A Voice in the Dark

It’s occasionally instructive to examine a taxonomy and formalise it somewhat.

Russ Nelson’s discussion of game theory in economics relied on what is, at heart, a fundamental misunderstanding of what a game is. He selects a subtype of game and acts as though this is the only possible type of game. In reality, he ignores many alternative forms of game, and thus needlessly restricts his thinking.

I’ve defined a few terms related to this in a way that I find useful and instructive. Your mileage may vary.

Play is the pursuit of fun. It is the normal reason one makes use of toys and games. Fun, however, is a fungible concept - everyone has a different definition of it. In order to accurately interpret the suitability of a toy or game, one must first establish the current definition of fun. Depending on that definition, it may not be necessary to employ toys or games - young children often play by simply making noises and wandering about at random, amusing themselves with no particular attachments to any specific activity. A more adult version of this is web surfing, where you begin by looking at some specific site, but rapidly devolve into repeated instances of “ooh, shiny” as you happily wander about the internet. (It’s been a long time for most adults since they got to just play. I’d be tempted to point at this as the single greatest value in the internet.)

Humanity, however - for better or worse - craves structure. And as a result, open play rapidly gives way to the toy. The toy is not necessarily a physical object; neither is the game, which we will cover shortly. What distinguishes the toy is the definition of rules. The minimal set of rules is “this is a toy”; it establishes a focal point for play, and play may be further self-directed by the addition of more rules: ”This small model car is a toy; it follows the rules natural and normal to real cars. It usually drives forward, does not drive in reverse for more than a few seconds, does not drive laterally at all, and is required to obey traffic laws and rules of the road. This particular car is being driven from home to work by a businessman who is late for a meeting.” There is no natural requirement for any of these rules, and they might be changed or removed at any time. When a child playing by these rules becomes bored by them (most children view “lack of boredom” as a definition of fun), he simply alters the ruleset: “Oh no, space monkeys are attacking! The businessman has to save the planet by turning his car into a spaceship!” The boundaries are essentially wherever the player chooses to put them. An adult variety of this might be Wikipedia surfing; rather than browse the entire internet, the user browses only Wikipedia. (One is rather less likely to find malicious or dangerous content on Wikipedia.)

But toys are for solo play, and human beings crave not only structure, but social interaction. And here is where it gets interesting: a game is when two or more people agree on a set of rules for play. They are not required to play together. They are not required to play for the same reason. They are only required to agree on the rules. And the rules do not necessarily require the players to interact at all: “six degrees of Kevin Bacon” is a game, even when only one person plays. The goal is to start with one actor in one movie, and identify a chain of co-stars such that one eventually reaches Kevin Bacon. Groups will often turn this into a social event by having one person identify the starting point, and having group members trace the appropriate chain. But all one really needs to play is the IMDB. A variant of this is “clicks to porn”; from a given site, how many clicks on hyperlinks must be made at a minimum to reach pornographic material? A single player plays to beat his high score on a given site; a group competes, sometimes with a bidding war akin to the old game show “Name That Tune”.

In group gaming scenarios, any player may ruin the game for all concerned by twisting the interpretation of the rules. By altering the rules of the game without consent from the other players, this player creates his own personal toy - which may be useful in pursuing his own brand of fun, but alters the ability of others to seek their own definitions of fun. In the party game “Killer”, for example, players sit in a circle and one is secretly selected (usually by dealing cards) as the “killer”. The “killer” may “kill” other players by winking at them, at which point the “dead” player must announce that he has been “killed” and leave the circle. The killer may win by “killing” all the other players, or another player may win by correctly identifying the “killer” before being “killed”. Once it has been agreed that this game will be played, it is not too terribly uncommon for one player to take the approach that if he is selected as the “killer”, he will commit “suicide” by announcing that he has been “killed” and leaving the circle. In the event that this player actually does end up selected, fascinating social dynamics play out, but the only one having fun is the recalcitrant “killer” who has invented and followed a rule nobody else knows.

This is often called “cheating”. It’s a violation of the fundamental social contract of any game: that everyone is playing by the same rules. There are a few special cases of this; in games with very few rules, the social contract often includes rules which are not explicitly stated, and the inexperienced player can violate these implicit rules (or infer entirely different rules) without intending to do so. In games with a vast quantity of rules, the social contract often excludes many of those rules, and an inexperienced player may then follow rules nobody else follows - or, more damagingly, demand that others follow rules that are not normally followed. In games where the rules are enforced by an automated mechanism - e.g. a computer - an individual who has significant familiarity with the mechanism may be able to partially or completely bypass it, thereby exempting himself from a rule to which others remain subject.

To relate all of this back to economics, game theory in economics is not about examining the play of the game, but the rules of the game: what are the rules governing play in specific situations? Where do individuals “cheat” in the game, e.g. through information asymmetry? When the game/market breaks down and no fun/value results from play/trade, have one or both parties in the transaction cheated by converting the game into a toy?

I’ve been criticised a little for this set of definitions, and accused of rewriting terms to match my own needs. I don’t quite understand the criticism; philosophers since Wittgenstein have long wrestled with the question of what constitutes a game, and Chris Crawford was trying to establish a formal taxonomy in his 2003 book “Chris Crawford on Game Design”. The most accurate criticism I’ve heard so far is that this reclassifies some games as toys, but this is intentional - some things we call games are toys, and need to be classified as such to draw useful comparisons and conclusions.

Not to sound arrogant, but of all the things I’ve come up with to date while sitting around thinking, I suspect this is the one for which I’m most likely to become known. It simply has the highest probability of being useful.

Filed under: Business, Gaming, Philosophy
No Comments

Russ Nelson doesn’t understand the purpose of game theory in economics. He says games have winners and losers.

This is not true. When you play “doctor” or “house”, everybody plays and everybody has fun. Most games do not represent free markets; they represent tightly controlled markets overseen by a rigid and implacable central body called “the rules”. Only in the earliest and most innocent of games do you find representations of free markets, and in those games, everyone who plays does indeed win.

More advanced games sometimes try to provide free market simulations, e.g. Dungeons & Dragons and its ilk, and one might even say Second Life or World of Warcraft qualified. But these are not really free. They pretend to be free, but instead institute the same sort of rigid heirarchy and implacable standards that allow people to practically lose where they cannot theoretically lose. You don’t really lose when your character dies in these games, because you can start over with a new character, but starting over is rarely fun once you’ve done it a few times. Even though in theory a loss is not really a loss, after enough deaths the player loses interest in playing the “start over” game again - and quits the game. Quitting is the ultimate loss.

Game theory investigates economic effects on the small scale. How do three or four people with known incentives and properties behave in a given situation? It’s where information asymmetry can be investigated and understood, giving one the proper tools to extend their understanding into the macro field and overall market theory. It is game theory that shows the budding economics student why this matters.

Filed under: Business, Economics, Gaming
No Comments

I made this comment to a post on IMAO and thought I should drag it over here. 

Conservatives do not hate gays. They simply want gays to “know their place”, in much the same way they once wanted women to “know their place”, and blacks to “know their place”, and immigrants to “know their place”. What people mistake for hate is the idea that these people have a place, which is on its face a prejudicial and arguably bigoted perspective.

Unfortunately, it’s also true. Everyone of every variety and in every group does, in fact, have a place. The place you have now may not be the place you have tomorrow, and a smart conservative recognises that. What these conservatives are trying to promote is the idea that while gays/women/blacks/immigrants certainly can and should aspire to have all the same rights and privileges and status that white men have, it is counterproductive to force the issue by ramming “EQUALITY NOW” down the country’s throat.

Historically, forcing the issue when the culture is not ready causes a backlash which results in more pain and suffering than the original state promoted. The similarities between the emancipation proclamation and prohibition are generally avoided as a dangerous subject, but they remain instructive.

I for one do not believe that gays would be better off mostly-closeted as they were in the 1970s. I believe they are better off now. I believe the entire American culture is better off now with more acceptance of gays. I believe that we will be better off still when gays are fully accepted as members of American society and nobody even considers one’s personal sexuality worthy of note.

But I do not believe that is what we will get if we make a law that says you MUST fully accept them (e.g. by extending marriage to include them). I believe what we will get is underground cells of homophobic jerks who go out and do real damage to real people to make a political statement about homosexuality. I believe that this is an unconscionable result that must be avoided at all cost, even if it means people accuse me of being homophobic for the rest of my life.

If we just wait, nature will take its course and homophobia will become a quaint oddity of 20th century culture. It’s the long view. Surely we can endure a few fag jokes in exchange for a culture that really does accept homosexuality, rather than one which gives lip service to the idea in public while continuing to fear, resent, and persecute homosexuals behind closed doors.

Filed under: Gay Rights, Politics
No Comments

Russ Nelson again hits a subject on which we largely agree, but doesn’t illustrate it quite the same way I would.

Let me illustrate America being better off… my way.

Thirty years ago, there were stray animals all over the place, and on more than one occasion my friends and I were bitten or scratched or chased through the neighborhood. Garbage cans in the back alleys - where I and my neighbors played - were frequently filled to a depth of two or three inches with maggots. Clouds of mosquitoes were easily visible over low-lying areas of the street where water collected in potholes. On more than one occasion, milk purchased from a school cafeteria or a local diner was spoiled due to improper refrigeration. My aunt found a cockroach baked into a donut. My grandfather routinely set out rat traps, roach motels, ant traps, and flypaper. While feeding the dog, my mother nearly had heart failure when a giant water bug crawled out onto her hand as she reached into the bag of dry food.

And we were not particularly poor or disadvantaged. This was simply how life was in the 1970s. Our homes and businesses were naturally and normally infested with any number of pests and vermin, and both dogs and cats ran rampant in our streets.

That’s what strikes me. I literally don’t remember the last time I saw a rat, roach, water bug, maggot, or stray dog. Somewhere in the last thirty years, these went away. They became emblems of poverty and squalor, not natural and normal components of daily life.

I suspect this is rather the same way it must have appeared to outlying territories of the Roman empire… who certainly rankled under Roman rule, but still had to admit that the roads and the plumbing and the hygiene were quite nice, actually.

Like those territories, we look at the upper crust with their billions of dollars, and we complain about inequality. But fundamentally, we’re all much richer. Why should I really care that Bob is three thousand times richer when I’m only twelve times richer? Bob was always richer than I was. Can’t I just be happy to be twelve times richer? Why does the degree matter?

Filed under: Business, Economics, Philosophy
No Comments

It is with some shame that I read this article on warnings.

I know this. I have known this for years. I have, indeed, lectured people on the superiority of “Undo” to warnings.

And yet, through all those years, I have continued to instinctively and automatically add warnings to my programs for one and only one reason.

It’s easier.

Adding a warning to a program is laughably easy. You say “this could do something bad”, and type something like “if(CreateOKCancelDialog(”Are you sure?”)!=ID_OK) break;” and you’re done.

It’s not as easy to make an “Undo” framework part of your foundation. When we start, the application is small, there are few operations, nobody really makes mistakes. It grows, we add more functionality, it becomes more complex. We don’t notice. We’re frogs in the boiling water. We find other people confused, and we create a training program, because it’s easy. This works for a while, but not forever. Eventually, “Undo” becomes unequivocally the Right Thing, and then it’s too complicated and there’s no time and it can’t be helped.

We need undo. Period. End of story. That’s what has to sit at the bottom of the application from day one, because we are going to need it. We should know that. Stop it with the warnings! Undo! If you don’t have it, you need it, build it! If you do have it, use it!

Those of you who work with me are going to start hearing a lot more of this.

Filed under: Business, Programming, Software
No Comments

It’s a serious question.

Let’s say two people are standing around expressing their views on, say, stealing cable.

“Stealing cable is immoral and wrong,” says the first. “Anyone who steals cable should be shot.”

“Stealing cable is normal and natural,” says the second. “Lots of people steal cable.”

Before the debate can proceed further, an observer pops up and brandishes documents that demonstrate both people are stealing cable.

Which person is more guilty?

It seems to me that both are equally guilty, and neither person can or should be held more or less accountable for his actions. But lately, I’ve heard a lot of left-leaning folk offer the opinion that the first person is more guilty because he is a hypocrite.

I don’t understand why this is somehow worse than simply not respecting the law. Why is it worse to say “please follow the law” and then go break it? Isn’t it worse in practice to say “screw the law” even if you don’t break it?

Filed under: Philosophy, Politics
No Comments

I’ve heard many people complaining that this study is not what researchers should be wasting their time doing. How about a cure for cancer?, suggests one blogger I read. How about a cure for AIDS?, suggests another.

Well, at the risk of being patently obvious, I suggest that the vast majority of scientific researchers are largely unqualified to seek cures for anything. Indeed, they are probably ill-suited for anything much more important than… well, fluff.

The researchers are still learning important things about how to conduct a research study, like how to select good participants, or how to draw rational conclusions, or how to publish convincing papers. These are critical skills that you had better have 100% dialed in before you start working on anything that really matters.

But what really annoys me is, why the hell are the news media wasting my time with this crap? Yes, men like hot women - DUH. Yes, women like men who are near their own attractiveness - DUH. Anyone who has paid any attention at all knows these things. That a research project was done on it and reached the well-known conclusions is not news. It is, indeed, what amounts to an extra credit project. “In other news, four largely inexperienced researchers demonstrated that they aren’t completely incompetent.” What? I don’t care. Tell me something that matters.

Filed under: Notices
No Comments

Giving some more thought to my argument with Orochimaru last week, I’ve coalesced an idea.

Whenever you are going to pass information over the network, strong typing on the sender side is irrelevant.

It doesn’t matter what the other side did, once you retrieve data from the network, it is to be treated as potentially hostile regardless of other factors. That means you need to directly verify and validate everything you get. You need to establish what the data is supposed to be, whether it was provided in the proper format, and whether the data itself is acceptable.

There are, of course, a great many ways to do this. But in the end, you still do the same things: establish the purpose of the data item, identify the rules regulating that data item’s permissible content, and verify that the data itself meets those rules. There is simply no way to avoid this. You can make the sender jump through a great many hoops to give you something that he has already validated, but all this allows you to do is create a mismatch between his rules and yours. You never actually avoid the necessity of validating the data against your own rules. It’s a false economy.

Filed under: Business, Microsoft, Software
No Comments

Actual Wisconsin state law (http://www.legis.state.wi.us/statutes/Stat0101.pdf):

101.58 Employees’ right to know.
(1) SHORT TITLE. Sections 101.58 to 101.599 shall be known as the “Employees’ Right to Know Law”.
(2) DEFINITIONS. In ss. 101.58 to 101.599:
[...]
(j) 1. “Toxic substance” means any substance or mixture containing a substance regulated by the federal occupational safety and health administration under title 29 of the code of federal regulations part 1910, subpart z, which is introduced by an employer to be used, studied or produced in the workplace.
2. “Toxic substance” does not include:
[...]
f. Lutefisk.

Now, maybe this is just me… but when you need to codify in state law that an ethnic food is explicitly not considered a toxic substance, perhaps you need to sanity check that assumption.

Why was I reading Wisconsin state commerce laws? No reason, really. I’m just a law and economics geek.

Filed under: Uncategorized
No Comments