GameDevBlog

Speaking of Arrogant

I once tried to claim that I invented the term "Heisenbug".

Michael Vance points out that you can find the term in Unix man pages dating back to 1973. I was four at the time.
To that I say: phooey. Somebody has clearly doctored the historical record.
You know what else I invented? Using right-click to get a context sensitive menu. This was in the MindCraft title Gryphon Masters, well before Windows 95. Unfortunately, MindCraft went under before the title could ship.
(Actually, it was very fortunate, for me. Gryphon Masters was going to be a colossal failure on my part - I wanted to compete with Ultima VII even though we had a team of, like, six people, who were splitting their time with other projects as well.)

I like puzzles. Is that wrong?

Working on my chess game, I've been playing the chess puzzles on chessclub.com. The "TrainerBot", they call it. I'm enjoying it more than actual chess, I think. It gives me the same warm fuzzy feeling when I play Zelda - actually, even better, because I've gotten so good at Zelda-type puzzles that they no longer thrill me to solve.
Which makes me think about the whole puzzles thing. Ion Storm has this anti-puzzle manifesto -- "problems, not puzzles" -- which I think has taken maybe too strong a hold on the English-speaking game design community. (Talking to an ex-Looking Glass guy who now works at Oddworld, I mentioned that I liked Abe's - he said, "Yeah, but it was puzzley.")
I've brought this up before. So I repeat myself. Hey, it's hard to keep this blog filled sometimes.
There are a number of things to like about puzzles, and by puzzles here I'm talking about "single optimal path" level design rather than Tetris type games.
- they are pass / fail. You get an A or you get an F. If you're the sort of gamer who pride themselves on perfect play, games with a single optimal path reward you better than games where you're presented with obstacles that you can overcome in a variety of less-to-more effective ways.
- a moment of empathy with the game designer. A lot of game designers think this is uncool; when they see a contrived puzzle in their game, be it Prince of Persia or Beyond Good & Evil or Zelda, it ruins their suspension of disbelief - it reminds them that they're playing a game and not living a simulation or immersed in a story. I actually like it, the feeling that the designer is communicating with me through the language of puzzle - sometimes it's okay for games to be games and not stories or simulations. And sometimes a magician game designer can contrive the puzzle so well that you don't even notice.
- a single optimal path is usually clear. "What am I supposed to do?" the player asks.
- you can't cheese them. Creating the opportunity for emergent strategy is great but sometimes a strategy will emerge that makes all your hard work meaningless. Usually these strategies are found in test (if they aren't, then you can rest easy assuming that not too many of your paying customers will find them either) but there's always the risk that the strategy will show up in Gamefaqs (the pausing vigilante missions in GTA3 exploit, for example). (Of course, even when you're creating a linear one-intended-solution puzzle, you can still be bitten by emergent strategies. Any system can create emergencies. Heh-heh.)
I'm not sure how I got from chess puzzles to the beauty of optimal paths in videogames, but here I am.
There's a big fat problem with puzzles. Namely:
- they are, by definition, shelf level events: if your puzzle is even slightly challenging, it is going to stop some percentage of gamers in their tracks. (Ugh, I'm having one of those states where either way I spell "their" looks wrong. One problem with typepad is no spell checker...) These people will not recommend the game to their friends. They might even return it. This is why games like Zelda, BG&E, and Prince of Persia have really easy puzzles. For some reason, I still feel clever when I "solve" them, but in general I could hand the controller to a retarded monkey and they'd get through the game as well. Steve Meretzky's theory -- that puzzles should be hard because someone who's stumped can always look up the answer on the internet, but people who like them hard can't make them harder -- does not seem to be borne out by the marketplace.
All I'm saying is: do not rebuke yourself if your game design does not hold up to Ion Storm's standards. Different strokes, and all that.

Notes on Beyond Good & Evil

A lot of guys at work had been talking about this game, but I didn't really know what it was until I tried it.
It's Zelda.
A science-fantasy Zelda. It's all here: the landscapes with animals, children, and heroic figures (see the Feb 12 post); the hero in green; the roleplaying lite; the puzzles; the different kinds of gameplay (combat, puzzles, stealth, and vehicle). It's like suddenly discovering that there was a secret stealth Zelda released that only a select few know about, already at bargain prices.
And that's a good thing. It takes balls to try to make a game like Zelda. Beyond Good & Evil is big, there's a variety of gameplay, and there's a lot of terrain.
So, how did it fare on the wife test?
"This is the best game I've ever played," she said after I brought it home and she played it for four hours. She's not a gamer, mind you. The only game she's devoted more than a couple of hours to was Animal Crossing. But after BG&E she may have crossed into the ranks of casual gamer. This is only partly because there were no big frustrations in the early game -- one of her biggest frustrations was not realizing that the consoles where you save your game are also the consoles that you read messages on -- an example of functional fixedness? Mainly, I think, it's because you play a girl -- a girl who doesn't have huge breasts -- with a camera.
I read somewhere that Wal*Mart did a study and discovered 50% of woman shoppers have a roll of undeveloped film in their purses.
Games have had cameras before: Metal Gear Solid 2, Zelda: The Wind Waker, and Dark Cloud 2, for example. But never has it been such a focus. One of the first things you learn how to do is take pictures. This pleased my wife to no end. She was beaming. Her first reaction on seeing a new enemy - even a menacing boss - was always to take its picture.
Later, when the game started getting harder, she'd hand the controller to me to get her through the hard part. Which says to me: this game must be alienating just about everybody. Most teenage boys would get alienated by the femininity, by shooting photos instead of shooting guns, and the girls will be alienated by the boss fights and stealth levels.
The stealth gameplay, by the way, reaches a new level in Stealth Lite. Zeldas have frequently had stealth sections that were usually the weakest part of the game. (Except for the sailing.) Beyond Good & Evil did these sections very well: they were room-by-room; when you fail they weren't too punishing -- you only go one room back; the patrol paths were short, so you didn't have to wait a long time for your opportunity to move; you aren't forced to sneak past but the guards are hard to kill, so sneaking is optimal; and the guards shoot very visible beams of light out of their eyes, so it's easy to tell which way they're facing even when they're far away; when there's just one or two guards you can exploit an achilles heel for a stealth kill, but when there's more than that it becomes infeasible.
Speaking of sailing, instead of a sailboat you have a hovercraft, which is much more fun. Instead of being in open seas, you're in lakes and rivers and chasms -- terrain matters -- and there's a high density of stuff to interact with. There are a few racing levels which at first seem like extraneous bonus gameplay but actually become tied to the story. A hovercraft is a great choice from a coding point of view, also: easy to simulate.
This is yet another game that successfully breaks my "Don't have computer controlled allies" rule. One of the main staples in combat is using your ally to set up an attack, bouncing enemies into the air so you can slam them into walls or each other.
One of the best moments I'd seen-- and you may want to stop reading now because this is a spoiler -- was hanging on the side of a crate as a crane picks it up and carries it, and you, to another area of the room, without the guards noticing. Some simple elements -- guards, crates, cranes, and wall-hanging -- combined to make a very cinematic moment, a moment that was contrived but not prerendered. Magicianship at its best.
Eventually Cathy gave up and I kept playing. I was hooked. But now we come to the sad part of our story. I get to a door I cannot get past. Thinking I'm not meant to get past that door yet, I move on, solving other challenges, and saving over what will turn out to be my last valid savegame. Eventually I return to the door, the only area in the level I have not yet been. It is the second time I have to consult Gamefaqs after about twelve hours of play. Gamefaqs says one of the enemies I had beaten should have dropped a key. I realize that I slammed that enemy into the very door I need to get past, and somehow the enemy dropped the key on the other side.
Bugs like this happen all the time. We had one in Spider-Man, caught not by Activision testers but by Capcom testers when they localized it for Japan. In Spider-Man, you could web-yank a thug holding a key through a waterfall that the key was supposed to open. Same sort of thing. Of course, you could always restart the level, and you wouldn't get to a save point until you had. With my BG&E situation, where I'd saved over that last happy save game, I am doomed. I can either quit playing or restart from the beginning. This is the worst kind of bug to ship. It is worse than a crash, because you only lose up to your last save game with a crash. The only thing you can hope is that only a small fraction of your users will get hit. (Even if it's a one in a thousand thing, and your game sells a million copies, that's still one thousand players who now hate you. A meaningless statistic but fun to say.)
So that's where my notes end.

Oh!

And Noel Llopis is blogging, too: http://www.gamesfromwithin.com/
(I'm not going to link to him quite yet...he only has a couple articles up so far. But I laughed out loud when I saw what he wrote about redundant include guards. Ah, geek humor.)

Blog Alert

News from the world of game blogs:

Ben Cousins is doing a thing at www.bencousins.com that looks promising. I hope he keeps it up.

Tim Schafer is updating the Double Fine Action News on a regular basis these days. There are even comics now, which...aren't exactly funny, per se...but I can't stop reading them. I hope they keep that up.

Czar Chart

A new Manager In A Strange Land is up.

Before I submit my articles, I e-mail out at work for opinions and feedback. This article generated more controversy than any other I've written, including the one on crunching.

Michael Vance pointed out that he couldn't imagine doing things any other way, and that I was basically attacking a straw man.

Tom Henderson, on the other hand, wants--not a rock star, necessarily, but a guy, or a small tightly knit team, in charge of vision, that cannot be overruled, that can get into any and all dishes and say, "This is the way it's going to be." And I'm fine with that: you could have a slot on the czar chart whose responsibility is vision, or you could have an org chart that's two layers deep: the one guy in charge and everyone else. I see the czar chart not as a replacement for the org chart but the chart you make visible and emphasize; the org chart is for emergencies. Also, the guy at the top isn't going to get his way by saying, "Look, see this chart? I'm on top. That means I get to say." He has to earn it, the way Tomo Moriwaki does on our team, by demonstrating his talent and knowledge and earning our respect and trust. Tom was in the military, by the way...so he may have a bias toward clear lines of authority.

Unless the guy at the top owns the company, of course, in which case he can do whatever the hell he wants, although earning authority will still be a more effective management technique than, "Do it or you're fired." (And if the guy at the top owns a single-team company - there's no reason for an org chart. Everyone knows who's at the top.)

Compensation

Something I said in my Crunch article can be construed as being very insulting:
"A profit-sharing plan that ends up distributing no profits may get one game out the door but all your employees will quit before the second one is done. "
Am I saying that if you're working at, say, EA--where there's no profit sharing--or at Ubisoft on one of their less-money-making projects that you should quit, what are you doing, what are you thinking?
I did not mean to imply this.
I have seen projects where profit sharing was used as a carrot, and then when the product didn't sell, the employees started phoning it in and quitting. I've also seen projects where the product didn't sell and the employees kept right at it, strong as ever.
I think one of the differences between these two kinds of projects is in expectations - if you start talking royalties or bonuses the employees might start getting visions of Ferraris and when they see the actual results feel cheated. So if you have a royalty or bonus program, give somebody an idea of what they're in for. Don't let them think Ferrari when they're going to get Yugo. Or maybe just be like EA - don't give out royalties. Don't set up any false hopes at all.
I absolutely did not mean "nobody in their right mind would work on a project that didn't have a sweet bonus plan."

Notes on Everything or Nothing

"Make up your mind. Do you want everything or nothing?" the sales clerk at EB asked me.
Funny guy. Ha. Ha.
So, first off, have you seen the credits on this thing? The driving team had like 30 engineers on it? Jesus Christ!
Some mild spoilers follow.
This really is two games in one - two separate teams made two separate games and stitched them together. I wonder if the code is in separate modules which swap at runtime - I couldn't help but notice that vehicles that are destructible in the driving game are represented by undestructible proxy in the action game. Unlike Halo, where the driving uses the same interface as the shooting, the control schemes for the driving and shooting portions of EON seem to have been invented by two disparate design teams. Only rarely did this cause a problem for me. (What button is shoot, again?)
I like to sometimes make my wife play the beginning of a game to see how hardcore it is. "The wife test", a variant of Blizzard North's "Mom Test". This game completely fails the wife test: there were four points in the first three levels where she would have thrown down the controller and quit if I wasn't there to explain it to her. Four shelf level events.
I've been arguing with my acquaintances about tutorials lately. There's three schools of thought:
1.) have no tutorial. Your game should be so simple that no tutorial is needed. (Diablo)
2.) have a tutorial which explains the entire game up front. Hey, when you play chess, you learn all the rules, and then you play, right? (Splinter Cell, Call of Duty)
3.) spread tutorial events out throughout the game (first third, first half, whatever.) Mark Nau calls this "Black & Whiting it" because Black & White was a shining example of it. So was the new Prince of Persia.
EON went with a modified #2. They did a very James-Bondy-get-you-into-the-action quick segment...which was very cool...you plug the CD in and the game opens on an arms deal and then you're fighting. But it wasn't until I played a second time that I figured out how to get to the multiplayer modes. Then after that segment there was a big fat tutorial which teaches you how to play pretty much the rest of the game.
Assuming your game is so complex that a tutorial is in order...seems like most games these days are...option #3 is clearly preferable to option #2. Option #2 has several problems:
- You can hit a shelf level event early, before the game has really hooked the player (once the story has kicked in and they actually want to see what comes next, or before they've seen what makes the game cool) [We had an example of this in Die By The Sword, where there was a running jump in the tutorial that failed the Mom Test. (It was Mark Nau's mom. Hey, she likes Heroes of Might & Magic, we thought there was a chance...)]
- People forget. God, I've been watching focus tests lately, and people sometimes seem dumb as posts. If you dump a large amount of data on somebody, context free, they're going to lose a big chunk of it. With EON, both a friend of mine and I forgot about "Bond Sense" - I rediscovered it by accident a quarter of the way into the game - he played the whole thing (on easy mode) without ever knowing how it worked. Option #3 doesn't totally solve this problem, by the way. Prince of Persia took the extra step of telling you everything at least twice...
- It's not playing the game. It's learning the game. It's a pill you're making people swallow before you let them play. By concealing it as part of the game the way Zelda does, it makes learning fun. Almost.
More on Bond Sense: this is one of EON's biggest innovations. Like GTA3, you lock onto a target with one button and fire with another. Unlike GTA3, you can slow time to a crawl - so slow it's almost stopped - rotate the camera, and pick a target. This means you can crouch roll into a group of people, Bond Sense, kill one, Bond Sense, kill another, etcetera, all the while taking very little damage. It represents a new level of pandering to the "I want to be an awesome badass!" crowd, which is a huge market that includes me. It's a whole additional level above Max Payne, because you're not just slowing time down, you're stopping it, and it's free. The only thing that makes it beatable is all you can do in this mode is pick targets.
Here's the problem with Bond Sense: they ran out of buttons, and tied Bond Sense in with your inventory. When you switch weapons, you're in Bond Sense. This has two drawbacks: it's hard to enter Bond Sense without accidentally switching weapons, and it makes it hard to learn - no doubt the reason why my friend and I took so long to figure it out. (Side note: putting two functions on one button is a bad idea. Another example of this is SSX3 - a brilliant game, but there's three different buttons that have two functions each and if you're in the wrong context pressing the button can be catastrophic. Bond Sense isn't that much of a problem.) The Bond Sense problem could have been easily solved by removing one of Bond's hand-to-hand attacks - I could not tell the difference between these two punches - and putting Bond Sense on its own button.
They don't have many save checkpoints in the action sequences, where you need to kill dozens of guys without dying. The frustration this caused me was my own fault, as I insisted on playing on medium difficulty up to the last level, where I finally caved and switched to easy. Which was too easy.
I've got one last gripe, the same gripe I had with Call of Duty: it breaks George Broussard's One Game Design Rule, which is Things In A Game Should Behave The Way The Player Expects Them To. There are a couple boss fights where the bosses take either infinite or a very large number of bullets before dying. (And there's one boss fight where it arbitrarily decides to not let you use guns, call that gripe 1A, give me liberty or give me another game.) The worst thing is stuff in the environment that a real James Bond could easily jump, vault, or even *step* over which your jump-buttonless James Bond cannot. Call of Duty had the same issue, so maybe this is becoming one of the conventions of gaming. Max Payne 2 did it right; if there was a place Max couldn't go, it *looked* like he couldn't go there. Fences or other obstructions really sealed him off. There's one level in EON where a waist-high brick wall prevents you from proceeding, so you have to jump into a pit and walk through underground mines, until you finally come up on the other side and can escape. I shit you not.
So, all that said...this game is AWESOME. It is the epitome of the funhouse-ride action-driving hybrid. It raises the bar to new level. Actually, it's better than a funhouse ride, because although you are on rails, there are forks in the road. You can choose one path one time and another path the next time. This works well for a console game without save-anywhere; you actually do get to see most of the paths before you win.
The 007 moment is the biggest innovation in this game. It works on so many levels:
You may have asked yourself, "How can we encourage the player to act like the character?" The 007 moment does it with simple recognition. "You acted very much like 007 just now." It gives you points which give you silly rewards but I think that's immaterial - the main thing is the recognition. That was enough to get me to keep asking, "What would James Bond do?" In any other game, my tendency was to shoot the enemy instead of shooting the explosive barrel next to the enemy - that was the most direct way to victory. The 007 moment induces me to shoot the barrel. And if I didn't get recognized for every explosive barrel kill, that was okay, it's not like I particularly minded shooting the barrel. It did make a nice explosion. (An intermittent schedule of reinforcement.) The moment also gives the game replay value - I probably will never find time to go back and play it again, but I wish I could, because I want to find more of those moments. They're more interesting than the typical "secrets" of yesterday's games, since they're sometimes set pieces which involve big things blowing up.
(There was one time where I blew this helicopter up and then jumped over the exploding helicopter in a motorcycle...I'm not sure if that was a set piece or a meticulously timed event or just dumb luck..."emergent narrative"? Not likely in an EA game...)
Another great innovation is the way the game deals with cover. When you crouch behind a crate or back against a wall (you can tell these are gameplay elements because there are buttons for them...two bad that on the Xbox they're black and white) you are nearly immune to enemy fire. Maybe this has been in another game, I remember MGS2 had a wall mode that was similar...though I don't remember if you could shoot from that position or not...I'm sure someone will correct me in the comments section. Then the levels are laid out with lots of corners and crates - almost like an arcade shooter - and the result triggers my gamist pleasure centers and my "this is just like a Bond movie" pleasure centers and my "I'm an awesome badass" pleasure centers all at once.
Games aren't really about reflexes anymore, have you noticed that? I'm not sure how long it's been this way. Maybe a decade. Even games that have a counter button make it optional, give you such a large response time that anybody can do it, or both. This game isn't about your FPS skills, either, with its lock-on mode and its Bond Sense. For the most part the action portion is a strategy game. "I'll cover here, take these two guys out when they run out, run around and get this guy, this guy isn't coming out from that crate so I'll use a grenade, I'm low on ammo so I better fight hand-to-hand some (trading health for ammunition), wow I'm clever, etcetera."
More on the Bond Sense: one of the big problems with games in general is that you have a 90 degree field of view and (unless you have a kick-ass sound system) very little idea of where sounds are coming from. I've been playing a lot of Tron 2.0 lately and I've been noticing getting killed from behind a lot - and Tron 2.0 does a cool little monitor-shutting-off effect when you die, so you don't even get the courtesy of seeing where the killer was that shot you, like you do with Quake. The Bond Sense solves this problem - if you're getting shot, you can flip to Bond Sense, turn the camera until you see where the guy is, and then either shoot back (if you had a weapon in hand) or get the hell out of the way. But it's not totally free - time is still creeping forward, so you can't dawdle.
The set pieces--a particular example being when two tanks and a helicopter are about to kill you and M. says over your earpiece "you better shoot out the gas station, Bond" and you shoot the gas tank between them to take them all out--have very little to do with games but a lot to do with entertainment. The only interactivity in this scene is "press the A button to advance"--it's barely more interactive than Dragon's Lair. Still, these moments *are* cool. Although it's hard to argue with "if I wanted a movie I'd go see a movie" when there's a bunch of cinematic elements in your game it definitely adds value. (There's one prerendered "stunt" - triggered by pressing the A button, again - that really was clever and I wouldn't have been surprised to see it in a Bond movie.)
Although EON doesn't pass the Wife Test it does pass the Gamefaqs test. How many times did I need to consult Gamefaqs? Zero. So it seems like they must have done a fair amount of gameplay testing. But if that's the case, how come it doesn't also pass the Wife Test? I'm guessing that EA only did testing on medium-to-hard core gamers. (A problem we have at Activision, too: casual gamers aren't really interested in coming to a videogame focus test.) This can lead to what Costikyan calls Grognard Capture, where you're struggling so hard to please the hardcore you end up alienating other customers.
The rumor is this game got delayed for around a year. As I understand it, EA does a thing where they hire reviewers to review their game in secret, then they look at the average, and if it's too low, they slip the title. I can only speculate, but I think EA must have spent a fortune on this. I'm guessing over $20 million on the development alone, with at least a million just on the voice talent. (If only games were like movies and the budgets were known...) It's clear they're putting quality first, speed-of-delivery second, and cost third in their scheme of priorities, something they were only able to do by making a game that isn't tied to a specific movie...which makes me wonder just what happened with Agent Under Fire and Nightfire...not that I ever played them, but Gamerankings isn't very stoked about them.
[Oh hey. I didn't realize until just this moment that Everything Or Nothing is the first Bond game with the actual (current) Bond cast. Now I feel like a sucker. I thought I was buying the game because it got consistently good reviews, but maybe I'm just succumbing to celebrity hype. (Hey, I was going to buy Beyond Good & Evil but EB was out of stock!)]
Do movies ever slip? Do they ever enter the editing stage and people say, "Well, this isn't as good as we thought, let's rewrite this part and reshoot this part and so on. We'll put it out next year."

Anybody Want To Play Chess?

It's Sunday night and I'm in the Intermediate Lounge #1 at games.yahoo.com, under the name jdfristrom. I'm a mediocre player.

Gaming Markets Balkanized

Game designers like to talk about games as if there are universal truths about what makes one game better than another. For exapmle, I play all the top games on gamerankings and try to find generalities, ask what they have in common.
Yesterday I noticed this on Amazon.com, under Spider-Man for the PS2:

Customers who bought this item also bought these items:

* Scooby-Doo: Night of 100 Frights by THQ
* Finding Nemo by THQ
* Harry Potter & the Chamber of Secrets by Electronic Arts
* The Hulk by Vivendi Universal

This shouldn't have taken me by surprise but it did. Where's Grand Theft Auto? Where's Metal Gear Solid 2? Where's Jak & Daxter? Ratchet & Clank? How come none of the popular, crictically acclaimed games for the PS2 are on this list?
It's like there's a group of customers out there who only buy games based on licensed IP. (But shouldn't The Two Towers be on this list?)
I haven't played a single one of these games. And suddenly I realize we're developing a game for a market that we are not members of. Which may be a problem, because we're basing our design decisions on what we like, and what gamerankings.com likes.
This shouldn't have surprised me, because there's over twenty-five million PS2's in the United States, and yet we only sold three million or so copies. Even GTA3, the best selling game ever, only sold around five million. If GTA3 was truly the PS2 killer app, it would be more like twenty million. Presumably all these PS2 owners need games to play. They can't all be DDR servers, can they? (According to DDR sales, no.) As Mark Nau puts it, "All games are colossal failures."
A couple things may be going on here.
1) A whole lot of renting. People don't actually buy games anymore.
2) A balkanized market. People who like movie games don't like platformers. People who like platformers don't like RPG's. People who like RPG's don't like Grand Theft Auto. People who like Grand Theft Auto don't like games with E ratings. People who like games that get high scores on Gamerankings.com don't like games that get low scores. People who like DDR...you get the idea.
That has the ring of truth, doesn't it? Look at yourself. There's probably a whole group of game genres that you have no interest in, whether it's platforming or hardcore strategy or racing or what. And you're into games enough that you're reading this blog.
Typically, when a game is wildly succesful, we say, "It must have appealed to the casual market." Implying that this is a game that appeals to everybody. If it were a flavor of ice cream, it would be chocolate. But what is a casual console gamer, exactly? This is somebody who likes playing games enough to shell out $200 to do it. That doesn't seem casual to me. When a game is wildly succesful, there's probably another phenomenon at work. It may have captured a whole island in the Balkans. Maybe it even captured two islands. Maybe it discovered a new island. What it did not do is appeal to everybody.
There are no generalities. There are no universal truths. Mark Nau keeps telling me this but up until now I refused to listen.