A non-deterministic system is not quite the same as a non-predictable one. This is where probability comes into play and makes me think we're really on the right track with currencies as tools for expressing/modeling complex, living systems.
[BTW, I still can't post replies, so I'm making this a main blog post even though it started out as a reply to Alan's post "Do currencies describe living systems?"]
Rain falling on a hillside is non-deterministic. We could never be sure about the path any drop will follow. However, taken as a whole with hundreds of thousands of drops, we can probably predict rather easily where water will run, where it will puddle, where much will soak in, etc.
That is what makes the design of a (monetary) currency SO IMPORTANT in the way it shapes behavior. It is non-deterministic -- it doesn't take away the free will of the actors. You might choose to burn that $20 bill in your wallet, but chances are good that you'll spend it... in a fairly bounded range of ways.
The design of a currency is like the shape of the landscape upon which the rain falls. Each drop follows its own non-quite-predictable path as each person performs their own not-quite-predictable transactions. But one of the things that I think fluency in flow language will make clear to us, is that taken as a whole, the probable pattern is VERY CLEAR.
Like with dollars being issued as debt which bear interest... That interest will compound, daily, forcing exponential growth until the debt load is growing faster than the economy can expand which causes collapse of the currency. The outcome is really as plain as day to anyone who looks at flow pattern.
However, it is not quite as clear exactly what day that collapse occurs. Or how hard the fall will be, because those factors are outside the inherent rules of the currency and are more like your chess game factors of how much sleep the players got or if they're playing for a big prize.
Even some of the pattern outside the explicit currency rules is visible when in play. The Bush administration as it greedily centralized wealth into the hands of the few, accelerated the collapse severely. Offering a "bail-out" which did more of the same (by handing it directly to the bankers) didn't slow the unraveling. Obama's bailout approach, which spreads short-term wealth around, will slow the collapse, perhaps for months or years.
However, they both are rooted in the same unsustainable core principle of dollars as a Ponzi scheme requiring us to perpetually mortgage our future with present debt. We cannot bear the tax burden to repay the "stimulus." It extends the downward spiral of compounding debt toward sure collapse.
Non-deterministic system --> Certain outcome.
Of course, this way lies madness for predicting on an individual level. It reminds me of Skinner's behaviorism and likely why his kids ended up institutionalized. Just because you can consistently get a pidgin to turn left to get you to release food does not demonstrate much flow language wisdom. The model is too one-dimensional to operate outside of a very artificially constrained environment.
Just like in most economic plans or corporate compensation plans where people think they're causing predictable behavior through monetary reward...but the plan is too one-dimensional to connect to the various layers of incentives of a whole human. This becomes painfully obvious when worker productivity drops instead of rises, or consumer confidence still falls even after $350B of bailouts. :)