« Pat Coa | Main | Movable Type migration update »

Monday, 19 January 2004


Concrete progress
on the Game of Life. Glad to see it.

Very interesting thoughts... It's been a theme in my head and my paper journal lately, though I haven't been able to really put very many coherent things together to say about it. Email, weblog comments, politics, altruism and goodwill. Is there any system that is immune to gaming, or is there any way to remove the incentive to metagaming?

On the other hand, reading Gödel, Escher, Bach again makes me consider that one of the basic principles of human thought is gaming any and all systems we encounter.

I need to think more about things to say...

what is a system? who are stakeholders?
Another framing...

* there's a group of people interacting

* there are unstated rules, "generally" agreed upon by members of the group, about appropriate behavior

* there are some stated formal rules which encompass a subset of that behavior

* "gaming the system" is behaving in a self-interested way within the boundaries of the explicit rules while ignoring the tacit rules and implicit values.

It's only effective if the other "players" don't notice, or you don't place much value on the future of the interactions with them.

I think gaming the system can actually be used as a bonding tool if you're gaming it with other people who are in the know.

Thinking about it in the context of object oriented programming, the explicit rules of the system are like the interface, and then there are the side-effects which you want to try to hide behind those explicit rules. You don't necessarily want people to connect the side-effect of "getting there faster" with the speed limit that they need to obey for their own safety. It's an implementation detail, an undesired side effect of speed since, in the context of speed laws, speed's primary effect is a loss of safety.

implicitly zero-sum?
I wonder whether any behavior we'd call "gaming" must be within a zero-sum game, because what's pissing us off is that the other person's gaming results in his gains, at our implicit loss?

Some counter-strategies:

* make actions Transparent (feedback, shared Reputation)

* make opportunities/benefits of longer-term relationship visible

* ???

An interesting observation in that book, Nonzero, is that most nonzero sum games become zero some at some point. For example, selling a product is a nonzero sum game for both the buyer and the seller since they're both getting more out of the transaction than they're giving. However, the exact cost of the transaction can become zero sum because if you're bartering over whether you pay $10 or $11 for the product, that $1 difference will be lost or gained in a zero sum way depending on the outcome.

So, I guess what I'm saying is that if gaming only makes sense in zero sum systems, I think we'll still find that it happens in almost all systems since almost all systems will have a zero sum component to them.

I'd like to follow up on counter strategies. It seems like one strategy is to make all actions explicitly allowed. Make drugs legal, have no speed limit, etc. Another is to make things less transparent--hide the side effects and only allow effects that support the system's explicit rules. That's probably not possible in many systems though, since you can't really prevent things from being effective--unless you do so by adding punishments to laws, and then the gaming component just becomes a matter of not being caught.

I think the benefit of long-term relationships and transparent histories is going to have a tough time beating the short term gains and pact-making of similar system gamers who will help keep their activities private.

zero or non-zero
I confess I haven't read Wright's book. It fell into the category of "interesting but sounds like so many other books and articles that I've read that I can't get myself to spend more time/money in that direction".

That price-negotiating example sounds weak:

* if it was costing $8 to make the good, and it had value to the buyer of $15, then both $10 and $11 count as "wins" for both players. Of course, if the cost and value are both $10.50 then things get different... but either way each party has the option of walking away (barring monopoly/monopsony)

* in individual negotiations, in particular, I find there's huge amounts of uncertainty and therefore imprecision in evaluations of cost and value. So "digging in" to a particular price is often a sign of a mentality that will *eventually* lead to problems either way (esp in an environment of ongoing relationships/transactions).

(This is a pretty fuzzy area - more narrow examples make for better discussions)

Also, re "make all actions explicitly allowed" does not seem realistic for human relationships. Can I bitch-slap you? (Again, trying to cover every type of interaction under the same labels seems counterproductive..)

I wrote something along these lines
Back when I was in high school, I wrote an essay about this topic and how it relates to the public education system.

The comments to this entry are closed.