoming attacks, or use them to start even more powerful attacks.  But what I really liked about him was that he could dodge (with style), left and right when the player prompted it.  This really helped sell the idea that you could use a dodge move even when rooted in place to avoid incoming attacks, and it was not only fun but also could be pulled off with some style and flair.

The Steve example (along with some Fight Night Round 3 footage) was convincing enough to give it a try, and we got the green light to prototype it.


Environmental Kills

The Standoff mode itself just in its raw form was pretty fun, but took a surprising turn for the entertaining when the level designers on the team batted it out of the park by placing all of the environmental hazards around the AIs, to make it so super thrilling to kill them in slow motion.  Propane tanks, signs, the habitually overused air-conditioner unit dropped on their head.  I think after awhile it became a challenge to out-do each other.

My personal favorite, though, was the 20 man Standoff that appears in Wong’s Estate just before the final battle.  It is insanely entertaining in its absurdity.  It’s the ultimate extension of the “puzzle” idea, in that each AI you need to shoot has less and less time on camera before the camera switches to the next guy, and if you don’t kill him, he’s there waiting to shoot you at the end of the Standoff.  If you leave 5-6 guys by the end, and you’re at low health (from having 20 people shooting at you in slow-motion), you could be in bad shape.


No Need for a Tutorial

One of the happier accidents that came out of Standoff mode was that we forgot we had put it in the game.  No, seriously.  We were running our very first official playtests on the game one night by having some focus testers come in and give the game a spin.  We just wanted them to play through the first level or two and watch them play to see what their reactions were to all of the different elements of the game.

We were also looking for anything else that might come up when you have someone play a game for the first time.  It sounds ambiguous, but that’s what focus testing can be sometimes.  Game teams know where all of the bad guys spawn in.  We know where we’re supposed to go at every point of the game.  When you bring a fresh player in, who doesn’t know where any of these things are or how they’re supposed to unfold when they play the game, you get a really fresh perspective on what you forgot to do as a designer/developer to help them have the least frustrating (and hopefully very enjoyable) experience with your title.

For instance, sometimes they (unknowingly) point out that you completely forgot that you put a Standoff moment in the game.  Everyone on the team let out a collective “oh noooooo” groan when a handful of playtesters all reached it at almost the exact same point.  We never made a tutorial for it yet, because we hadn’t really planned for it to come up in this playtest; we had forgotten to remove it from the level.  But there it was, rearing its ugly head in front of us, with every player in the room (eventually) having to get through it.

Just when we were trying to figure out what to do (do we just say nothing and let them play it? Do we ask people to pause so we can quick explain it to them?) the sequences started and they began getting shot at.  And the players began dodging bullets and shooting back.  They even realized that they could aim for the gleaming objects in the background for even more explosively entertaining killshots.  There were audible cheers coming from the testers.

The mode, we were shockingly surprised to find, didn’t require a tutorial.  It was intuitively playable by default.

We all collectively breathed a loud sigh of relief, and shared some pretty incredulous “can you believe that?!?!?” type of jawdropped stares.


From Initial Idea to Final Completion

After that playtest, we were pretty convinced that the mode was a hit, and we could leave it as-is while we finished the rest of the game.  Unfortunately, we seemed to have forgotten that it was just sort of stubbed in and prototyped, and not really cleaned up and finished.

The Standoff in its prototype phase worked exactly the same as what you saw when we shipped, only it was much simpler.  We didn’t have a lot of time to prototype it, so I remember we just had the animators give us one set of aimgrids, copied and tweaked to be two additional dodge poses.  Basically, we used a single pistol stance of Tequila standing with just one gun out (his right one), and when he dodged left he’d just lean forward, still aiming with his right gun, and when he dodged right he’d just lean back, still aiming with the right gun.

Functionally, that was great.  It totally worked.

Unfortunately, it didn’t look anything like a John Woo movie.

I remember playing it again (we’d usually skip the Standoffs if we weren’t testing them specifically when playing/testing levels or mechanics in editor), intentionally not skipping it this time, maybe a week or two before art lock, and had that sinking feeling when I realized that “oh god we’re about to ship a game with prototype stubbed in animation for the signature mode.”  I think I grabbed Brian, the EP and Team Lead, and Neill (the Lead Designer), Patrick (the Sr. Designer on the team) and A-Rod (the head gameplay programmer who built the Standoff in the first place), and begged them to let me fix/finish it.

I knew what I had to do to fix it (namely introducing 3 new proper looking aimgrids), but the problem was that it required tech we didn’t have which would have to be created.  The problem, was that I needed Tequila to start in a dual-pistol stance, and if he dodged left, he needed to lean to his left, and only use his left pistol, and if he dodged right, to lean to his right and only use his right pistol.

We had code for the leans; that just used existing AnimTree/Finite State Machine tech.  We already had “just shoot with your right pistol” for when Tequila or an AI only has a single pistol to shoot at you, and we even had “shoot with two pistols” code that allowed for alternating between your left and right pistols, and even for those two to work in conjunciton with each other (say you picked up a 2nd pistol while shooting with just the one).  But what we didn’t have, was code that knew that you only wanted to shoot with your left pistol, or at this moment wanted to shoot with alternating left and right pistols, or in another moment only wanted to shoot with your right pistol.  And we sorta had to have this feature.  The mode just wasn’t shippable without it.


Dead Akimbo to the Rescue

What saved it, code-wise anyway, was a failed special “Tequila Bomb” that we never got working, which was just simply called “Akimbo.”  We were trying to emulate moments in the Woo films where the hero just starts shooting everything in the room with crap-tons of style.  Maybe he’s aiming at guys on the left side of the room with his right gun, and vice versa.  Maybe he’s running through he room with his arms spread-eagle.  We wanted it to be dynamic (it had to be; there was no way you’d ever encounter the exact same position of enemy AI twice), in that the player just sort of had to aim Tequila in the general direction of the bad guys, and Tequila’s arms would intelligently just know where to aim, and it would look super sexy and sleek.

But we could never really make it work.  The arms would work for awhile, and then BAM they’d freak out and go all gangly and go through each other.  The right arm would be aiming high left, and the left arm would be aiming low right, and suddenly they’d decide to switch heights (because they’d both re-acquire new bad guys), and just go straight through each other.  Or worse, an arm might try and go through Tequila’s body because that was the shortest distance between two targets.  The mode was pretty much a disaster no matter how we tried to help his arms.


What the mode did have was the ability for the guns to fire intelligently.  Maybe at one moment only the left gun had a target.  A moment later, the right gun had a target.  And another moment later both guns had a target.  It had the code we needed for Standoff already built.  While it still wasn’t a trivial thing to merge with Standoff, it was still in a better place than having to write it all from scratch.  Failed Akimbo saved Standoff.  It sort of became an organ donor on its deathbed so another mode could live.


Wrapping it Up

In the end I think Standoff turned out pretty great.  Everyone involved was pretty happy with how it played out.  Rogers’ article isn’t the first I’ve heard praise for it; whenever people found out I worked on Stranglehold, the Standoffs were usually the next topic of conversation.  People just seem to like it.  It was intuitive, and I think it captured what a Woo movie feels like while still giving the player the ability to control what’s going on in the scene.  I’m pretty proud that something I originally thought couldn’t possibly be done well in a video-game was executed in a way that became one of the highlights of the game.  Most of all, though, I like that it doesn’t come off feeling like a Quick Time Event.  But don’t take my word for it.

This is as exciting as QTEs can possibly get: the action fits story context, character context, and game control context, and the payoff is visceral and instant. Much as Half-Life phased out the cutscene by making the narrative “happen” in the world as the player plays, Stranglehold shows that QTEs can be part of a game and not be sudden, intrusive, demanding situations. — Tim Rogers.

I think in the end we succeeded with these modes (especially Standoff) so well because the team approached these not as a QTE, but really just as another part of the game.  Apparently it shows.


* when you shoot someone in Stranglehold, especially in the nuts, there’s about an 8-out-of-10 chance you’re seeing one of my hit reacts.  I do my own stunts.