You are one of two people in an experiment. The other person is handed $100 and has the choice of either splitting the money 50-50 with you, or taking $90 and giving you $10. You can accept or refuse the offer.
The other person takes $90 and offers you $10. Do you accept?
Chances are that you will reject the offer and thus end up with no money even when you could have had $10. This is completely irrational behavior, but you were disgusted by the behavior of the other person and wanted no part in it.
Businessweek writes about neuroeconomics:
According to the new science of neuroeconomics, the explanation might lie inside the brains of the negotiators. Not in the prefrontal cortex, where people rationally weigh pros and cons, but deep inside, where powerful emotions arise. Brain scans show that when people feel they’re being treated unfairly, a small area called the anterior insula lights up, engendering the same disgust that people get from, say, smelling a skunk. That overwhelms the deliberations of the prefrontal cortex. With primitive brain functions so powerful, it’s no wonder that economic transactions often go awry.
We can extend this to games. Players do not always play rationally, but sometimes they do.
Here’s a prediction that I have not tested:
It would be nice to empirically test this.
The first two parts of your prediction have definitely been tested, and they’re basically right. Whether this behavior is irrational is another matter. For instance, maybe people act as if they are playing a repeated game against their society rather than one-shot games against individuals? In this case, refusing small offers is (rationally) training members of society to be fair later on, at the cost of a small amount of money.
As for the second two points, shouldn’t we be able to test this without varying how experienced the players are, simply by providing one detailed and rich fictional world/fictional character and one character who is very flat and invites consideration solely as game rather than also as fiction?
See Colin Camerer’s book Behavioral Game Theory for these and other experiments, and see that link for mention of Camerer’s recent talk at Penn and some of my brief thoughts on behavioral game theory and video games.
I can’t give you a proof of your hypothesis, but here’s some anecdotal evidence:
Vampire colon The Masquerade colon Bloodlines is the first game that made me care about what characters thought about me, even though there was no real system in place to model anything like “trust” etc other than that scripted into branching dialogue/story structure (hardly very flexible).
Although I was still doing the whole “power game” thing, trying to guess what my reward or punishment would be if I let a mass murderer (who, given the context, I sympathised with) off the hook, this is one of the very few times when I’ve cared enough about the characters for emotion to well up and affect my judgement in any serious way. I was impressed, and sorta wondered what kinda suggestive gasses Troika managed to make my keyboard emit.
Perhaps it was a simple case of wanting to gain the affections of the lady with the biggest boobies in the game, or feeling sorry for a man whose wife and family were killed right infront of him (holy crap, am I ever susceptible to the cliche’s?!). Perhaps it was the facial system, which were used a hell of a lot (even if there were some perculiar emotions being expressed). There was enough there, though, to imply that these people were real characters with motivations, agendas, feelings, and a place in the world. It was faked, but it did enough to get the point across.
So, your prediction sounds good to me!
The explanation for this can be found in any introduction to evolutionary psychology, such as. The seemingly irrational behaviour acts as a sort of doomsday device – threats will not be taken seriously unless you are willing to back them up, and having “if you cheat me, I’ll hurt you back, no matter the cost” hardwired into your genes is an effective deterrent.
“Irrational” from what or whose perspective? From a population or system’s perspective perhaps a little bit of individual-irrationality may be necessary to enhance benefit for all – consider “Tit-for-Tat” (Prisoner’s Dilemma) as an example.
Well, it is irrational behavior in the experimental setup as the participants presumably do not know each other personally.
I completely agree with Nick and Nathan that it may be a rational (optimal) behavior in real life – for the individual and/or for the group.
Nick, I missed that you had blocked about such things already. Late to the party.
The point about experienced vs. inexperienced players was based on the assumption that experienced players are quicker to see the game as rules and ignore the fiction, and hence experienced players presumably behave more (classical) rational.
The interesting prediction – in terms of video games – is:
“The role of disgust in decision-making when interacting with NPCs depends on how much the player thinks of the game as a fictional world. If the player believes in the fiction of the game, disgust will be a factor, if the player thinks of the game as a set of rules for which to optimize his/her personal performance, disgust will not be a factor. ”
To test it, you’d have to measure two variables, right?
You want to measure “fiction mindset yes/no” and “non-disgust/disgust” and then test the correlation. But how to go about testing “fiction mindset yes/no”?
Jesper, you are right about the rationality thing.
Nathan, Tit-for-Tat is hardly individual-irrational even if it happens to have collectively beneficial effects (under the concrete rules where TfT did well).
– Jonas
Jesper, I understand why you want to vary player experience — in order to test what Jonas calls “fiction mindset.” I think it’s good to test this, I was just suggesting a different experiment design that might be easier.
Without relying on a particular pool of subjects of a certain sort, you can build a system that has one richly fictional world with characters who are well-developed, and one that where the world is much less rich. This way the virtual world itself will provide significantly more or fewer affordances for the fictional mindset.
Recruiting one group of experienced players and one group of novices might be harder — what if everyone who turns up for your experiment is in one group? There also might be other differences across those groups that don’t have to do with willingness to embrace the fictional world; the novices may not understand the basic interface as quickly, for instance. So you can’t control the differences between those groups as easily as you can control intrinsic aspects of the system you develop for your experiment.
Nick, agreed, there are two parameters – the game and the users. You’re right that it is easier to use the game as parameter. Now, where are all the research assistants that should be doing my research for me?
I was wondering if, somewhere deep down, we find traces of Descartian dualism in your predictions Jesper… (The Horror).
Not so sure about disgust in particular (interesting things about that in Paul Bloom’s ‘Descartes’ Baby’, though) but would the above be implying that strategic thinking is unemotional?
Acting rationally/sensibly often involves emotions, thus we should not expect game play to be emotionless. But of course many emotional responses _may_ depend on the player’s perception of the opposition (human, machine, nature etc.). So you could be right after all.
Hmm…
Regarding the strategic role of emotions see Damasio’s ‘Descartes’ Error’ and Robert Frank’s ‘Passions within Reason”.
– J
Jonas: The website for the book on “Behavioral Game Theory” that Nick refers to states that “Game theory, the formalized study of strategy, began in the 1940s by asking how emotionless geniuses should play games,”
This is true enough, right. Game theory is the study of optimal strategies, the research question I posed is to what extent and under what conditions players play game-theory-optimally.
I read Damasio as saying, not that can never implement an optimal strategy, but that we have emotional motivations for selecting the goals that we work towards …
Jesper: Ok, I didn’t actually realize that was the original question.
By players playing rationally, do you then mean
A) Players happen to act in accordance with what a game theoretical analysis would prescribe
or
B) Players analyse the game rationally and THEN employ a “rational” strategy
?
I guess I read Damasio more as saying that in real-life human existence (and we may well consider some games quite different from that) it doesn’t really make sense to think of “rationality” and emotions as separate things.
Haven’t read the books cited, but I don’t think the question of “when do people play optimally” can be reduced to whether they accept the world as fiction, or whether there is an NPC involved. Some people will always attempt to play optimally (in either circumstance), whereas others (many others, I think) see the play experience as more important than the play outcome. So you might start a chess game with a new gambit, for instance, not because you think this will result in winning, but because you think it will result in more interesting game play.
I should add, though, that yes — I think you’re right in both hunches (no solid evidence) that 1) people have certain ethical expectations of human actors that will result in a refusal to deal when they believe those expectations have been breached (and this is *not* irrational, btw), 2) that people who are immersed in a fiction are more likely to role-play. If we peg role-play as irrational (again, skeptical of whether that’s a good semantic move), those people are irrational. So, yes, I think you’re right. :-)
Single iterations of experiments are not valid in a game perspective. Only a multiple iteration situation is applicable in a MMORPG (in particular), or a general gaming frame of reference. Thus not accepting an unfair split is signaling that such behaviour is unacceptable and that in the case of future co-operative situations the ‘giver’ may be ot of luck. A game setting is also inherently more complex than what basic game theory evolutions, as given in the example, can deal with.
As for there being a difference between experienced/inexperienced players, it will not be in their reliance on disgust as a mechanism, but in understanding when an offer is ‘disgusting’ to use your terminology.
The level of accepting the fictional and disgust in dealing with NPCs is a function of personality, not rationality. The optimal solution in a game is that which provides the most fun, thus what is rational to one, is entirely irrational to another. Experience does not, as far as I can tell factor in. Game theory can only account for this if ‘fun’ is given a value.
I’m pretty sure I’m coming at this from a different angle than some of the rest of you, so I’m not entirely sure how applicable my comments are in the overall scheme of things.