Over the two weeks I listened to Super Crunchers: Why Thinking-by-Numbers Is the New Way to Be Smart by Ian Ayres during my commute back and forth from Dallas. The book is very interesting. Most of what I got out of the book directly was inspiration and ideas on ways to use number crunching in everyday life. Much of what the book is ultimately about is prediction – trying to look into the future to gain some advantage. I think prediction is a really interesting subject, whatever side you’re on (whether you believe in predicability or randomness).

One of the biggest points Super Crunchers trying to make is that people are computing scenarios all the time, trying to predict what will happen – this applies to betting, the movie box office, retail sales and all sorts of other things. The author basically says you can be on one of two ends. You can either be a luddite who is afraid to engage with numbers and computation, in which case you will be put at disadvantage. Or you can be enlightened and use number crunching for your benefit and take advantage of it’s uses. You will be on one end or the other – you might as well make a conscious choice. Those who embrace number crunching stand to benefit from it.

A few examples from the book:

• A wine collector uses past data to predict which wines will be good based on weather and other current data.
• Someone creates a website to predict the prices of seats on airplane flights.
• A company creates an algorithm to help predict revenues and suggest changes to improve profitability of movies.

So for me the book Super Crunchers was really interesting. I liked the audio version, but I think it would still be a worthwhile endeavor to actually read the book. I would give it a rating of 4 out of 5.

I must confess, I have become a bit of a skeptic. I have been reading the writings of Nassim Taleb and his work makes sense to me. Let me explain.

I invested in some stock and recently sold it for a nice profit. I was invested, I am now divested, and I need to invest once again. However, not just any stock will do. I think the stock market is misvalued according to the state of our economy and I don’t trust these prices. I find it difficult to find good value. This is not a new thought for me – I do not like the idea of speculation and have not since Benjamin Graham opened my eyes. I prefer intelligent investing.

Taleb talks about risk of the unknown unknown. What we do not know, but think we know, can hurt us far more than what we know we do not know. That’s confusing I know. Jeff Hawkins would say that the neocortex works by making predictions about the future – that is it’s basic function. We all do it, there is nothing wrong with that fact. The problem comes when we make predictions about situations we know nothing about. And usually we think we know more than we actually do. Those predictions set us up for failure. We will be rocked when something completely unexpected comes along – something we did not take into account. Basically, most of us are very prideful in our predictions. This is why the housing crash, 9-11, and other events hurt us so badly. Those are the times when our pride is revealed as misguided. And that is a chance for humility to set it.

But that doesn’t happen much. Most of us never set aside our pride. Most of us continue in our misguided, prideful predictions made on limited amounts of knowledge. Taleb ultimately advocates a defensive stance, an intellectual humility, which I find intelligent.

Taleb talks about information in terms of gausssian (or normal) distribution vs fractal distribution. I do not pretend to understand all of it, but I am beginning to see it play out. For instance: Last night Ezra, our German Shepherd Dog, wanted to go outside in the middle of the night to go to the bathroom. This is not uncommon. So Angie took him out then brought him back into our room. Not long after that he started whining and scratching at the door. We tried to tell him to go back to sleep.

Now, if we assume normal distribution of time between bathroom breaks compared to dog age over the lifetime of the normal dog, you would see something like a bell curve. An unpredictable, shorter time between bathroom breaks when they are a small puppy, a gradual curve up to predictable longer times between breaks around mid life, then a gentle slope back down as they age. The same could be modeled in humans.

However, this is not really true of real life. The model is flawed. The model is a broad sweeping generalization that does not take into account outliers and “random” events. There is a small chance that something is actually wrong with my dog and that he really does need to go back out. Maybe he is sick. In this case the Gaussian curve is a poor predictor of behavior. Even though the curve would say Ezra needs to go out no more often than every 5-6 hours, this might be a random event.

Usually the random events are identified post-disaster and make sense. Taleb calls these Black Swans. I do not know that Ezra’s needing to go back outside after 5 minutes is a true Black Swan, but it does illustrate the need for a fractal distribution. In order to make the correct choice (taking him back out because he is sick), I must assume that there is information I am missing. I must come to the situation with intellectual humility – otherwise I will have a mess on my bedroom floor and it will not be the dog’s fault, it will be mine.

I am not sure that makes sense. I hope you understand what I am trying to say, which is basically that we all need to make decisions and predictions from a place of intellectual humility. We need to act defensively and assume we do not know as much as we may think. If we do this, we will be much less surprised when something beyond our radar happens, and perhaps even be in a place to take advantage of such a situation. If you had bet against the stock market in 2008 you would have made a killing as a defensive, skeptical, intellectually humble, intelligent investor who knows the future is not nearly as predictable as it seems.

AI Techniques for Game Programming is the book that got me interested in Genetic algorithms. I was originally interested in Neural Networks and went to the public library looking for anything involving that subject. In this book it also talks about Genetic algorithms as they are applied to games. I started experimenting with Genetic algorithms after that, but have never really implemented them into a real life situation.

My project does not involve Genetic algorithms as they appear int his book, but the idea of genetics and how it is applied is what is relevant to me. The nice thing about this resource is that it’s written in a way that’s easy to understand. The bad thing would be that the code is in C and the design/readability is awful. But this was a very helpful resources (and continues to be) as I’ve tried to understand genetic algorithms.

On the side (of web design and development, teaching, etc) I like to read and think about investing.  The Intelligent Investor is my favorite book on investing, written by the brilliant man Benjamin Graham.  The reason it is my favorite book is that it makes practical sense to me.  It seems wise.  It rails against speculation.  It is about as conservative and grounded an investing strategy as you can get.

As I read through Taleb’s The Black Swan, I recognized an idea that is complementary.  It is all about limiting your risk to the unknown – which is really what Graham was getting at.  Here’s Taleb’s approach:

I am trying here to generalize to real life the notion of the “barbell” strategy I used as a trader, which is as follows.  If you know that you are vulnerable to prediction errors, and if you accept that most “risk measures” are flawed, because of the Black Swan, then your strategy is to be as hyper conservative and hyper aggressive as you can instead of being mildly aggressive or conservative.  Instead of putting your money in “medium risk” investments (how do you know it is medium risk?  By listening to tenure-seeking “experts”?), you need to put a portion, say 85-90 percent, in extremely safe instruments, like Treasury bills – as safe a class of instruments as you can manage to find on this planet.  The remaining 10 to 15 percent you put in extremely speculative bets, as leveraged as possible (like options), preferably venture capital-syle portfolios.  That way you do not depend on errors of risk management; no Black Swan can hurt you at all beyond your “floor,” the nest egg that you have in maximally investments.  Or, equivalently, you can have a speculative portfolio and insure it (if possible) against losses of more than, say, 15 percent.  You are “clipping” your incomputable risk, the one that is harmful to you.  Instead of having medium risk, you have high risk on one side and no risk on the other.  The average will be medium risk, but constitutes a positive exposure to the Black Swan.  More technically, this can be called a “convex” combination.

The Black Swan, pages 205-206

In addition to it’s application to investing, this idea can be applied to life.  And, indeed, Taleb details what that might look like at a later point in the book.  But the idea is to be conservative in your job, in your pursuits, etc with all but 10-15 percent of your life.  In that other 10-15 percent you take risks that have maximum upside.  As in, look for opportunities that “scale” (as a business person might say).  You look for things that allow you to input some amount (time, money, etc) but can return gains that are non-proportional.  I can put this idea in the negative better than the positive: if you plan on making widgets, your earnings will always be proportional with the number of widgets you make.  Sure, you may be able to raise the price, but you will always have to work harder to make more widgets and make more money.  If, instead, I could write a book, there’s a chance that book will sell millions of copies (*cough* Harry Potter *cough*) and give me a disproportionate return.  As in, I write one book which takes me x hours (say 200 hours).  The more books I sell, the more I make for those 200 hours – I do not (necessarily) need to work more hours to make more money.  David Heinemeier Hansson from 37 Signals talks about this same idea.

Anyway, it’s a very interesting idea to me and I wonder how this might apply to both my career and my investments.

I am currently enrolled in ITGM 723, Human-Centered Interactive Design.  Generally, I don’t love “clinical” user testing. I don’t like bringing someone into a room, giving them a task, observing them doing the task, recording the results and creating the analysis. I would rather talk to a few people, talk through the process, what they might like in a product, what’s important to them, etc. However, I know from research a person’s spoken desires and functional desires are two different things. In other words, people have a hard time articulating what they want, but they can usually identify it (in the negative or positive sense) when they see it.

Paper prototyping is a traditional, clinical way of reviewing user needs and refining user experiences. I’ve not used paper prototyping (as a testing mechanism) before. As I read about it, and as I put together my first paper prototype I felt the whole idea was very stupid. To put that much effort into a “functional” prototype that did not move the development needle seemed ludicrous. Even still, I picked a task and created a functional paper prototype.

I went ahead and read other chapters of the book Paper Prototyping by Carolyn Snyder besides the one assigned. Specifically I went through the section on “Deciding Whether to Use Paper.” I begrudgingly agreed and understood what the author was getting at and why she might be a proponent of paper prototyping. I read through the research that shows that low and high fidelity prototypes generally reveal the same problems. And I feel like I understand a little better now.

However, it still bothers me that paper prototyping doesn’t really contribute much to the actual development process. Digital interactive wire frames are more involved to put together, but I think there’s some merit to the fact that I can actually expand on that development and it’s not “wasted time.” Especially in the recent years with the advent of XHTML and CSS – moving information around on the screen is not difficult and layouts can be changed very easily – that’s the point of those technologies. And I understand the author of the book is advocating “rough” paper prototypes and rapid development of those materials, but it still seems like a waste to me.

So all in all, I see how paper prototypes can fit in. I understand the merits, and I see some of the drawbacks. Will I use paper prototyping in my professional career? Probably not much if I do at all. Was it a good exercise to go through? Absolutely. Is it a good tool to have access to? Sure, it’s always good to have another tool available.