In my previous life I had to give talks on a regular basis usually running from around 15 minutes to 60 minutes (questions included), although a few were 10 minutes and 120 minutes respectively. I found the size of the audience (usually 10-300) never really matters as much as its composition, here experts are easier to “handle” than non-experts, because the latter may misunderstand the subject and ask strange questions(*).

(*) My favorite one being the nuclear physicist who was trying to use the optical theorem (quantum wave scattering) to try to understand a binary star collision.

There is one golden rule when it comes to giving talks and that is “Don’t get derailed!”. Other than that, there are two ways to give talks.

First there’s the linear process. In the linear process, you cover 1/3 of introduction, 1/3 of intermediate stuff, and 1/3 of advanced stuff. Normal attention spans, even for people who have trained for years to sit quietly and listen, is around 17 minutes. Hence, try to make your introduction as good as possible … 17 minutes is just about your introduction in a standard seminar. Of course this could also suggest that people lose their focus once they no longer understand what you’re talking about. Personally, I have, therefore, broken this role and done 2/3 of introduction and 1/3 of advanced stuff. I feel that if people aren’t learning something, they’re wasting their time. The problem with the linear approach is that it is very susceptible to derailing. Every slide needs to be superbly timed. Also, if people get lost, they are really lost.

The other approach is the onion approach or the spiral approach.
The idea here is to start be describing the complete approach as simple as possible leaving out a lot of details. Then process is described again in somewhat more detail. And then a third time.
This approach can not be derailed (mainly because there is no rail). On the other hand it can not make rapid progress. In fact progress is hard to quantify.

In my martial arts training, I have noticed that karate uses a linear approach to teaching. You drill a set of techniques for a certain time and then you get tested in those before you start learning new techniques. Each new technique is quite different from previous techniques. Conversely, shinkendo uses the onion approach to teaching. Each technique has several levels starting with a couple of cuts and then progressively adding more cuts on the more advanced levels. While I have only tested for the first level, I know (a few) techniques from the fifth level.

The linear approach is the specialist’s approach whereas the onion approach is the generalist’s approach. A specialist needs to be competent in a single area, and thus always moving in one direction is the most effective approach. A generalist must be widely competent and thus can not afford not to know the wider environment, he is moving through.

The onion approach to teaching is much harder to implement because it has few precedents in our educational system, which of course is mainly designed to sort students into various vocations and teach them specialized skills. Hence, few people have much experience in teaching in this way. As well, few people have experience learning this way and will tend to find the onion approach unstructured and wavering.

For this blog, I have tried the linear approach already in the 30 day challenge and the problem here is clearly that people will get derailed by their inability to, say, move out, get rid of their car, or similar, basically making the rest of the “talk” a waste of time. I wonder, therefore, if the onion approach is not the superior way to teach something like early retirement. What do you think?

Originally posted 2009-06-19 12:11:46.

Game theory is the mathematical description of the structure of strategic decisions when the choices of two agents directly influence the other. Much work has been dedicated towards finding optimal strategies; in fact much military and political strategy is built on mathematical foundations, but it also governs economic decisions.

Consequently some knowledge of game theory can be very helpful.

The simplest way to describe a game is by using a payoff matrix. Suppose player 1 has strategies A and B and player two has strategies C and D. Then there are 4 outcomes: AC, AD, BC, and BD each with a certain pay out.

The most famous game theoretic problem is the prisoners’ dilemma, where two criminals are interrogated by the police. In the prisoners’ dilemma, a prisoner has the choice of ratting on his partner or keeping quiet. Here is the payoff matrix.  Hence quiet-quiet = 3 years for both, quiet-rat = 10 years and 0 years. rat-quiet = 0 years and 10 years and rat-rat = both 5 years.

What is the optimal strategy? It turns out that the optimal strategy for the individual is to rat on your partner. Ratting on your partner is a so-called Nash equilibrium after John Nash who was portrayed in the movie A Beautiful Mind. A Nash equilibrium is the optimal strategy for an individual given full knowledge of the other player’s strategy. In other words it is the choice of strategy where the payout of the individual can not be improved by changing strategy. In the case of the prisoner, since 0 years is better than 3 years and 5 years is better than 10 years, a prisoner can not improve his position by keeping quiet

This is an interesting case where the global optimum will not be chosen. This is what makes it a dilemma. People with too much sparetime can now make a game theoretic model of whether for two hostile nations it is in the best interest to for a non-nuclear nation to pursue a nuclear program given the idea of a Nash equilibrium (answer: it is).

Now, bripblap asked if the same holds in terms of wealth.

Let’s look at that. I’d like to rephrase the question in terms of whether trade has a global optimum (it does). When trading people are often confused by vague definitions of price and value, so let’s define those first. Price is the objective measure of at which a trade took place. If I traded 10 hogs for 10 cows, the price of 10 hogs is 10 cows and vice a versa. Value is wholly subjective. I am trading my 10 hogs for 10 cows because  I believe 10 cows are of greater value than 10 hogs. The person I am trading with believes the opposite.

Drawing a payoff diagram in terms of value shows that the optimal strategy in a free market is to only accept a trade when I believe that I am receiving more value (wealth) than I am delivering. The trade does not take place unless both counter-parties think the same. Hence, value is always created. That is the beauty of a free market system.

Conversely, if the trade was coerced (taxation), at least one party would think he is getting less value than he is receiving. This can lead to zero-sum or even negative-sum games. So why do we do this? The reason is that payoff is not strictly linear. Most humans, for empathetic reasons believe that a situation where a majority of people are slightly discomforted is optimal if it means that one person avoids severe pain (disutility). In other words, when it comes to extreme hardship, the needs of the one outweighs the needs of the many.

Another thing that can lead to negative payoff is buyer’s remorse. Although a rational person would base decisions on all available information at the time of the trade, few people are rational. Hence people will often choose suboptimal strategies. This, as well, can lead to zero or negative sum games. This does, however, not necessarily mean that the other person gets more value, but it is often so.

Originally posted 2008-11-07 07:36:27.

I just watched this genius(*) movie for the first time after having borrowed it from the library. You may be able to watch it for free here.

(*) I like to believe I don’t use that word lightly.

The movie has no dialog and no plot but a powerful message which is achieved by using different time and dimensional scales (something I got used to in my professional career which might be why this works so well on me) to put humanity in context to the planetary timescale, mostly the atmospheric timescale which runs about a thousand times slower than human life does. From this perspective human life begins to resemble that of a of an insect colony, human motion becomes the blood vessels of a super organism which is technology society that spreads like a cancer (or a mold) on the surface of the planet. Humans are organized very similar to how components are organized on a microchip. This leads us/me to consider whether these are natural patterns, specifically whether anyone is in charge of whether technology just works the same on different scales (it probable does). Then if this is indeed true who do we really serve when we go on about productivity and quotas; is me pumping out blog posts really that much different from a NAND-gate doing its thing in a computer? Who serves who?

Note that Koyaanisqatsi has two sequels.

Originally posted 2010-03-22 21:38:39.