The genius of Pac Man

Below is an article that I wrote a while back, that briefly examines Pac Man as a role-model for game design perfection. I've always liked examining older games, which weren't nearly as reliant on graphics for their success, with the idea of learning from their gameplay design. I may write a few similar articles in the future, looking at games like Asteroids and other classic games. In the early 80's I worked in several arcades and was quite the game player, even winning several tournaments here in Dallas. A benefit of working in an arcade was being able to stay after hours and play endlessly for free.

Anyway, here's that article...

~~~~ begin ~~~~

The genius of Pac Man

It's the best selling arcade game worldwide, with over 250,000 machines sold. When it came out, I played it occasionally, but for me it wasn't in the league of Space Invaders, Missile Command, Scramble, Asteroids, Defender and other high-action arcade games I preferred during that time. George, my partner at 3D Realms, though, loved it and eventually memorized patterns that allowed him to play endlessly on a quarter. (These patterns can be found in a book George and I co-wrote in 1982, "Shootout: Zap the Video Games.")

The years passed and other arcade games faded away, yet I still found Pac Man in arcades and restaurants, and never being a fighting games fan (the only other games that seem to be around nowadays), I found myself playing more and more of Pac Man and eventually appreciating the game's genius.

I believe there are only a few software games ever created having perfect and pure gameplay. Tetris and Pac Man are two examples, but Pac Man is the better example because it's a more complex game, and therefore the greater feat. Every aspect of Pac Man has gameplay repercussions and provides the player with decisions that involve trade-offs. Also, the game is extremely quick to learn, but nearly impossible to master, and has just one ultra-intuitive control, a single joystick.

Just the act of eating dots is perfectly designed, because eating them slows Pac Man down, slightly slower than the pursuing ghosts. This makes the act of eating dots a trade-off decision for the player because to eat dots – to accomplish the game's primary goal – is to put the player in greater danger. While staying in lanes where there are no dots allows the player to move slightly faster than the ghosts, but does not accomplish the game's goal. A lesser designed game would have not slowed down Pac Mac just the right amount when eating dots.

When Pac Man eats one of the four power dots, it turns the tables on the ghost, not only taking away their threat to Pac Man, but making them valuable to eat. Again, there's just total perfection in the way this is handled. First, the pursuit of these harmless ghosts takes away from the player’s primary goal of eating dots, but it's a perfect trade-off decision for the player (greed for points vs eating dots and moving to the next maze). Plus, with each of the four ghosts eaten, the points double for the next ghost the player eats – but the trade-off is that going for the next harmless ghost puts the closer to the danger of that ghost turning back to the normal kind, and killing Pac Man.

Finally, the appearance of bonus fruit, which rises as mazes are cleared, is just another simple trade-off for players because those fruit are not required eating. But they are so compelling that players often take dangerous risks to eat them, and end up dying. And dying in this way the player knows full and well it was their own greedy fault.

There are other perfect design issues with Pac Man, such as the design of the maze itself, which includes dangerous areas where it's easy for Pac Man to get trapped, the tunnel that leads through the screen, and the different personalities (different AI) for the four ghosts.

A perfect game that we can still teach us, over 18 years later.

~~~~ end ~~~~

I think the key idea to take away from Pac Man is the principle of trade-off decisions. Every key decision you make has both a positive and negative side. With simple games it's easy to design with this principle in mind. With the more complex games of today this principle too often gets back-burner treatment. For example, we've all played a shooter in which the later weapons you collect are superior in every respect to the earlier weapons. The result is that the player ignores the earlier weapons. In Duke Nukem 3D, we tried to rectify this by having the later weapons all have both positive and negative attributes, and we succeeded to a worthwhile degree.

One final thought, as I will often talk about "principles" in this blog. A principle is a basic truth, a proposition that is reliable in most cases. I will rarely talk about game design "rules" because the word "rule" suggests a principle that's written in stone, and without exception. I prefer the word "principle" because it allows for some flexibility.

Comments

[Charybdis]

This is quite interesting:
http://jongy.tripod.com/GhostPsychology.html

[JP]

"One final thought, as I will often talk about "principles" in this blog. A principle is a basic truth, a proposition that is reliable in most cases. I will rarely talk about game design "rules" because the word "rule" suggests a principle that's written in stone, and without exception. I prefer the word "principle" because it allows for some flexibility."

Bully for you! I am quite sick of hearing Noah Falstein's "400 Game Design Rules" cited as some kind of gospel truth. A collection of looser, expandable, open-to-revision/clarification principles that speaks in a consistent lexicography, like the kind ION Storm seems to use, seems like a better fit.

I am quite adamant about developing more empirical ways of talking about game design and the theory behind it but laying down absolutes - rules - then having to retreat from them immediately with convoluted "trumping" meta-rules, seems as inflexible as it is counter-productive.

And yes, Pac-Man is a good little piece of game design with some nice lessons. I'm partial to Ms. Pac-Man, though, with the "turbo speed" dipswitch enabled... you are much faster than the ghosts and the game moves at blistering speed but still retains an incredible balance.

[steve]

"I think the key idea to take away from Pac Man is the principle of trade-off decisions. Every key decision you make has both a positive and negative side. With simple games it's easy to design with this principle in mind. With the more complex games of today this principle too often gets back-burner treatment."

This is the fundamental appeal of Civilization, a game that's hardly simple. It's a series of interesting choices, each with their own positive and negative impacts (just changing governments causes a cause/effect). SimCity is another game that thrives on this dynamic. Perhaps designers of first-person shooters just need to consider genres outside their own for inspiration? It'd be good for the genre.

To some extent, I think Halo adds this dimension by limiting the player to two weapons. You're always forced to make interesting choices for what may or may not happen ahead. Each has its upside and downside; take the limited ammo heavy weapon if you think you're in for a big battle, for example, but if you have no more ammo, it's useless.

You could argue that you don't have enough intelligence to make a proper decision, because you don't know what sort of battle you're about to face. That's true, but that's also part of battles, one area of "realism" that works as a gameplay dynamic. (And the fact you could pick up alien weapons helped keep the game from being impossibly hard if you made the wrong weapons choices.)

[Gestalt]

Steve - "To some extent, I think Halo adds this dimension by limiting the player to two weapons. You're always forced to make interesting choices for what may or may not happen ahead"

Personally I just found that annoying. As you said yourself, you often don't know what kind of weapon you're going to need at any particular point in the game, so you end up making poor decisions through no fault of your own and ending up saddled with a gun that's no use.

For example, at one point you go into a network of underground tunnels, probably expecting to need some good close range weapons. Then you find yourself in a huge cave where you really need a sniper rifle. Do you waste five minutes running back to where you dropped the sniper rifle outside, or try and muddle through as best you can with your assault rifle? That's *not* an interesting choice. :) In fact, limiting me to just two weapons isn't giving me more choices, it's limiting my choices of how to approach the challenges I meet along the way to whatever is possible with the weapons I happen to have with me at the time.

[Charybdis]

"As you said yourself, you often don't know what kind of weapon you're going to need at any particular point in the game, so you end up making poor decisions through no fault of your own and ending up saddled with a gun that's no use."

That's the old problem with 'save gems' and their ilk too. Horrible as it may be, tactically, I got through Halo by grabbing more or less whatever I saw lying on the ground and occasionally thinking "Ah, that's handy." When a weapon ran out of ammo, you'd promptly find it on the floor. That kind of dynamic has always worked much better for me in games like, say, X-Com, with options like stocking up on rocket launchers capable of smashing straight through walls, versus sniper rifles, or of course Rainbow Six when you could see the level plan and knew what your objectives were before so much as getting onto the bus. The catch of not knowing stops it becoming so much of an interesting choice as a potentially interesting gamble, especially when the gameworld suddenly starts playing silly buggers on you (for instance, Max Payne's bosses being able to soak up more bullets than the Atlantic Ocean)

[Brad Renfro]

"As you said yourself, you often don't know what kind of weapon you're going to need at any particular point in the game, so you end up making poor decisions through no fault of your own and ending up saddled with a gun that's no use."

The idea is that you adapt to your choices. You play differently based on what weapons you chose or were forced upon you. This would be less likely to happen with a persistent stockpiled arsenal. Also, generally you should know which weapons are most effective from the first couple of minutes in the level. And like Charybdis mentioned, you are meant to be constantly conserving and scrounging for weapons with more ammo in them (especially the plasma weapons).

"I am quite sick of hearing Noah Falstein's "400 Game Design Rules" cited as some kind of gospel truth"

I've always wondered what hardcore designers thought of these rules. I've often found that people have their own general set of "rules" which ultimately leads to different personal styles. Having a core set of general shared rules makes game design sound like a paint by the numbers. But I bet there are similar "rule" lists in, say, painting, architecture, music, etc.

[Joe McGinn]

I couldn't agree more about risk/reward balance and decisions leading to fun gameplay.

I wonder if this is why games like Prince of Persia don't resonate past hardcore gamers ... it lacks this basic dynamic. The platforming is pretty cool, but empty of any meaningful decision making.

[Shahar Eldar]

Noh Falstein's 400 project while a good idea in theory (collect known principles which apply to games) dosnt attempt any kind of serious analysis of those principles once they are collected.

Once these principles are collected they should be properly analyzed for which areas of game design they impact, and how they relate to each other (with more depth than "what it trumps" and "what trumps it")

For Example - the rule presented in the Pac Man analysis would be probably formulated as something like this. "Each choice in games should have both an advantage and disadvantage." which would then be accompanied by the various "situation used" "trumped by" and so on... However I would take it further
analyzing this rule I would derive various suggestios which could then be confirmed/denied by testing and or further research. One of these (and I'm coming up with this as I type, so give me some leeway here) is that the player does not feel he has made a significant choice if there is only one advantageous resolution.

I allso think principles are important but it's allso important to base those principles on something objective, a way to really analyze gameplay from a point of view where everyone could understand it equally. I have come up with various "elements" which make up gameplay and help in understanding the parts of it which I think the principles should then adapt to.
These elements are:
Field, Figures, Rules, Goals, Feedback, Choices, and Chalanges.

where basically these are what I see as the main building blocks of any game, if anyone is interested, I explain each of these elements in my site far better than I can without creating an enormous post here
http://www.silverlime.ca/users/pizzaman/index.php?page=GameTheory
tell me what you think?

[Charybdis]

"I couldn't agree more about risk/reward balance and decisions leading to fun gameplay."

Was thinking about this in relation to Megaman the other day. You didn't know what you'd find on each level, but you could work out that having - say - Fireman's weapon would be useful when dealing with Woodman, and thus plan the order that you hit the bosses to gain the advantage. Freedom Force had something similar, but made itself a bit too open - you could work out precisely what you needed and then the villains couldn't even hope to compete. Some sort of balancing is clearly required, but at least that adds a strategic edge beyond the individual levels.

[Scott Miller]

Shadar, you will not find me analyzing my principles or whatnot with academic integrity or precision. My stuff should be viewed as fluffy opinion -- take it or leave it. ;-)

BTW, I've done a similar breakdown of games, with my slant being, What qualities are important to ensure the success of a game? I've worked on this for some 18+ months, and likely this year I'll have a substantial blog entry covering this topic.

I've also developed a scoring system that can fairly accurately determine the success potential of a game. But, unfortunately, this system is rather subjective and therefore only tuned to my particular judgment calls. Still, it might be of interest to readers and so I'll post it along the way.

[Greg Findlay]

I'm looking forward to those two blogs Scott :).

On risks and rewards:
I think this is one of the reasons why linear stories seem so disconnected from the game. All of the risks the player takes don't effect the story except to progress through it. Once the user is integrated into making decisions in the story, like say take the short path through the dark scary woods or the long path through the fields with hopping bunnies, they have a stronger tie to the story and makes them feel like they are in control. The plot can still be linear, but the user has the option to take risks. This could also be a way to avoid having the user set the difficulty at the beginning of the game as they could choose along the way to take more risks if the game is too easy for them (and is well complemented by an ADD system).

Ideally the number of decisions the user can make is kept very minimal, not only because it becomes easier to test but because it adds tension to the story. I find if you have too many decisions you get distracted from the story because it takes longer to decide what to do and as a user you start assessing too much.

This is all of course, easier said then done and it by no means replaces linear story games, but if your goal is to make an immersive story I think that would take you one step closer. Another tool for the tool belt ;).

[Shahar Eldar]

those sound like very interesting reads, by the way, just as a side note, I'm a sophmore in a Digital Art bachelor's program in Otis school of art and design, and for one of my classes right now I'm required to find somone in my chosen interest and conduct an interview, I havnt yet decided on the questions, but I was wondering (Scott) if you'd be interested in giving me this oppertunity? this is connected with a rather large research paper about history/lexicon/hirarchy of the industry

[Scott Miller]

Shadar, I've been turning down a truck load of interview requests lately, but for you I'll make an exception, as long as you don't barrage me with questions, and as long as the interview doesn't become public. Please send directly to my email.

[Jake]

This might be a stupid question to bring up, but how do PC adventure games like Grim Fandango, Leisure Suit Larry, etc etc (Myst?) fit into the idea of positive and negative tradeoffs in gameplay decisions? I guess they don't?

[Shahar Eldar]

adventure games specifically lucasarts adventure present almost no risk, the game is baseed entirely on exploration, you are given a world/situations and the chance to explore it's every nook and cranny (from dialogue trees, environments to item combinations) in this way adventure games are more like interactive comic books, where you go through scenes at your own pace and explore content not necesarily to do with the main plot, then they are like other games.

[Tim]

I know this might not be much of an important topic im trying to point out, but still... Im an independant game developer, Level Designer in particular, and every time I do anything, i keep on coming back to half-life. Particles, structure, and levels interest me almost too much, its dangerous. Yet after you look and study the levels very seriously, you might call me crazy, but after long plannin, i mapped out the whole game. I have it all in my huge notebooks. (4 and a half of them) Im being detailed, and they dont have anything special, yet whats making it so great, and so popular. I dount anyone studied it as close as i did. If u look very carefully, their colors often dont match togather, and gameplay is basically kill or be killed. I dont even get half of the idea.
Yet its a bomb and still is. It doesnt have a decision structure u explained in your post, yet as u play, if u think about it, u cant change anything, that just one thought, made me save, and quit at once. Yet i came back to it. And one more thing, its a huge and a very long game. Not many people past it. Yet they still can tell their friends what a great game it is. The ADD structure is interesting, i will definatelly consider it. I will have to make some slight personal modifications to suit my game.
Please, can anyone tell me whats so great about Half-Life?

Thanq
This post was made By Amigo
at 11:26 eastern time, friday, 11:27pm.

Good luck to all

[RodeoClown]

I remember reading that when they were designing it (the second time - they scrapped the game and restarted from design when they finished it) they decided that it would be 2 (or 10, can't remember) seconds until the next 'thing' happened, whether that be an enemy or a puzzle. This meant that a player could sit in one spot and rest but as soon as they wanted action all they had to do was move forward, so there were no 'boring' spots in the game (I think 'Zen' was actually pretty cool, even though most people seemed to find it boring). I think that the constant action made the game more fun, you were driven onwards because you would think, I'll just take out one more bad guy... and then it's 4am.

[MacD]

Dunno if you still read comments to an article this old, but here goes:

Good blog; insightfull and interesting. Thanks for that.
As for game principles and game rules: I have the one rule to rule 'em all (and to think I dislike the LotR books :)). It's been covered before in different aspects, and you describe an aspect of it with the 'trade-off' principle in pac-man, but the generalisation of that principle, the goal any game dev should be striving for and asking him/herself is this:

"Let the Player decide."

All the frustration you describe (towards save points, difficulty levels [sorry, pet pieve: Homeworld (2) is THE example; great game, beautifull game, great gameplay...but it punishes the player who doesn't have the time/inclination to really learn the single player tactics to the utmost], whatever) is covered by this rule.

You know, I was going to illuminate this point, but I think everyone gets this (especially with all the articles having been written on this subject): the player buys enjoyment, not punishment. The gamer must have a consistent choice in all matter.
GTA:VC proves this, as does The Sims; you can restrict the player, but it must be selfconsistent with the gamerules and must not detract from their choice of gamestyle.

And pissing the player off for reasons other than their own mistakes (the greedy going after the ghost/fruit in pacman or the 'I must finish another mission' in GTA3 are acceptable) just is not done (like the 'game over' due to an unavoidable deathpit-unless-you're-psychic-or-have-played-the-level-before or the 'screw you, you'll play the game at whatever difficulty level we feel apropriate' mentality).

Sorry for the rant, but it does boil down to human mentality: 'I payed for this, so you better let me have fun'.

And I did, so you better :)