Friday, December 13, 2013


"If it looks like a duck, swims like a duck, and quacks like a duck, it's probably a duck." ~ the duck test.

Actually the wikipage on the duck test is pretty amusing, largely because a potent picture-caption combo.

The duck test is a heuristic, a rule of thumb. It may lack the precision of say, DNA testing, but it's good enough.

What is good enough? Something that works often enough to be useful, and when it goes wrong it's of negligible consequence. A heuristic survives because of it's imprecision, because it has a bunch of exceptions, it gets naturally selected out if these exceptions are actually of any concern.

Go ahead, try and imagine a situation, a dire situation where not knowing if something that looks like a duck, swims like a duck and quacks like a duck, is of dire consequence?

Notice, that the duck test is not for example going to get Elmer Fudd confused betwixt firing at a duck or firing at his own daughter, or even confusing a duck for a rabbit for that matter.

As Douglas Adams parodied:

"If it looks like a duck, and quacks like a duck, we have at least to consider the possibility that we have a small aquatic bird of the family Anatidae on our hands."

Now I'm sitting with my friend Rod and he is talking about the launch of the Apollo mission. If the trajectory is off by 1 degree, just 1 degree, the rocket misses the moon by some hundred-kajillion-billion-miles. 

What's the solution?

I facetiously suggested you launch rockets in 180 directions, so one of them is bound to hit.

The answer though, is (hopefully) obvious. You build a rocket that can adjust it's course. 

I'm sure the apollo missions had a more specific trajectory than 'up' but essentially, when you are able to adjust that's really all you need.

NNT would cite domain dependent thinking, in that while it seems obvious to build rockets that can adjust course and cars with steering wheels and ships with rudders and pencils with erasers. There's times when we strive to an ideal of perfection, precision, as if the best approach to the problem is simply to get that trajectory absolutely right.

What are Carl Thompson's secret's to making the worlds best bass guitars? 'Use Good Wood.' not particularly helpful, but I think if you come across a number of master craftsman, practitioners, artists, chess masters you'll find they just have broad principles to convey. No specific technique like Kung Fu Panda's Wu Shu finger hold.

While presumably there's a lot of skill in balancing a bass so it sits right on the body, and placing frets so they achieve the right harmonic frequency, Carl Thompson pointed out the hazard of overdoing such precision work - he'd seen bass builders pet a fret board on that was thin as paint, so that it got scratched and was impossible to remove because the veneer was so masterful. Or wiring up the pots so the wire was the perfect length, then one of the pots busted and you had nothing to play with, nothing to cut away so you had to rewire the bass rather than just replace a knob or pickup.

NNT divides these principles into a dichotomy of being 'approximately right' or 'wrong with infinite precision'. Delicious.

Let's recap on what has been covered in this blog before.

So Miyamoto Musashi arrives late at Ganryu Island by boat, where he has carved a 'bokken' wooden practice sword from an old oar in the bottom of the boat. Sasaki 'Ganryu' Kojiro whose day goes badly enough for an island to be named after him is furious at Musashi's tardiness for their duel. According to legend, Kojiro drew his sword 'The Drying Pole' and threw his scabbard away.

Here remarked Miyamoto Musashi 'You've lost.'

What makes the legend plausible, is if you read the Book of Five Rings, I can't account for what is lost in the original language, but from the translations I've read, Musashi's confidence is about as complete as one could get. I can totally see the author calling a fight for himself, to his opponent as it is being engaged.

Assume the legend to be true then, Musashi knows none of the details, couldn't possibly know, how the fight would go down. He just applied a heuristic, a man who throws away a perfectly good scabbard doesn't expect to win. A victor would make sure to keep their scabbard, for being alive, they would need it to sheath their sword in.

The other (perhaps) heuristic or reverse heuristic to apply is that a man that turns up with a wooden oar handle to face a Japanese katana doesn't expect to lose (or even be hit).

Move past Musashi's voice in his book of 5 rings and you'll find it filled with vague hueristics, including his 'body of a rock' untouchable defence. I think it's a not-unreasonable inference that rather than some mystical technique that hardens Musashi to stone, the principle was to not get hit at all by an opponents weapon - based on the fact that Musashi in legend would fight naked blades with wooden practice swords, sticks, fence posts, scabbards etc.

Switch over to his contemporaries in the Yagyu clan, who devised the 'no-sword' school of kenjutsu, and their invincible defence breaks down to getting a good judgment of your opponents range and never letting them get in range.

I share all this to point out that those adolescent minds that seek out esoteric martial arts treatises for secrets to hidden human potential are often dissappointed that far from specific actionable techniques to dominance, the martial arts treatises are usually just a collection of heuristics.

But that in itself is telling. Because martial memes are subject to the same darwinian forces that apply to biology. Ineffective martial memes kill their host organism when stressed.

Sun Tzu's art of war contains no concise, comprehensive strategy of dominance. Just a set of rules that evidently over the ages and abstracted to other contexts work 'well enough' most of the time. The Art of War breaks down to a collection of duck-tests.

Get to the battlefield first and you'll win.

Take the higher ground and you'll win.

Always punish deserters and cowards.

Don't eat provisions provided/discarded by your enemy.

And so on for a bunch of pages, of which I'm sure history would find exceptions that would reject every single thing discussed in the Art of War as a hard and fast rule. But I guess that's why it's called the 'Art' and not 'Science' of war. These are just generally good strategies and tactics to adopt. They work well enough that they are a good basis to adjust from. And if you are aware that they are fallible, then they actually help you see that the problem is not as straightforward as it first appears.

Because what makes a rule of thumb smart, is that they have a built in stupidity, they are fallible. They inspire a different kind of confidence to say Modern Portfolio Theory - it's genuine confidence - the willingness to take risks rather than hubris - a belief in riskless-profit.

As President Ike quoted 'Plans are nothing, but planning is everything' as good a military heuristic as there is. Nothing goes to plan, planning allows you to recognise what isn't going to plan.

I also like the heuristic 'you know you've won when the enemy conforms to your strategy' you don't need to know the details, you can't know the details, you just need them to start behaving like you want them to.

I haven't read far enough into NNT's book Antifragile, he certainly included the memorable phrases of 'approximately right' vs. 'wrong with infinite precision' in the futile attempt to eliminate all risk from life. But I'm sure he's coming to the fragilista's and could predict that they will never use heuristics, or if they do, won't treat them as soft general principles but hard and fast rules - subject to 'insults from reality' as Sam Harris would say, an author I suspect NNT doesn't respect much.

Anyway, due to my current reading, I learned a new word, a word central to who I live my life and interact with my environment.

Precision is at best a training exercise. Heuristics are for going live.

No comments: