Superstition
B.F. Skinner is well-known for his theory of behavioral conditioning, but one of his quirkiest studies involved inducing superstition in pigeons (1948). 8 pigeons were placed in a reinforcement contraption (i.e., Skinner Box) and were given a food pellet every 15 seconds no matter what they did. After several days, each pigeon had fixated on a particular superstitious behavior. One pigeon danced counter-clockwise, another two developed a left-to-right head-swinging motion, another attacked an invisible object in the top right corner of the cage, and so forth. This phenomenon has also been replicated among high-school students (Bruner & Revuski, 1961). And given that MMOs are a kind of Skinner Box that offer some random rewards (e.g., rare drops), it's not surprising that superstitious behaviors emerge in MMOs as well.
During our weekly meeting at PARC, we stumbled upon the issue of
superstition in MMOs. Cabell Gathman mentioned that the current "trick
or treat" event in City of Heroes/Villains has some players convinced
that there are ways to increase the odds of a "treat" when knocking on doors. Eric then
mentioned that in his WoW guild, some believe that it's best for
Hunters to be the first to enter an instanced dungeon because Hunters
supposedly have better loot tables. I was also reminded of Raph Koster's account
of the gibberish ghost language in Ultima Online and how many players
were convinced that ghost language could be deciphered and that ghosts answered player's questions.
But the incident that stuck out the most in my mind was the amount and intensity of speculation in Star Wars Galaxies as to how someone unlocked a Jedi character (before the anti-climactic truth was revealed). As in the case of the pigeons, what spurred these speculations was that the game showed system messages to certain players that they were one step closer to the truth without telling them what it was they had done that brought them closer. The sheer creativity and conviction that people had towards the main competing theories was mind-boggling. There was a very well thought out theory involving completion of a chain of specific NPC quests; another involved describing the different requirements for different professions; others focused on unique geographical landmarks that had to be visited. I thought this was actually the most interesting period in the SWG universe because of the player productivity. Unfortunately, the truth (i.e., grind levels, abandon profession, grind again, etc.) was much more banal than any of the main theories.
What's clear is that it's not that hard to create superstitions among MMO players. Several questions come to mind:
- Is encouraging superstition a good thing for an MMO (i.e., leads to player productivity and player-generated content)?
- What is the most interesting case of superstition you have seen?
My favorite is still the stuff that surrounded the wisps in Ultima Online. I wrote about it in a blog post about NPCs:
FWIW, I tend to agree on the productivity of that period when nobody knew how to be a Jedi. I think giving the clues was a mistake -- not only because it revealed the system which did not live up to expectations, but also because it commodified what was mysterious. Ideally, people would have reached Jedi without having any idea what they had done.
Posted by: Raph | Oct 31, 2006 at 02:41
Does Mr. Bigglesworth help you get more Epics from Naxx trash?
Posted by: Wolfe | Oct 31, 2006 at 05:27
Nick Yee: Eric then mentioned that in his WoW guild, some believe that it's best for Hunters to be the first to enter an instanced dungeon because Hunters supposedly have better loot tables
I am not sure if calling these things "superstition" is fair. Players understand that code can be messy and that code can lead to strange behaviour. In fact, code can lead to anything.
In LP-MUDs where the code is available to the wizards, players might call on one to get the code examined for bugs or to find out if his theory is bonkers.
Raph Koster: deally, people would have reached Jedi without having any idea what they had done.
If random() > player.playtime Then player.make_jedi()
Posted by: Ola Fosheim Grøstad | Oct 31, 2006 at 06:22
Ack... invert that test. *blushes*
Posted by: Ola Fosheim Grøstad | Oct 31, 2006 at 07:28
I would agree with Ola that "superstition" may not be a fair term to use.
I think referencing "urban legends" is better as (1) it's about something that's not proven, (2) it captures the receiver's imagination and (3) the receiver develop their own version of "the truth" of the matter.
Given that these speculations and "urban legends" do capture the imagination of players, it should be considered for proper inclusion in MMOs.
I think the current crop of TV shows such as Lost, Heroes, X-files, etc. uses these techniques to keep viewers engage.
So rather referencing "superstitution" (which is catchy as a phrase), I think a better reference is "urban legends" or actions based on guesses, which occurs often between the time when a quest is added and when the actions necessary to solved the quest is successfully repeated.
Frank
Posted by: magicback (Frank) | Oct 31, 2006 at 07:44
Better question: Is superstition merely our attempt to make a reasonable guess about the algorithm of the universe?
Posted by: EdMcGon | Oct 31, 2006 at 08:00
"A Superstition is the irrational belief that future events are influenced by specific behaviors, without having a causal relationship."
The thing about behaviour is that it is always related to a cause. When Skinner ran his experiments, his pigeons were in cages. They were being observed, but were also able to observe.
So there's every chance that the behaviour of the pigeons was pre-empted by a corresponding response from the observers, anticipating the pigeons "superstitious" response. A response which although unrelated to the food dropping event, may have been a rational response given the limited data available.
In WoW, there have been issues raised concerning raid loot generation, given statistically unlikely sequences of loot being dropped for the same raid leader.
IIRC, the development team acknowledged that originally loot generation was seeded (at least partially) by the raid leaders name. Other factors were also involved. I wonder if one of them was class/race?
I'm not a mathematician (unless it's my round) but I've run into enough real life statistical situations where convergent patterns emerge from seemingly random initial data, so this isn't particularly far fetched.
Do I think that Hunters get better loot tables as raid leaders? No. What I do think is that people are extraordinarily bad at detecting causative factors.
As further thought, consider the role of hidden mechanics as part of gameplay folklore. In the real world, we have Masons, witchcraft, Select Committees and other such quasi-mythological entities, knowledge of which is alleged to have significant impact on our daily lives. The important point is that whilst their effects may be largely overstated or imagined, the above entities all have an ability to impact events.
The direct allegory to this is cheat codes and game mechanics. Games are inherently comprised of codes. Understanding and manipulating these codes is an inherent part of gameplay and gaming. Game culture is full of references to both real and imagined codes that convey significant benefits to the able cryptologist.
I'm currently one of a handful of people who currently stack their character with +weaponskill items. Few other players understand how the crushing blow mechanic works in WoW and there's no official literature in or out of game (short of delving into the ever changing, ever censored WoW forums).
If a person looked at my character to derive an idea of how I was doing so much damage, they would first notice the amount of +hit items my character has, since I only have weaponskill on one or two item slots at a time, by comparison to +hit on every other.
That player, if then trying to ape my damage output, would then attempt to replicate my +hit itemisation and failing to duplicate my output, would probably ascribe the significant shortfall to some theoretical "+hit cap" that I have managed to reach.
Yes. This is a belief that an event is influenced by a non-causative factor.
The decision making behind it, however, is perfectly rational.
*Sorry about quality of language, I'm writing this in fragments whilst I beaver away in a call centre. Hope it makes sense.*
Posted by: Gareth Eckley | Oct 31, 2006 at 08:02
First, I am glad to see Skinner's study referenced as I found it quite informative. I disagree with the claim that the pigeons were reacting to some other factor. The pigeons merely suffer from the same problem humans do: An assumption that the universe is governed by predictable and understandable rules.
When we encounter something that is random, or over which we have no influence, we are loathe to come that conclusion. We'd rather build an increasingly complex set of hypotheses that fit the observed data so far.
Second, to the Game Design question. Is superstition a good thing? YES! Superstition is the game equivalent of "reading the book vs seeing the movie". Superstition is the player projecting complexity onto your game mechanics that isn't there. Superstition is, in effect, the player playing a *better* game than you wrote. Indeed, sometimes the game that they think they are playing is not even possible for you to write due to technical limitations.
What I think a good game designer should do, however, is troll the forums silently looking for superstitions. Then, where possible, they should secretly change the code so the superstition becomes true. Ie, say a lot of people are claiming that by sacrificing gold in a certain ritual at a shrine increases their luck attribute. Perhaps one should then add code so that actually occurs.
There are two ways to find out how to make a game more worldly. One is to wait for someone to try something that doesn't work because it isn't implemented. You then know that you should implement that feature. The second is for someone to try something that doesn't work but that they don't realize doesn't work. In this case, you get the benefit of implementing the feature without having to admit that you never thought of the possibility.
Posted by: Brask Mumei | Oct 31, 2006 at 09:17
There's always Malinowski's functionalist analysis of the supernatural in Magic, Science, and Religion. I don't have a copy at hand, so I've had to Google up some quotes. To wit:
This
"The classic study in this area was Malinowski's (1925 - reprinted 1948) analysis of Trobiand Island fishermen who combined magic with "scientific" knowledge - both deployed to enhance food-gathering activities. Malinowski lived with the islanders he observed and felt that when events couldn't be explained "scientifically" magic or ritual was deployed to reduce uncertainty. Support for this notion was found in what he felt were the different orientations of the inshore (in the lagoon) as compared to the more dangerous offshore fisheries. Offshore fishermen performed more elaborate rituals and were more deeply superstitious than those who worked in the calmer waters of the lagoon. Wherever the outcome was more uncertain, the greater the need for magic. Hence "we do not find magic wherever the pursuit is certain, reliable and well under the control of rational methods and technological processes. Further, we find magic where the element of danger is conspicuous" (1948, p. 139-140) [...] Malinowski's was a functionalist analysis wherein superstition served to reduce anxiety where there is uncertainty and a risky situation."
That
Malinowski, in Magic, Science, and Religion and Other Essays, pointed out that Trobriand islanders—far from living in a perpetual fog of magical thought—hunted and gardened with empirically-honed skill; they only turned to magic when they reached the limits of their practical knowledge. Evans-Pritchard, in Witchcraft, Oracles and Magic Among the Azande, argued that a belief in witchcraft did not preclude the Azande from understanding empirical relationships in the world around them and that their witchcraft beliefs had an underlying logic and plausibility that impeded falsification. While there were reflexive implications to this work, Westerners too turned to magic and religion when they felt powerless."
Posted by: Sarapen | Oct 31, 2006 at 09:22
This last part was cut off for some reason.
Posted by: Sarapen | Oct 31, 2006 at 09:23
What's with this? Okay, the quote I was trying to put in:
We find magic wherever the elements of chance and accident, and the emotional play between hope and fear have a wide and extensive range. We do not find magic wherever the pursuit is certain, reliable, and well under the control of rational methods and technological processes. Further, we find magic where the element of danger is conspicuous.
http://atheism.about.com/library/glossary/general/bldef_malinowskibronislaw.htm
Posted by: Sarapen | Oct 31, 2006 at 09:24
What is interesting for developers is whether they should add random code, or code that follows complex patterns, to enrich their world. I worked on an events team for a couple years, and the best stories we came up with were ones that offered a bare minimum of information, and gave the players a chance to puzzle out the rest. One of our events involved a "scrambled" communications that was actually us hammering on our keyboards randomly. The producer told us to follow grammar and linguistic patterns similar to real language, then we would add in occasional references to what the players said. The result was months of puzzling over what we said, and several people claiming to have broken the "code." It would be interesting to have something similar within the game itself, a mechanism that would randomize rewards based upon changing parameters.
I'm a believer that players want to understand what's going on in a virtual world, but don't want to be fed it on a plate. There should always be something they don't understand, so they're never "finished" understanding how it works. One way is by continually adding new content you've developed for the game; another is to add some random sequences that can be puzzled over by players; the third (and best in my opinion) is to let players create dynamic content, so they continually innovate in the world.
Posted by: Gabriel | Oct 31, 2006 at 09:26
I second that, Gabriel. This is easier to achieve in single-instance worlds though, where you can walk in lock-step with players as they progress. Albeit somewhat expensive in terms of human resources.
Posted by: Ola Fosheim Grøstad | Oct 31, 2006 at 09:41
It reminds me a bit of "scuttlebutt" in the military or in other similar campy settings (college dorms, airplanes?, concerts/events?). I think it comes from a mass of similarly situated people who have some feel for how decisions are made but are essentially out of the loop. Rumors then become incredibly salient, with the lack of hard information... there's a model here somewhere.
Posted by: patrick | Oct 31, 2006 at 09:53
I stayed up late at night to play my crafting characters in UO, because I'd heard that skill gain was based on the number of characters currently online with that skill. I also ate a lot of fish steaks, because everybody knew that skill gain was improved when your character was full.
(Were those true?)
In WoW, there have been issues raised concerning raid loot generation, given statistically unlikely sequences of loot being dropped for the same raid leader.
Are they actually looking at hard data with reasonable sample sizes? Random item generation makes people think funny things.
On the last live game I worked on, every once in a while, players would start swearing up and down that we'd changed one of the game's random item generation systems. One person would seed the idea on the boards, and it'd take off like wildfire. It felt like every single player who ever posted on the boards was 100% certain that I was lying to them when I said the system wasn't changed.
It was true. That system hadn't been touched in years. And a few weeks after it started, the players would stop talking about it. Even though they let it die, I suspect they held a grudge. It was all because one asshole started a rumor, and people are willing to believe anything about random item generation.
Posted by: Sara Jensen | Oct 31, 2006 at 10:37
Online Casino, Online Casino Games,Free Online Casino Games,Real Vegas Cnline Casino,Online Casino Reviews,Online Casino Gambling,Online Casino Blackjack,Online Casino Bonus,Royal Vegas Cnline Casino,all slots Online Casino,Grand Hotel Online Casino,Golden Palace Online Casino,Club Player Online Casino,Las Vegas Online Casino,Online Casino News,River Belle Online Casino,Carnival Online Casino,Platinum Online Casino, The Gaming Club Online Casino,Top Online Casino,Online Casino Free Bonus,
Posted by: Online Casino | Oct 31, 2006 at 10:47
Actually, as I recall the Skinner superstition study is not entirely straightforward. I seem to remember that attempts to replicate it in the 90s indicated that it's not just a case of "randomness produces superstition" there was another factor or two. The overall verdict, from a student of Skinner's, was that Skinner kind of fooled himself on this one. I'll see if I can dig up the later study.
When it comes to players, I think that superstition isn't the problem, it's unlearning superstitions. It's perfectly fine if players decide that a hunter has to enter an instance first, but it's not ok for players to avoid running instances if they don't have a hunter. If a significant portion of players are consistently doing something that makes the game less fun, it's the designer's problem and needs a fix.
It's very difficult to train someone out of a behavior which is non-functional but which they believe is associated with a reward, simply because they don't want to take the "risk" of skipping the behavior. You can't reward players for not doing the superstition, so you have to wait for the lack of reward to extinguish the behavior, which can take quite a long time and may extinguish interest in the main task.
Posted by: John Hopson | Oct 31, 2006 at 11:08
Some superstitions have a basis in fact.
I can't help but be reminded of the situation with the monk intimidate skill in the original EQ. One or two people were able to exhibit the skill, and nobody else was.. There was a huge volume of speculation about secret trainers and quests to get the skill. Later there was a backlash in the other direction with claims that someone who provide screenshots of skill in action has fabricated them.
Eventually it came out that the skill was real, was supposed to be available from regular trainers, and that GMs thought it was in place because characters who had their level changed artifically had the skill correctly unlocked. Later it was fixed to be available from standard trainers and I doubt if many people even recall the incident now.
Its too bad they didn't do what Brask suggested and make the code fit the player mythology.
Posted by: Thabor | Oct 31, 2006 at 11:21
I stayed up late at night to play my crafting characters in UO, because I'd heard that skill gain was based on the number of characters currently online with that skill. I also ate a lot of fish steaks, because everybody knew that skill gain was improved when your character was full.
(Were those true?)
The first one was based on truth, but was false. the chance of gaining a skill was based on the overall frequency of skill usage for that skill. Skills that were used infrequently by the playerbase as a whole would go up faster than skills that were used constantly. The reason for the origin of this system is that something like swordsmanship was queried really often and quickly, and something like spirit speak was queried really infrequently. Were there no table to adjust the advancement rates, you'd max out swords in far less time. Overall across all players, the total time to 100 in each skill was exactly the same.
Unfortunately, the system wasn't baselined somehow, so it spun into a feedback loop, making everything slower and slower and slower. Explanations of how the system worked led people to think that playing when no one was on would mean that everything would be faster. This was erroneous.
The fish steaks thing was complete hokum. No idea where that came from.
Posted by: Raph | Oct 31, 2006 at 11:23
My favorite MMO superstition was the Wi Flag. Found some history here:
http://www.vitaerising.com/modules.php?op=modload&name=News&file=article&sid=40&mode=thread&order=0&thold=0
The fascinating thing about the Wi Flag is that it was rooted in truth:
"Normally--if we had rolled between 0 and 3 in the example--your order in the list should have no effect on how likely you are to be chosen. But because we only rolled between 0 and 1, the earlier you appear in the list, the more skewed your chance of selection is. And as it happens, in AC code, your position in this list is determined by the InstanceID of your character, which is assigned when you create the character and never changed."
And when your wildest superstitions turn out to be based in fact, it's hard to argue in the future that rotating counterclockwise doesn't have some affect on the seed supply.
Posted by: Axecleaver | Oct 31, 2006 at 11:53
In Dark Age of Camelot, crafting skill-ups were essentially random, as was the quality of the final item. Statistically, the odds were proven; it was basically a 1/100 chance for a perfect 100% quality item, a 1/50 chance for a 99% item, etc. (I don't remember the numbers exactly, but you get the idea). Creating "masterpiece" items was just a random event. The more you tried, the more chances you had.
The DAOC crafting boards were laden with threads discussing the various ways to get the most out of crafting. There were a plethora of theories, most of them abominable. For example, make lots of low-level items (thereby saving money on retries), then switch to making high-level items when you get close to 100 attempts to cash in on the "inevitable" masterpiece that the system owes you. People would become angry if their masterpiece "came early", i.e. while they were still crafting useless low-level items because then their masterpiece chance was "used up" and they'd have to start all over again.
It was also amusing to hear crafters arguing with one another about "stealing" masterpieces. The thinking went: if I do a lot of crafting and get a lot of masterpieces, then I've won the masterpiece item, and you therefore won't make any. It was thought the system had a finite number of masterpieces in it, and if I created them all, then they wouldn't be available for you to create.
No matter how concisely or how clearly one tried to explain random independent events to these folks, they KNEW FOR A FACT this was the way crafting works.
Posted by: Informis | Oct 31, 2006 at 12:53
Well, they aren't superstitions, for one thing. They are patterns of behaviour that work, just like in real life. For example, if I ever have a parcel not renting in Second Life, I know how to ensure that it will rent instantly -- I first have to spend an hour building or assembling a house on it, then it will rent, and five minutes later, the customer will ask to have the prefab removed to put out their own. Works like a charm, but I try to ration the charm's use.
Posted by: Prokofy Neva | Oct 31, 2006 at 12:57
That got me wondering what an "anti-climatic truth" might be. Meteor impact? I used to giggle every time I saw the Professor's city "Climactic Research" in Sid Meier's Alpha Centauri.
Anyway, the fundamental problem is that not a lot of game players know how probability works. If you try to organize people to collect data on how a particular game feature works, they immediately launch off into anecdotal evidence. Some people are happier with myths than facts, and I guess that is something a game designer or a politician has to accept and take into account.
Posted by: CherryBomb | Oct 31, 2006 at 13:53
Everquest had tons of this stuff, too. People were convinced that mob respawn was affected by whether you fully looted the corpses of previously killed mobs (it wasn't). People were also convinced that if you stood too close to the mob spawn point it wouldn't spawn (it would). After every patch, people were convinced that they'd nerfed the run speed boost of J-boots (they hadn't).
It all comes down to a game like EQ having really inscrutable mechanics and humans being terrible at pattern recognition and understanding of randomness.
Posted by: Brent Michael Krupp | Oct 31, 2006 at 14:20
Heh - Thanks for catching the typo CherryBomb. I've gone ahead and corrected it. I guess the opposite typo is more embarassing :)
Posted by: Nick Yee | Oct 31, 2006 at 14:30
first of all, i think "speculation" is a better term to use than "superstition".
players speculated on what would get them a Jedi.
(by the way, putting player Jedi into SWG was THE single biggest mistake, which ultimately killed the game... jus'sayin')
one of my biggest gripes with MMOs is all the Math involved.
it seems that in most MMOs everything's about numbers. every skill, every item, has numbers attached. most players know most of the formulas which go into the game's mechanics.
everything is right out there in the open!
this takes away so much of the mystery of how things work.
everything is broken down into formulas. number crunching reigns supreme.
i think that making MMOs which promote speculation amongst players is a very good thing.
as was seen in SWG, player's came up with some fairly imaginative ideas on how things worked.
if you remove the visible math from MMOs, players are left with vague ideas.
ideas are malliable (in the players minds).
math is not.
those vague ideas will become imaginative ideas.
people aren't used to seeing all of the underlying mechanics to their World.
we've just got vague (varying degrees of vagueness) ideas about how things work.
instead of: "this sword has 12 damage (+3 vs. Orcs)."
how about: "this sword is well crafted and sharpened, and coated with a subtance known to be poisonous to Orcs."
instead of: "i need 12,000 more Archery XP to make my next level, and earn the 'Faster Aiming' skill, which increases my firing rate by 5%."
how about: "i've been practicing with this bow and arrows for four days now. i hope that the next time i speak with my trainer, he will notice my improved abilities, and maybe teach me a new and helpful technique."
i think vague ideas and speculation would make for a far more interesting MMO, than another number crunching grind.
Posted by: Kohs | Oct 31, 2006 at 14:30
I think the term "speculation" is fine for the situations where people are indeed just theorizing on a message board, but I think when you have people repeatedly performing an action in the belief that the action causes a specific event outcome (and in the face of many non-confirming outcomes), then it's fair to call it "superstition". For example, in the case of obsessive looting of mobs in EQ to "help" respawns, that feels more than just a speculation.
Now, if they were systematically testing this (i.e., experimentally), then that would be hypothesis testing, but in most of these cases, they aren't testing the speculation. They just do it without any good evidence because they are convinced it does happen.
Posted by: Nick Yee | Oct 31, 2006 at 14:52
The point to temember about environments with complex contingency, like MMOGs, virtual worlds, and everyday life, is presicely that statistical risk management approaches break down in them. Humans in them therefore rely on something which, all things considered, they are quite good at -- developing a reliable disposition of how to act. Sure, sometimes this generates laughable "guesses", but this is neither "anecdotalism" nor "superstition." It is an open-ended readiness to learn, always in the process of becoming.
Posted by: Thomas Malaby | Oct 31, 2006 at 14:53
fair enough.
one of my favorite superstitions from real life, is when Major League pitchers hop over the foul lines on their way to the dugout, thinking that stepping on said lines causes them to pitch poorly.
for some reason, the actions players take in MMOs have never seemed comparable to that. or at least haven't seemed to be done in the same "spirit" as something like that.
i suppose they are done so afterall.
i still say that ADDING to the superstitions and speculations (by removing as much visible mechanics as possible) will make for a much more rich and colorful world.
you'll have players saying things like "if you visit the Blue Forest at night, you'll become much stronger the next day."
or "never cast the fireball spell while in the presence of a star-crystal. it will backfire!"
and statements like those could be based on kernels of truth (which are in the now-hidden mechanics), or they could be completely false, and based on the experiences of a few talkative players.
you could even have players psyching each other out, or using trickery.
for instance, if one group of players possesses the so-called "star-crystal", they could spread completely unfounded rumors about its effect on the "fireball" spell.
thereby causing their opponents to neglect using that spell in battle, giving the advantage to the group who spread the rumor.
that may be a far-fetched scenario, but it would be possible in an MMO which encouraged superstition and speculation by hiding the mechanics.
Posted by: Kohs | Oct 31, 2006 at 15:17
Nick Yee: in most of these cases, they aren't testing the speculation. They just do it without any good evidence because they are convinced it does happen.
The problem with code is that they cannot reliably test their hypotheses/thories anyway, unless they understand the premises for the code. In order to predict they need data, but developers can prevent sufficient data from sipping out by using an odd distribution (e.g. brownian motion) or a wide range of nonintuitive variables as parameters. Random drops doesn't have to follow a white noise pattern... Systematic testing presumes a "somewhat deterministic" model.
Posted by: Ola Fosheim Grøstad | Oct 31, 2006 at 15:46
Or to phrase it differently: the only reasonable model players have access to is the model they can obtain from reasoning about what kind of programs a programmer would bother to write.
The physical world is different: we can inspect the underlying mechanisms. Variables usually leave visible traces. The physics of the universe provide constraints. Code is inherently different.
Posted by: Ola Fosheim Grøstad | Oct 31, 2006 at 15:59
"Variables usually leave visible traces."
Not just visible traces, but obvious, in-your-face traces like health bars, and new abilities that literally offer you "+3% chance to hit." What if, even though it may be based on numbers, your MMORPG offered no numeric feedback to the player? This would include gauges, life meters, and progress bars. Coming up with meaningful feedback would be a challenge -- and a possibily prohibitive one, at that, as you try to invent a metaphor for every game mechanic -- but I bet it would foster superstition.
Posted by: Informis | Oct 31, 2006 at 16:28
I have a friend who swears up and down that in UO, the random chance of your fizzling a spell or resisting one was seeded by your location in the world (your tile). Therefore, if you were getting a lot of fizzles, you should shift a few tiles to get a new seed.
I heard that theory as well when I played, but it always sounded very urban-legendy. My friend played a lot more than me (ahem), so I deferred to his opinion on it, but I have always been curious whether it was really true.
Incidentally, this same friend thinks that most MMOGs use a similar system now, so if he is wrong, he may be considered a poster child for the Skinner's Pigeons theory.
Posted by: Jayce | Oct 31, 2006 at 16:39
Informis:but I bet it would foster superstition.
Yes, but my point is more that the idea of "superstition" only holds under the assumption that the implemented models are dead simple, accessible and predictable.
The theoretical point is this: if the computerprogram is unknown, then the program's output should be considered to be arbitrary. Under these conditions MOST models should be considered to be _equally_ probable. Hence "superstition" is a meaningless term from a purely theoretical view.
Granted, developers encourage players to develop certain models over others, but can you trust developers to be "nice"?
It is also known from psychology that human beings tends to favour evidence that support their existing believes, and ignore contradictory evidence. This is a good thing too, otherwise we would never be able to agree on anything... I think "superstition" is a misnomer, unless they have solid evidence against their beliefs. I consider this type of behaviour to be "pragmatic rationality". Some people favour insanely complex models, others prefer very simple models.
Posted by: Ola Fosheim Grøstad | Oct 31, 2006 at 17:12
Indeed, Ola. This is also talked about by some as a "reliabilist" epistemology.
Posted by: Thomas Malaby | Oct 31, 2006 at 18:34
I have the same experience as Sara Jensen above with Meridian 59; even though we haven't patched for months, people swear something changed or some secret nerf went in.
There are also cases where it's not mere superstition. There was a case in M59 where you could become invulnerable if you put on and took off a certain set of equipment. People thought it was some former programmer's easter egg. In reality, it was a particular piece of equipment not resetting the resists for a weapon type correctly when it was removed. So, the main part of the sequence was actually putting on and taking off a particular piece of equipment.
Personally, I wonder how much of this "superstition" is players wanting to feel some control over something that is almost entirely random, and therefore out of their control. Nobody likes to feel that they are completely out of control in a situation.
An offline example is when I play console RPGs, I usually tap the button on the controller when there is no option during combat. As a kid, it seemed that the game would work better when I did that. Now it's just a habit I have that fills the time while my character does its 30 second animation for its attack. But, I think part of me wanted to have more control over the combat between the meaningful choices. On the other hand, this behavior gets me in trouble in the Mario RPGs where tapping the button at a specific time is important. But, you do have more control over the situation between choosing your attacks.
Some thoughts,
Posted by: Brian 'Psychochild' Green | Oct 31, 2006 at 18:39
Not exactly; the chance of advancing in a skill was seeded by your location.
See "Use-based systems" for further details.
Posted by: Raph | Oct 31, 2006 at 18:59
Whoops, that was in reply to Jayce.
Posted by: Raph | Oct 31, 2006 at 19:00
I do remember 8x8, but I had moved on by that point too :) I first heard the fizzle/resist rumor well before 8x8, during the T2A era. Interesting to have it debunked once for all.
On topic, I think Psychochild is onto something with his statement about people wanting a feeling of control. Corollarily to the law mentioned in Raph's use-based system article, people will figure out every algorithm, etc in your world - but they will also "figure out" some that don't actually exist.
Posted by: Jayce | Oct 31, 2006 at 21:39
> i still say that ADDING to the superstitions and speculations
> by removing as much visible mechanics as possible)
> will make for a much more rich and colorful world.
Heh. Design approaches like this are clearly covered under Rule 1 - "The Professional Game Designers Trap", that is; creating a game that *you* want to play, rather than creating the game that *players* would like to play.
There are many types of players who will not like such ambiguity at all (I suspect about ~25%), and who instead will set aside their willing-suspension-of-disbelief in order to min/max stats and performance. I don't happen to be one of those players personally (story, ambience and setting are important to me), but I know I must make games that keep the rest of the players happy too. If we create a game that hides cause-and-effect too deeply (which is all too easy when we're dealing with pixels on a screen and not real world senses) we give rise to a form of aberrant-conditioning... which will not create a pleasant user experience. Downplaying the numbers-game should be important, but not at the risk of giving mixed messages about progress to players. Sigh... would that our technical ability to convey the richness of our worlds had kept pace with our designs; then we'd see some great things emerge!
Posted by: Michael Steele | Oct 31, 2006 at 21:43
actually, i keep that in mind, and try to avoid that as much as possible.
if your definition of "players" is "those who already currently play MMOs, and have been doing so for many years", then yes, you would do well to keep the math and math-based feedback visible at all times.
but if your definition of "players" is "any and all human beings", then you've got a lot more room to work with.
of course i'm not suggesting that ALL feedback is obscured to the point of confusing and bewildering players, so that they can't figure out how anything works.
there's no need to give mixed messages.
just more meaningful messages, without hard numbers to serve as evidence.
you're obviously not going to decrease a players XP as they practice something.
even the dumbest player can wrap their brain around the concept of "the more you do something, the better you get at it".
so why show them the XP bar?
all that does is make their experience predictable (meaning, they KNOW exactly when they'll achieve their next skill or level).
and while humans enjoy a sense of predictability, i believe they enjoy surprises more (at least the good surprises).
as it is right now, MMOs are practically ALL hard numbers as very visible evidence.
so we can only take it in the other direction.
and i believe we need to, in order to capture the players from that second definition i offered. because very few human beings enjoy number crunching and min/maxing.
Posted by: Kohs | Oct 31, 2006 at 22:15
"The physical world is different: we can inspect the underlying mechanisms. Variables usually leave visible traces. The physics of the universe provide constraints. Code is inherently different."
Strong assertions, all problematic. I think my key objection is that, unlike the rest of us, Grøstad's experience of the universe seems to be dominated by basic physics. But the universe I live in is dominated by human behavior. The directly relevant "underlying mechanisms" in this universe are much less available for inspection. Most variables are not only "invisible", some believe they are in principle unavailable for "inspection".
Furthermore, the assumption of the regularity of the universe is an assumption. It's a necessary assumption, yes. But then it's also a necessary assumption for anyone existing in a virtual world. And while we can never hope to know that the real universe is regular, as it happens we can know that a virtual universe is regular.
If anything, it seems to me that the direct opposite to Grøstad's argument is more defensible: a gaming universe is more comprehensible than the real universe, its mechanics less obscured and more likely to be known to be regularized.
Posted by: Keith M Ellis | Oct 31, 2006 at 22:28
Well, in that case Keith, simulation of the physical world becomes an impossibility
Posted by: Ola Fosheim Grøstad | Nov 01, 2006 at 03:25
"Well, in that case Keith, simulation of the physical world becomes an impossibility"
You need to fill-out these statements of yours more. I don't know what you're getting at. On its face, this statement is both true and false, depending upon how "simulation" is interpreted. In theory, at least as it seems to me off the top of my head, you can simulate reality to any degree of detail excepting complete detail. The level of detail which is practically possible is another matter entirely. And of course you have to assume the universe is regular.
I'm not sure how you intend this to relate to our argument, but I'll just wish you luck simulating me to any level of detail beyond the most superficial and for purposes beyond the most trivial. If your original statement about the comprehensibility of the real universe was applicable to people—and assuming that you meant your statement to apply in the practical realm (which it seems to me you must have intended for it to have any relevancy in this discussion) then the problem of AI would be relatively solvable and already solved. But it is relatively insolvable and emphatically unsolved.
But that's only the most extreme example. I can only guess (well, I could do a web search on your name to dig a bit deeper), but it seems like you must be relatively unaware of two histories essential to your point: the history of science and the history of simulation in science. Simulation of even basic Newtonian physics is hard. Need I remind you that even something as exquisitely simplistic as the newtonian three-body problem is, as a practical matter, unsolvable?
And the argument can be turned on its head, too. The very reason that simulation has become an indispensable tool in the modern physicist's toolbox is because simulation allows a selective ignorance of variables that make theoreticaly models unworkable while, at the same time, allowing a level of complexity with regard to selected variables and the model itself such that useful results are generated. Which, by the way, is really nothing qualitatively different than what expressing a theory mathematically is doing.
There is a deep irony involved in the fact that people almost universally think of phsysics, relative to the other sciences, as "difficult". The truth of the matter is that the reason physics is "difficult" is because the description of the universe it aims for is about the easiest, or most attainable, description of the universe available. Because in this sense physics is "easy", the likliehood of it being successful is high, but it still takes about as much rigor and genius and hard work as we have available to achieve that success. In comparison, all the other sciences are only slightly more advanced and rigorous versions of superstition. Thus, we think of physics as "hard".
But the universe is essentially incomprehensible if we take as our goal of comprehension something closer to the god's eye view we aim for in basic physics rather than the extremely limited and contextual comprehension we've evolved to possess. When you compare the real universe to a virtual world with your focus on what's "really" happening in the code, you've defined the rhetorical context unambiguously as being that omniscient view. But with that view as the standard, our actual comprehension of the real universe is practically nil. And with that view as the standard, or actual comprehension of a virtual world is relatively very large.
Posted by: Keith M Ellis | Nov 01, 2006 at 10:10
I wish people wouldn't say things like "simulation of the physical world becomes an impossibility" when clearly we can simulate the physical world to a very high degree of accuracy, and regularly do.
The world is simulatable to very high degress of accuracy, but at the same time we know it also contains random events (e.g. wavefunction collapses) which are either unpredictable or the information required to predict them is unmeasurable.
Posted by: Peter Clay | Nov 01, 2006 at 10:23
That seems to be pervasive throughout software. QA people seem to evidence this almost as much as regular users. On occasion even designers are vulnerable to that. I think "superstitions" are actually stronger when they occur on development teams, because they have a much stronger belief that the system will meet their expectations..
Posted by: Thabor | Nov 01, 2006 at 10:59
"...when clearly we can simulate the physical world to a very high degree of accuracy, and regularly do."
Well, no, we don't. We simulate to a very high degree of accuracy carefully selected and carefully limited portions of the real world. That seems to me to be a very long way from "we can simulate the physical world to a very high degree of accuracy". You can simulate me to a very high degree of accuracy if by "me" you limit yourself to "statements of similar character and content as those made by Keith M. Ellis in this blog". Which is, for select purposes, a useful simulation. No one would claim it's a simulation of me in any large sense. Neither is any simulation of the "real world" a simulation of the real world in any large sense.
Also, as best as I can parse it, Ola Fosheim Grøstad's comment is, in the context of his argument, a contrafactual. He's saying it's the consquence of my argument. It's not his argument.
Posted by: Keith M Ellis | Nov 01, 2006 at 11:05
In my eq2 guild we mentor to the lowest person in the raid prior to zoning then unmentor after all have zoned. Some think that this will lead to a better loot table if loot is determined by the server during zoning In theory, if our level is lower we are fighting more difficult mobs and will be rewarded for this.
Others argue that this is an error. Loot tables are based on levels. So if you want to loot a level 70 armor piece the raid needs to be an average level of 70.
When I played Project Entropia (now Entropia Universe) superstitions were common. Some players would not hunt when another player was on the radar. Some theorized that better loot was given to higher level players in the same area. Another superstition was the longer a mob was spawned the larger the loot so players would not kill a freshly spawned mob.
One society established themselves as a church to the god that controls loot (Church of Lootus) and players would offer up gift prior to hunts. Probably tongue-in-cheek...but why not...just in case.
Posted by: dave | Nov 01, 2006 at 11:22
"I think "superstition" is a misnomer, unless they have solid evidence against their beliefs. I consider this type of behaviour to be "pragmatic rationality". Some people favour insanely complex models, others prefer very simple models."
I agree. Superstition is just an incorrect (though not necessarily illogical or irrational) belief about how the world works. A player might arrive at a superstitious conclusion by deducing it for themselves based on what they've observed, or a more experienced player might deliver it to them pre-packaged.
And this leads back to the original question: "Is encouraging superstition a good thing for an MMO?" The presence of a superstition amongst players means they either haven't completely deconstructed your game yet, or they have and they think there's more there to discover. Both are good, but the former is clearly a better position for a developer to be in. Once (YOUR MMORPG HERE)wiki.com appears, you're nearing the point where the mysterious workings of your game have become commonplace, *understanding* is no longer a goal of your players, and your game has to survive on the merits of its game mechanics alone. In this sense, superstition is an indicator of a healthy game because you've created a game that is resistant to deconstruction.
Posted by: Todd Ogrin | Nov 01, 2006 at 11:29
To me, that's the essence of game design as an art form -- the ability to calibrate these multiple sources of the unpredictable in a way that is continually compelling for the participant. But I would add that, while many games that are designed are closed systems, in a sense (such as chess), nonetheless it is the creation of games that are so complex as to never be fully deconstructed by *anyone*, not even the designer, that are particularly impressive to me, and which characterize most of the MMOGs that we talk about here (yes, even WoW, which may be the most shallowly contingent of them all). It is the ability of a great game designer (like a great GM in pen and paper RPGs) to manage complexity on the fly, without ever controlling it utterly, that continually amazes me. It's certainly hard to do well.
Posted by: Thomas Malaby | Nov 01, 2006 at 13:26