The ludic fallacy

 Posted by (Visited 16998 times)  Game talk  Tagged with: , , ,
Dec 092008
 

I was just pointed to this wonderful essay by Nassim Nicholas Taleb, author of The Black Swan and Fooled by Randomness.

First Quadrant: Simple binary decisions, in Mediocristan: Statistics does wonders. These situations are, unfortunately, more common in academia, laboratories, and games than real life—what I call the “ludic fallacy”. In other words, these are the situations in casinos, games, dice, and we tend to study them because we are successful in modeling them.

–Edge: THE FOURTH QUADRANT: A MAP OF THE LIMITS OF STATISTICS By Nassim Nicholas Taleb

It’s not the only ludic fallacy I can think of. Recently I had a discussion with a management and leadership consultant, and we were discussing the generational characteristics of Millenials versus Gen X in the workforce, and we were talking about how a gamer mentality may have affected the way Gen Y behaves in the workplace: more likely to follow the rules, more likely to work in teams, more needful of reassurance, less creative and risk-taking, less likely to see the full scope of irreversible consequences of a choice, and less likely to see things in shades of gray. In a way, these sound like thinking trained by games.

I’ve talked in the past about whether games are, because of their inherently mathematical nature, limited in conveying certain types of information. But I think it is also worth asking whether the scope of their models is effectively misleading, or even actively lying, about how the world works.

In Taleb’s books, the bogeyman is the extremely rare event with a high impact, which he calls a black swan. These are historically visible, but exist on timeframes long enough that we consider them arbitrarily rare: essentially, a non-event in our planning. Yet they also have a highly disproportionate effect on the environment in which they occur. In the article he offers this example:

There is a measure called Kurtosis that indicates departure from “Normality”. It is very, very unstable and marred with huge sampling error: 70-90% of the Kurtosis in Oil, SP500, Silver, UK interest rates, Nikkei, US deposit rates, sugar, and the dollar/yet currency rate come from 1 day in the past 40 years… This means that no sample will ever deliver the true variance.

A picture of Mission Valley, flooded in 1916

A picture of Mission Valley, flooded in 1916

What Mission Malley looks like today

What Mission Valley looks like today

I am reminded of an article I read recently about the San Diego river, which is an “upside-down” underground river almost all of the time. It runs through a beautiful valley in San Diego — or would, except that there is no sign of water. But the Mission Valley area is still in a 100-year flood plain, and we silly humans built a mall there and a whole bunch of hotels. It’s on record that the flooding once reached the steps of the mission — which is on a hill above the valley (by Qualcomm Stadium). It’s been 92 years since that flood. Are we worried yet?

Humans have mental trouble with a lot of aspects of statistics. One of the commonest ones mentioned is that people tend to extrapolate linearly, not exponentially, so out gut instincts on rates of change are typically wrong.

Core to Taleb’s reasoning is the notion that we can live more safe and secure lives by being incredibly pessimistic and always planning for the absolute worst. By always making decisions based on the potential downside, rather than the potential upside, you end up with a far more conservative, but also more successful investment portfolio over the long run, whereas people who make decisions based on upside end up at a very high risk of losing everything in a dramatic day.

Graph of rising action in an audio play, by Yuri Rasovsky

Graph of rising action in an audio play, by Yuri Rasovsky

Games have characteristic difficulty ramps that have little to do with real-world statistical distributions. For example, we are most familiar with linear curves, Gaussian and Poisson distributions, exponential curves, and the new modern classic, the punctuated curve. This last one is the difficulty curve of a typical game these days, which rises linearly or exponentially to a peak, then falls back somewhat, then rises again, in order to provide an extra challenge at a boss, followed by a breather. It is, of course, also the classic graph of tension in a novel or other form of dramatic writing.

What would a game with true black swans look like? It would throw something unpredictable and monstrous at you. Well, why couldn’t Tetris occasionally drop a monster block? Just one time in a thousand. Why couldn’t Fallout 3 occasionally deliver a randomly rolled instantly fatal death blow? Because it’s not fair!” “Because it would make the game less fun”? Perhaps, but the game might more accurately be reflecting reality, and thereby teaching better strategies.

In fact, we have seen this in action — in early Everquest, there were high level mobs that were aggressive and roamed the newbie areas. They slaughtered newbs who simply had no ability to fight back. Successful players during this period learned to be conservative in terms of how they moved around the zone. And it may be that the reason why multiplayer PvP games always come back up as an idea is their sheer unpredictability. Other humans are pretty good at providing black swan events, as UO demonstrated over and over again, and as EVE and Shadowbane continue to demonstrate today.

In light of today’s Pew report in which we learned that more adults game than don’t, it’s worth asking what other lessons games are teaching. Games have relatively finite palettes, because they exist on the basis of providing complete models — even if they are not NP-complete in the strict sense. They don’t tend to surprise, but rather reveal the expected, in many ways, because that is what signals to the player that they are mastering the problems presented.

I have mentioned before that one of the possible great strengths of games is that they can present complex problems as being amenable to analysis.

…games are by definition tractable. They teach you to chunk up, to chop apart, to disentangle. Perhaps this is where games can most change society: a way to look at the problems so that we no longer throw up our hands…

Games for Change, closing address

But when we look at that list from the top, about Millenial workplace attitudes, it’s hard not to draw the analogy to a generation raised in a system where there are always rules, always correct answers, where you geta pat on the head every time you do something right, and curves are always easily projected.

[Edit: it was pointed out that all this sounds like I am slamming Millenials; in fact, I did write a whole book about how good games are as teaching tools, so let me point out the clear benefits of gamist thinking as reflected in Millenials: team-oriented practices, clear goal orientations, high self-confidence, a firm belief that things are solvable, a lack of the cynicism and skepticism common to GenX… it’s not all bad!]

Rather than just bemoan this, I’ll instead issue the challenge: what is the fun game that features black swans, phase transitions, and the catastrophic 100-year flood? How do you sculpt a system that does this without chasing away the newbies?

Because it is they, above all, who need to know that their job and bank are far from being institutions — a lesson everyone is learning at a rapid pace these days; that yes, even something like the Fashion Valley Mall can end up underwater. That would be turning a feature of games against themselves, and making the medium — and our players — more robust.

  38 Responses to “The ludic fallacy”

  1. I’d argue that, when it comes to MMOs, patches and expansion packs fill that role. In the big ones, the physics of the world change overnight and everyone has to scramble to re-establish themselves.

    In general it seems though that these kinds of events are more bothersome to established players than new ones — catastrophes put everyone on equal footing again. This can be very healthy for the playerbase if it’s done right and goes a long way towards preventing elite players from running everyone else off. But the problem with blending too much catastrophe into a game is that people can quit games, where in reality they have to live with it. It’s hard to teach a lesson about overcoming adversity when it’s an option to go do something else.

  2. Would SimCity count? The trouble with a Black Swan event is that it becomes part of the game, and thus anticipated. If Godzilla attacks are in the menu, is it really a Black Swan event?

    I like the idea of having a creature (or spot) in the newbie area of a game that can’t be beat at that level. LOTRO sort of does something similar — not in the newbie areas, but there’s a quest that involves sneaking past an ogre to pick up something. It’s an ordinary mob and can be killed, but it’s also 7-8 levels about the player at the time the quest is given. A quest that says “Kill X, but don’t go near Y” could be an interesting way of suggesting mortality at an early level of a game. (And proper warning in quest text means that for slaughtered newbs, it’s their own darn fault.) Plus there’s always something sweet about coming back to an area ten levels higher and slaughtering the mobs you had to avoid before.

    But that’s a bit off topic.

    Would the recent zombie plague in WoW count? I think it does. And in fact I would argue that any true “Black Swan” event in an MMO would have to be person-driven — anything otherwise will end up in thottbot. Not that I wouldn’t love to have more randomness “built in” to game events, other than the classic “spawning” and loot issues.

  3. There are plenty of (modern) board & card games that include “Black Swans” or radical changes. The great old card game “Nuclear War” has a huge phase change from the Peace phase where you steal people from each other with Propaganda cards to the War phase where you blow each other up. Magic: The Gathering’s exception mechanic means that new cards and new strategies can substantially disrupt and reshape game play in very interesting ways.

    Games with card/event-based mechanisms can vary wildly from one session to another.. Fluxx being an extreme example, but there are other family and strategy games that use events to incorporate extreme “what if” scenarios either for players individually through new capabilities or environmental changes.

    One can also look at some of the Avalon Hill games such as Republic of Rome where the management of chaos is key to the game (Players can benefit by increasing disorder in the game for their individual benefit, but if the republic has too many wars at once, then all of the players lose to the game). I think this is also true of Reiner Knizia’s Lord of the Rings game.

  4. The trouble with a Black Swan event is that it becomes part of the game, and thus anticipated. If Godzilla attacks are in the menu, is it really a Black Swan event?

    I’d say no… and similarly, I think a lot of the board game cases simply aren’t rare enough. Even in a game like Fluxx or Magic, you don’t see a single play capable of wiping out the entire opposing field in one move and winning the game — that happens only once every 10,000 games. I mean, Fluxx isn’t even that large a possibility space; you learn to live in the midst of chaos, but its fairly controlled chaos.

    That’s the thing about black swans — they aren’t defined by catastrophe, but by rarity & magnitude.

    It might be that the patch/zombie type things are closer.

  5. Magnitude is a key issue, it seems to me. Instead of one huge catastrophe that wipes everything, maybe catastrophes that affect those who failed to plan, or just failed, and on levels based on how much the failure was.

    In some of the medieval strategy games, where you try to increase herds of cattle and grow fields of crops, you can be affected by disease or blight. Keeping cattle separated into smaller herds and rotating crops can reduce the chances of bad events, and reduce the effects should it happen. Bad weather can come and affect both regardless of what you do. I’m not entirely sure this is a classic example.

    I guess the point I’m making is, since gamers don’t like to be “you lose”ed, maybe a division of what that player represents can allow for black swans of various types.

  6. The first thing that popped into my head was Crusader Kings. I mean, overall, it might be more accurate to call it a historical simulation than a game, since all sense of balance is thrown out the window for the sake of historical realism, but anyway, you basically spend the whole time trying to fight the Muslims (theoretically – you actually spend most of your time trying to invent claims to steal territory from other Christian rulers), but then at a random point between 1200 and 1600, mongols will appear, with a random number of troops. Sometimes they just get killed off by muslims and you don’t even realize that they’d appeared, and sometimes they overrun half of Europe, and there’s no way to know whether or when it will happen.

    In this case, the very fact that the game sets itself up as simulation rather than a balanced game seems to make this work.

  7. Does the gamer generation take too many risks (according to Taleb) or not enough (according to the consultant)? Or are you saying they have the worst of both worlds: they think they’re being conservative, but in fact they’re taking risks?

  8. Remember the Battle of Trinsic? It was an Ultima Online event, a bloodbath of a fight that the players could and did lose on some shards, and which had lasting consequences on the game world. It was a once-in-a-lifetime experience, one that many of us remember years later.

    Now granted, it was a scheduled disaster… we were gathered in orderly ranks and loaded for bear when the undead hoards arrived (for all the good it did us). But it seems to me that some of the lessons discussed above — despite the best preparation and following all the rules, failure IS an option, and there are consequences for actions or the lack thereof — were brought home with unholy glee by the GM-controlled enemy.

    Any newbies popping into the city that day would have had a memorable if somewhat short introduction to the game. I doubt that many were dissuaded.

    But as a gamer, I don’t want to be incredibly pessimistic and always plan for the worst. I want my cookie when I jump through the right hoops, and I want the big cookie with chocolate chunks. I have real life handing me daily reminders that I am limited and fallible — I want my diversions to make be feel superhuman and competent. If the game hands me lemons, I want plenty of water, sugar and pitchers to make lemonade.

    Would anybody play a game featuring random flea bites -> bubonic plague -> permadeath for 50% of the player base? It may be fine training for the next pandemic, but it doesn’t sound very entertaining to me.

  9. Yes, games with a purely flow-oriented designed tend to teach the kind of lessons that public schools have been trying to teach more crudely for a century, and that makes dystopia much more possible. This kind of blindness to model-risk is what made the financial crisis possible, although self-blinded folks on both sides the trades were lured into by low-interest rates and the Basel regulations.

    The notion of a black swan is actually an extreme version of the variations on algorithmic theme that give games their content-longevity, it raises an interesting notion about the dynamic content-creation problem. By creating algorithms that produce black swans (or by creating rule structures that invite players into creating them, like Basel did with disaster finance) we could tap into another order of process intensity and interactivity.

    I´m big on games that warp and distort pure agency, using the flow as a hook to lure you into traps, because I think these games can not merely teach a certain kind of decision making, but encourage critical thinking by reflection on those decisions. I think Oiligarchy does a fine job of this, for a recent example. I think these kind of games can both be the organizing implement of the modern world, but also the hedge against either a dystopic New World Order or a New WoW Order, whichever is worse.

    Finally, you could implement black swans with a simple randomizer function with a low period, but it´s better when they´re systemic events. I saw September coming in May, for example, because the causality behind the financial crisis was actually quite simple, the equation and data were public enough for any twelve year old with a salt of critical thought to connect the dots. That´s what´s really significant about phase transitions, the transcendence of the model was implied by the model all along.

  10. Does the gamer generation take too many risks (according to Taleb) or not enough (according to the consultant)? Or are you saying they have the worst of both worlds: they think they’re being conservative, but in fact they’re taking risks?

    Hurm, how to put it. The gamer generation tends to see things as safe sandboxes — there’s a bit of entitlement there based on how they were raised, and the prosperity they have always known. (For example, data shows that they ask for raises and promotions sooner than previous generations).

    On the other hand, they are less likely to jump outside the box, much less “think outside the triangle.” (All in terms of statistical averages, mind you!).

    This may lead them to paradoxically be more vulnerable to black swan events, which do not fit neatly into the perceived model.

  11. Turns out Wikipedia has an article on the ludic fallacy, which may help the discussion for those who have not read the book.

    http://en.wikipedia.org/wiki/Ludic_fallacy

  12. The other aspect of Black Swans is they create true believers, those who will stick to some task or mission in hopes of the great single game changer, eg, the talent agent that mysteriously appears at the gig to give you that record deal. (They do appear sometimes and the rottenness is something else takes the deal right out of your hand when they do, usually someone in the band goes bozo from anxiety, or their signficant other does.)

    They do this because as behavioral science shows, if the payoff is big enough, almost anyone will keep pulling the lever at a slot machine as long as the bet isn’t that big. It is considered a root cause of repetitive obsessive behaviors. It will cause people to vote for a wildcard candidate, chase an unattainable girl, swap spouses, and so on.

    There is also a justification behavior that comes along with it to explain away why one is playing a game they always lose, or will vote for a candidate who can not logically or reliably make good on their promises. In one sense, the Black Swan is one of the reasons that civilizations do improve: hope is a black swan and it will engender risk taking that even if the payoff is not immediate or direct, the lightly coupled side benefits or bets take effect and there are positive unintended consequences.

    It’s just a bad way to run the train system or schedule deliverables.

  13. Well, catastrophic out of the blue events can and do occur in the metagame. One day you wake up and your favourite grinding trick has been nerfed. Or your account deleted for cheating.

    Then players bitch and whine.

    Sure this is human activated but it is similar.

    Really I think the problem is just getting gamers to accept loss as part of playing. If it’s win all the time then win has no meaning.

    Shirens 50:50 love:hate reviews is at least 50% heartening. Roguelikes do of course throw black swans at you due to their nature.

  14. There is another approach to this issue: creating a noisy or signal laden game with measured ambiguity to enable the player to solve a problem not solved within the game or by the game designer. You can think of this as emergence farming.

    Read this:

    http://www.fooledbyrandomness.com/LudicFallacy.pdf

    Compare it to the problems of terrorism, voting for wild card candidates, and repetitive compulsive behavior. The point to be made regards education and gaming is that games and statistical theory teach a limited or false model for risk management. To tie this to Governor Blagovitch, he is the example of a true believer, that is, that he is destined for a high payoff so he keeps pulling the lever despite all advice that his risky behavior has a high certainty of a bad payoff in his current environment.

    NOTE: The author is wrong on one point. Unknown Unknowns (unk unks) are an old term taught in the management courses at General Electric for decades. I covered the topic in Beyond the Book Metaphor (1990) under ‘catastrophic failures’ as part of the explanation of why we study chaos theory and problems of superstitious learning even in controlled environments.

    To tie this to innovation, one can create an environment that cultivates innovation. There is an old science fiction story about that in which noise theory is used to create an information rich environment laden with clues about the problem to be solved (anti-gravity), and a con played to convince the scientist/players that the problem had been solved but lost, so they would give up doubt that it could be solved. The game is designed to trick the player into thinking outside the box.

    That is high education done well.

  15. Raph –

    I’ve got to disagree. A game can only meaningfully explore “Black Swan” events by making them much more probable than in “real life”. If your point was that game’s can’t “teach” about these incidents(i.e., dealing with catastrophe or highly anomalous situations), my previous examples certainly address the issue as do some of the others mentioned above.

    Other commentators have noted actual cases where “black swan” events have occurred in games.

    If I was to create the “Mission Valley” game and it covered hundreds of years, your “Black Swan” flood (and major fires) could happen with some regularity.

    Can games address Black Swans? Yes. Can they have unintentional Black Swans? Yes (oops) Would you design a game with a Black Swan that occurred extraordinarily rarely. Probably not (it would not be worth the designer’s effort) unless it was a procedural Black Swan (say, a funky game of Dwarf Fortress).

  16. I’ve got to say, this comes off as an incredibly sour-grapes view of millennials. If anything, I’ve seen more to suggest that millenials are the MOST likely to think outside the box, and to question how things have been done in the past. What you’re saying about them is counter to the majority of research and reports, which generally cite their desire to learn and be challenged, not their inflexibility. (http://www.generationsatwork.com/articles/millenials.htm and http://en.wikipedia.org/wiki/Millenial are as good starts as any.)

    I like the post, and I think you’ve made some excellent points regarding the value of true “Black Swans” within a ludic model, but the largely unfounded anti-millennial spin left a real sour taste.

  17. I believe not all people would welcome learning into their gaming habit. A lot of this industry’s customers just want to “play”, not be challenged in any way or learn about the world.

    Also, there are many experiences that gaming could almost provide in the form of almost-being-there understanding of situations too dangerous or otherwise inaccessible to the gamer, but… the randomness of real world may not be one of them.

    In general, unpleasant truths are best believed when experienced in the real world, rather than taught or talked about in media. Adjusting to unpleasant reality is expensive psychologically and I believe most people will only adjust if needed, i.e. if reality, the real reality proves certain unpleasant truth to be really true.

    If I see mobs of higher level in a newbie zone, I will most likely feel that the developer is a cruel bastard, not that he’s prompting me to meditate about the uncertainties of human condition. And that’s because it’s the more comforting (if shallow) thing to do.

  18. Nabil, I am just passing along what I have read and heard. There was no intent to be overly negative on Millenials. I should have spent more time on the positives of gamist thinking — though it does seem like I wrote a whole book about that. 🙂 Maybe I will edit the article to reflect this.

    Overall though, the Wikipedia article you link to makes basically the same points:

    According to a survey by BusinessWeek of almost 4,000 readers, Generational Tensions ranked in the top 6 issues impacting the workplace. With four generations in the workplace, these issues can grow from tensions to larger generational divides and conflicts. The 2008 Gen Y Perceptions Study[15], conducted by the Cal State Fullerton Career Center and Spectrum Knowledge, measured how Gen Y views themselves in the workplace in comparison to how the Boomers and Gen X view Gen Y. Some key findings include the following:

    * Gen Y Wants Instant Gratification: Managers and employers often complain that Gen Y “wants instant gratification” and it seems like Gen Y itself recognizes this, too. 89% of Gen X and Boomers agreed that Gen Y “wants instant gratification,” while 73% of Gen Y agreed with the statement as well.
    * Casual and Professional Can Co-Exist: Though almost three-quarters of all survey respondents agreed that Gen Y dresses and behaves casually, the generations agree that professionalism may not necessarily be a “casualty of casual.” Out of all 22 perceptions regarding Gen Y that we provided on our survey, Gen X and Boomer respondents disagreed with the statement that Gen Y “lacks professionalism” the most. Of course, this may not mean that Gen Y is viewed as professional, but the majority disagree with the view that they’re unprofessional.
    * Even Many Gen Y Think They Feel Entitled: Two-thirds of Gen X and Boomer respondents agreed that Gen Y feels entitled. Surprisingly, however, almost half of Gen Y respondents also agreed that their generation “feels entitled to job benefits they’ve not yet earned.”
    * Willing to Pay Their Dues?: Our study found that almost twice as many Gen X and Boomers agreed with the statement that “Gen Y lacks willingness to pay their dues” compared to how Gen Y participants rated their own generation. In fact, Gen Y was 17 times as likely to strongly disagree with the statement that the generation lacks willingness to pay their dues.
    * Masters of Multitasking or Misperception?: We often hear anecdotally that Gen Y is great at multitasking, working in team environments and self-directed learning. Though Gen Y agree that these are some of their strengths, their Gen X and Boomer managers and supervisors don’t agree that they excel at these working styles.

    And my reading of the other link you supplied does too.

    On the “think outside the box” aspect — that came from the management consultant guy, and I think it was an outgrowth of the “craves structure” piece.

    BTW, on the last one, multitasking is a myth; all the cognitive research says so.

  19. That said more about classic generational conflicts and older generations perception that about the real Gen Y trend, precedent generations established workers always see younger new workers how ambituous and claiming for underserved job benefits too fast, and all time young people want grow up and have these benefits quickly. Nothing new under the sun. When generational tension was not a problem in workplace?

    A lot of this industry’s customers just want to “play”, not be challenged in any way or learn about the world.

    Pople wants only play without any challengue? Then, what is the sense for succesful competitive games or PvP games? Online games with the randonness suministred by other real people proves that really they wants some unexpected challengue in their games.

    On the other hand, they are less likely to jump outside the box, much less “think outside the triangle

    How can explain then that many millenia voted for Obama in the last election?

    For me Gen Y still be very young to have a clear picture, the present economic situation can changue greatly their perceptions in few years.

  20. Well. I believe it is a matter of point of views.
    I will tend to see gaming as a whole, rather than concentrate on a single game. From this point of view is easy to see black swan events, in terms of games changing the rules and the usual clichées. For example, the battle with Psycho Mantis in Metal Ger Solid, the role of the princess in Ico, and maybe the accent in the construction kit in Little Big Planet.

    And I think I disagree about the gaming generation being less creative than the others, for at least three reasons: if it’s true that the gamer is more keen to stick to the rules, it is also true that, from a whole-gaming-POV, they are really fast in adapting to changes in the environment/space of possibility.
    Second, I think that to be creative you have to see and know as much material as you can, in term of games, books, images and so on. The gamer generation is overexposed to stimuli that can resolve in increased creativity.
    And third, I believe that sandboxes are pretty good in giving you constraints and limits. And it’s easier to train your creativity starting from a limited environment, rather than from a white sheet of paper in which you’re free to do what you want.

    My two cents.

  21. I don’t see how anybody can “think outside the box” since the box is merely the conceptual model of the problem. Oh they can think in a different box than provided them, or even expand the box, but to think “outside the box” would be to think the unthinkable. Unfortunately for poets and dreamers the world over, you can’t simply ignore paradox and think the unthinkable or we wouldn’t have unknown unknowns; thanks for the information on the term Len.

    But overall, isn’t it somewhat strange that we would be participating in a discussion on how to move away from overly platonic models by applying one to a generation of people. I find it difficult to believe that an entire generation has anything apart from being human universally in common. Perhaps as game designers, thinkers and professionals, we should be expending some of our energy on how to better avoid the pitfalls of our own platonic models of players. Besides, we’re designing for tomorrow, not today.

    I would agree with Patrick in that if you want the players to learn how to deal with a black swan, it’ll need to be systemic to some extent. As the example you gave of Mission Valley, it’s not that the event is “random” it’s that it’s rare and powerful. However, if our intent is to teach a player a proper way of handling it, it may require a certain amount of randomness to prevent players from disabling their critical thought while handling it. On the other hand, you can use that to your advantage. For instance, you could design the game so that it warns the player immediately of the coming black swan, the exact moment and type of the event, and even the most effective tools for stopping it. The challenge to the player would be to maintain order/profits/score on both sides of the event, as it is often the reconciliation of short term and long term planning that is lacking. Even better players could have their performance ranked, encouraging them to plan more creatively and explore new ways to solve the problem on multiple play sessions.

    A way of combining traditional design with a changed focus on black swans, in single player games, would be to use the above method for the tutorial. Then each subsequent level the player is warned of less, until finally their just warned that there is a coming event. For MMO players you would simply skip the earliest step, give an incomplete warning for those ahead of the pack, as it were, to figure out. The change would be frequency, ever lengthening between events. In an MMO we have an excellent working model of the large problem and the indifferent masses, the challenge then would be to solve in a safe environment how to mobilize and effect change in a large group for changes coming ever further down the line. Truly the hardest step would be to convince the players that it is in fact a solvable problem, no mean feat considering we don’t actually know if it is.

  22. […] Links on 10 December 2008 with no comments The ludic fallacy Koster muses on Taleb's concept of the ludic fallacy. He worries that games are not preparing […]

  23. “How can explain then that many millenia voted for Obama in the last election?”

    Because a new box to which they wished to belong was defined for them. Semiotics played smartly is a matter of creating attractive frames in which to interact. The Obama campaign went Rove one better in that game; otherwise, and regardless of like, it is the same game.

    Means matter.

    “Truly the hardest step would be to convince the players that it is in fact a solvable problem, no mean feat considering we don’t actually know if it is.”

    In the story I referenced, they brought in a group of researchers and showed them a film that was damaged by a fire, so only parts of it were understandable. In the film, the lone inventor presents his anti-gravity device with broken or noisy explanations of its operations, then it blows up after proving it can work thus killing the inventor. Some of his research is discovered and the team is left to fill in the holes. One finally gets a solution that doesn’t work at the same scale as in the film, but does provably work. Then the team is introduced to an actor who played the inventor. The explosion and the original device were a hoax. The explanations had been seeded with domains where the answer was likely to be found, then they were given unlimited access to other information. By making them believe it had been done, the controllers removed doubt.

    There are certainly holes in the approach, but it is similar to live simulations with problems inserted with the twist being the con. It seems to me that online games integrated with internet indexing could be applied similarly.

    The question of the ludic fallacy is payoff: is it worth playing the game repetitively at those low odds of success if the payoff is high enough and over repetitive playes how does the player increase the odds of success?

    One envisions a game world where certain questions yet unsolved are expressed in a library as a set of games the premises of which if satisfied would lead to an answer to the question even if not an implementation or with one depending on the question.

  24. I think a big part of the problem, at least in terms of the “thinking out of the box” thing is that games (and some other things too, like the internet and software development methodologies) do not train creativity in finding solutions to problems, but rather train iterative processes to solving problems. “Thinking outside the box” is usually typified by lateral thinking, applying methods and finding solutions that sort of come out of left field; it’s not a process that benefits from analysis or iterative processes, as it’s mostly subconscious and benefits more from ignoring the model that you’ve been developing than by actually developing a model at all. The problem, at least here, isn’t about rare events with high magnitude, it’s that there is an assumption that everything is knowable and categorizable, and if something isn’t part of the model it’s assumed that it doesn’t exist. The model is not questioned, there’s not enough introspection about assumptions.

    We aren’t going to necessarily reverse that trend by having high magnitude but rare events, that’s not really what the problem is; we need to actually make it so that there are tangible incentives for breaking the rules of the game instead. We need to train people to expect the model to be incomplete, faulty, and abusable. To have gaps in it that you can slide through to accomplish bigger and better things. And how you do this with something that’s effectively a complete mathematical model is very difficult. It works in real life because the model is not known, at least completely. But with games the designers would need to be building systems that even they don’t truly understand the full extent of. You can do it with permutations, but the permutations have to be large in number and valid and not repeats of themselves. It’s a tricky issue.

    As an aside, while functioning in a “plan for the worst case” state of mind is beneficial in certain respects and in certain fields, it’s also a pretty terrible way to live your life. Most of the things that we find meaning in are pretty senseless if we’re talking cost/benefit analysis. So while it’s a great idea in a financial system, it’s a pretty terrible idea when applied haphazardly to everything.

  25. Interesting piece! I appreciate any attempts to spread suspicion concerning induction (which is what the black swan story is about) – Hume was more than two centuries ago, how are we so slow to accept this idea? Perhaps because we don’t want to give up our attachment to science, and science is almost entirely dependent upon induction. 🙂

    I take brief task with: “people tend to extrapolate linearly, not exponentially”

    Actually, although this is just the overall trend. If you constrain your interest to mathematically-biased intellectuals (Rational temperament/orbito-frontal cortex-bias) you’ll find a surprising tendency towards compound exponential extrapolation – that is, while the most common response is to underestimate by assuming linear progression, scientists and geeks often overestimate by assuming exponential progression. But either way, “our gut instincts about rates of change are typically wrong”, as you say. 🙂

    I think the problem here isn’t that we pick the wrong extrapolation, it’s that extrapolation by only one rubric is almost always wrong, since in the real world we experience multidimensional influences too complex to track. Isn’t that, in essence, the whole ludic fallacy in a nutshell?

    Best wishes!

  26. You’ve completely lost me. You seem to be confusing traditional AAA game pacing (which in turn parrots conventional narrative structure) with the overall limits of the medium, which are quite more vast.

    If recent generations are inculcated with any harmful assumptions it’s because they’ve been raised on big-budget games that are attempting to be movies, not on the medium of games in general.

  27. Either that or the tiny subset of gamers who apply a ruthlessly statistical outlook to everything, a subset you might have spent too much time locking horns with back in your MMO days.

  28. JP, a couple of things…

    1) The post is not about pacing; you’re pulling out one detail and misreading it as the point.

    2) All gamers apply a mathematical model to games; the difference is whether they do it consciously or not.

  29. “All gamers apply a mathematical model to games; the difference is whether they do it consciously or not.”

    I agree that everyone applies a model (to everything!), but calling it mathematical even if the viewer/player/listener experiences it as highly sensational/emotional/sequential sounds like it’s coming more from the perspective of an outside observer than some fundamental truth of how people engage with things. Obviously as a game designer you have to assume players will apply a “mathematical model”, especially with the Sisyphean balancing task of an MMO. I just don’t feel like that’s a very descriptive model for what’s going on in peoples’ heads when they play most kinds of games. Even something as formalist as the MDA framework acknowledges an aesthetic layer that is no less central than the mechanical.

    In short I think saying games are “just math” is akin to saying music is “just sound” or books are “just words”. Assuming prejudices or perspective limitations of the audience based on that seems harmfully reductionist.

    The conventional wisdom that says good game design has strong perceivable consequence, which you’re totally right in pointing out sits in tension with the desire to force players to deal with the unpredictable, is just that, conventional wisdom – a rule waiting to be broken for artistic effect rather than some immutable limitation of our medium. I’d love to see more working examples of that idea.

  30. Raph, would it make sense to approach this by taking the black swan event concept and applying it to the *mechanics* instead of the gameworld? As in, at some point in the game, the old method of doing things simply fails completely, forcing players to reach for something else.

    You wouldn’t necessarily want to do this by denying access to existing verbs or skills the players have developed, but by changing the way those verbs or skills are applied, and making the previous method of application no longer valid. Could be kinda tricky to make the feedback for this look natural enough, but it could promote more creative uses of abilities while demonstrating that sometimes things change abruptly.

  31. I agree that everyone applies a model (to everything!), but calling it mathematical even if the viewer/player/listener experiences it as highly sensational/emotional/sequential sounds like it’s coming more from the perspective of an outside observer than some fundamental truth of how people engage with things.

    The issue isn’t how they engage with things in general — it is how they engage with games in specific. And at their core, games are about math. This doesn’t mean the aesthetic layer isn’t incredibly important — it is! — but you can have the game with minimal aesthetics, and call it a game, and the reverse is not true.

    Eolirin, you could, but if it happens simply narratively, you are not necessarily teaching the underlying lesson about black swans… you’re once again displaying a linear evolution rather than a discontinuous one.

  32. “And at their core, games are about math.”

    What I was saying is, they’re only “about math” to game designers, and some kinds of players. Many kinds of modern games (Passage, Mirror’s Edge, to grab two random examples) invite players to engage first and foremost with their aesthetic layer. Many players of those games take that offer.

    Again, I’m not arguing that any game cannot be dealt with as being “about math”… I’m saying that many aren’t in practice.

  33. @Raph, It’d have to be done to the model of course. You’d want the model itself to change and abruptly so. Say, you were able to defeat the enemy using a certain tactic, and it’s going well, and then all of a sudden, use of that tactic becomes a losing state. Not just less effective, but actually countered directly. No narrative warning ahead of time, though narrative explanations may need to be there at some point, because you don’t want it to seem too arbitrary… but the model that was being used for overcoming the mechanical challenge just plain ceases to function all at once. This doesn’t necessarily help with teaching everything that’s missing (more focus on creative solutions isn’t necessarily implied in the above, and I think we need that), but as part of a combination of things it could go a long way.

    @JP, even if it’s only unconsciously, the way that a player interacts with a game like Passage or Mirror’s Edge is basically to absorb the mathematically model and apply it. It doesn’t matter if the reason they’re doing this is for the aesthetics, the only actions they can take have to do with the model, so in order for them to do anything at all, they need to be taking the model into account. Even if the model is really simple and easy to master and the greater purpose is the aesthetic layer, if you were to remove the math you’d have nothing at all.

    So at the core, it’s about math. Anything else is layered on top. Aesthetic gives meaning to the math, and meaning is perhaps more important than the underlaying model. But it still exists as a function of the math.

  34. I’m surprised you don’t think Black Swans are common. They seem like a fairly standard game (and plot) device to me. Perhaps I am interpreting your meaning incorrectly. You teach a player some skills in one context, then significantly change the context/ruleset to force them to adapt their skills to the new environment. It can also be used to force them to explore skills that were available but little used in normal play, for whatever reason. For example, Halo 1 seemed on the first play-through to be entirely about fighting Covenant. The marketing of the game focused on it, and the first ten hours of game play reinforced it. As soon as the player was comfortable fighting Covenant, the Flood arrive, requiring new tactics using the same tool set. Ditto games where the avatar looses all items or skills part way through the game (most recently, Fable2).

  35. One more thing: Perceiving games as math is no different from perceiving life as math. Min-maxing your retirement savings or grooming your kids for law school are similar behaviors. Those folks are just ‘playing the game.’

  36. Is a previously unencountered capability in an antagonist sufficiently ‘non-linear’? For example, in sea battles, aircraft changed the playbook entirely by changing the value, pace and order of maneuvering as well as the kinds and types of ships. Radar changed the battle strategy from one of putting ships next to one another into finding the fleet first then maneuvering. The difference is when both sides have aircraft, therefore scouts, and one side has radar and the other does not.

    Or would disruption only be considered a black swan if both had sailing ships and one of them had radar?

    I have some trouble discerning what is sufficiently disruptive as to constitute a black swan that is discontinuous.

  37. […] I ran across this entry at Raph Koster’s blog: Recently I had a discussion with a management and leadership consultant, and we were discussing […]

  38. […] Nicholas Taleb’s The Black Swan, you better grab the book now. He advanced the notion of ludic fallacy which is so prevalent in education through standardized tests, the gamification of learning through […]

Sorry, the comment form is closed at this time.