The Toilet Door

So there's an election due in September and this means that both major political parties are starting to have some 'serious' debate about the state of the economy, debt, policies and so forth. One thing that I've noticed is the way all parties use numbers as talking points as metrics for success or failure.

A hypothetical example would be, "We've created 150,000 jobs and unemployment has dropped to 5.2%". At face value this seems like a good statement to make, but the more I think about it, the more it seems meaningless. Creating jobs is obviously a good thing but it's the numbers they throw around in a debate that concerns me. What does creating 150,000 jobs mean? Did the public service grow by 150,000? Are the jobs in industries or fields we should be investing in or pulling out of? Are they full time jobs or is this casual work? Similarly with the unemployment rate dropping to 5.2%, what was it before? What is the underemployment rate?

These are all questions I'm thinking of when I hear someone mention something with numbers, what do those numbers really mean in the larger context of everything else. Now obviously numbers are important, having any metrics, even flawed ones are better than having no metrics, part of the issue I realise now is that people don't really understand numbers.

A study was performed testing if people would behave differently for a reward if it was $3 or 300 cents. The results, surprisingly indicated that some people preferred 300 cents. Even if  cents and dollars were switched around, people were more easily swayed by the larger number.

This knowledge has some interesting consequences for looking at political discourse. Are the politicians aware of this effect and use it to mislead or confuse citizens about the state of things? It's possible, there's limited time in media segments to accurately and adequately describe what a number truly represents, it's probably more important that the reader or viewer simply remembers that it was 150,000 jobs created or that a policy will cost $94 billion.

Speaking of policy costs, it's interesting to observe that the cost of everything is often put in vacuum. $94 billion sounds like a lot and it rightly is for an individual to own, but in the context of an entire country that has a yearly GDP in the order of $1.5 trillion ($1500 billion), it doesn't seem as large, it'll seem even smaller if instead of stating the total cost over 10 years and comparing to a yearly GDP, we state the yearly cost $9.4 billion.

I'm going to keep an eye on how it progresses and see if there's a correlation between the way the numbers are presented and how they are meant to be viewed. Obviously positive achievements would be promoted and negative achievements downplayed.

So one thing I enjoy is games and exploring the core concepts and skills they test when taken to high levels. One thing that nearly all games have in common is the concept of trading.

Now trading in a traditional sense could be two parties coming to a mutually agreeable set of terms to net benefit of both parties. In most competitive games however, trading isn't mutually agreeable. In Chess say, both players start with the set of pieces. Each turn a player is trading positional advantage for material advantage, or vice versa. Sometimes a player may initiate a trade by taking a piece. The opposing player may even the trade up by taking a piece back of equal value. The most common example of this is in Chess, where it's rare to not trade Queens. You don't want to give up your queen without getting a significant material or positional advantage out of it.

So onto Monopoly, it has a trading element core to the rules of the game. Players can make trades for property, cash at any time. The main reason to do this, is to gain a cash advantage or material advantage over your opponent. Now, a lot of people when they play monopoly use a flow of logic when it comes to trades along this line. "I don't want to trade with you, because you're making this trade to get an advantage over me, no matter how much you are seemingly offering me". So if there's no trading, then monopoly is really a game of chance, you're hoping that your dice rolls are favourable on average, and the opponents are unfavourable on average.

Where it gets interesting, is that monopoly is what I'd call a 'solved' game. A number of people have calculated the probabilities of landing on any given square, the average number of rolls for a certain investment in a set of properties to pay off. The limitations on human players would to be to remember and evaluate all the tables of data to know the likelihood of winning from a given board position. So taking it to the next level, let's suppose that we have perfect players that know all the probabilities for any given game state to calculate a winner. Trading properties then becomes simply betting. Investing heavily into mayfair and boardwalk has a low probability of paying off, so most people wouldn't make it, but it essentially becomes an agreement between players that in the next x dice rolls Player A thinks they wont land on it, and Player B thinks they will.

http://www.amnesta.net/other/monopoly/

http://www.tkcs-collins.com/truman/monopoly/monopoly.shtml

I think it's interesting that many games and sports rely on limitations in various skills or abilities to make them fun. Monopoly can be fun because most people are incapable of correctly evaluating the probability of any given player winning from a certain position, but even taken to the extreme, it's betting on dice rolls.

So I've been playing around a bit with LaTeX because I like the idea of what it can do. It sorts out all the tricky formatting depending on how you want to present your work. It's not ideal for collaborative editing though. Microsoft Word wins in that department with comments and tracking changes.

One of the ways I'm going to investigate using LaTeX is with some custom tags. This xkcd comic illustrates which characters interact with each other as a function of time over a story. Now that was produced by reading the books, and manually going through and making note of which characters are where and with who at each time. Ideally though, if that information is put into custom tags at various chapters or at points in the story it can be invisible to the reader, but a simple program or script could extract the information and produce a plot like that. Similarly, if we could generate one for characters, we can also do it for action, or emotional content.

Having the ability to produce graphs to demonstrate the action during a story or comedy say gives an author new tools with which to view the overall story. Is the front action heavy? Is it too dry? Is there enough story progression? Humans are visual creatures, so I'm hoping I can make something to produce these graphs, so that when I begin writing, I have a large number of tools available to guide my story.

So I need to come up with a name for a phenomena I've observed.

Basically I tell my students at the start of the practical the same things. Check the marking scheme, don't forget to talk about errors, work quickly.

Inevitably they ignore me and forgot to talk about errors, so I started telling the students. Check the marking scheme, don't forget to talk about errors, work quickly. I tell all the other students this and they usually fail, so maybe you'll be different.

Inevitably they think I'm joking or something, and they do poorly. So I started telling the students. Check the marking scheme, don't forget to talk about errors, work quickly. I tell all the other students this and they usually fail, and I tell them what I've just told you now and they usually fail, so maybe you'll be different.

So we can really just state it's

Message Content. + self referencing warning

I need to print something like that out for my students and just show them what happens 🙂

So I just read a substantial paper that debunks the argument many people use to support legislation or technology that would infringe on privacy. "If you've got nothing to hide, what's there to worry about?"

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=998565&

Basically he states that privacy is a fairly ill defined term,  and it has a range of classifications. He goes on to define the problems that the term privacy can encompass. The two main ones are Orwellian surveillance and Kafkaesque problems. Kafkaesque being the kind of situation where another party has so much information about you that before you can take any actions, you're already limited in what you can do.

Anyway, it had me thinking about trying to break down privacy to what it really means. All I've come up with so far is that privacy is like a currency denominated by information. The majority of interactions we have in the world are essentially trades in information. With friends, we share secrets, which is an information trade to build up trust. Similarly, when we go for job interviews, or are looking for romantic partners, we have the ability to release and share information in a way that makes us more desirable.

Where privacy is important is that it allows us options in how we reveal information and when. If you remove privacy, it allows other parties you may want to interact with to assess you on either a) incomplete or incorrect information or b) correct information presented in an unfavorable order. For most cases it's probably a), which is why privacy is important.

I guess an appropriate analogy would be suggesting that life is a bit like a card game. You can spend some time trying to trade for better cards, you can play cards in a beneficial order, you can make intelligent guesses and inferences about other people based on what actions they take. The privacy component is that your hand is hidden information, there should be no desire to give up the advantages that privacy affords even if you're doing nothing wrong, because it can in no way help you in the future. What is happening with big data and data mining is that we're getting so much information from everyone about what they are discarding, what they play and what hands they have that we can start building fairly accurate models about what moves you'll make next.

They've already built a computer program that will consistently beat most humans at rock paper scissors based on big data http://www.nytimes.com/interactive/science/rock-paper-scissors.html?_r=0

It's an interesting thing to consider, but for the time being, fight to hold onto your privacy, otherwise you're giving up your social currency too easily and gaining nothing for it. As I'd say in most games, that's a bad trade.

So I haven't updated for a while. I think I was perhaps a bit ambitious to think I could make a meaningful entry every day. I either need to rapidly increase the speed at which I can produce an entry, or I need to make them less frequently. Part of the issue is unbalanced work weeks, but anyway. On with the content.

So one thing I find interesting in Game Theory. I'm talking about things like the Prisoner's Dilemma, Pascal's Wager. They can be applied to all sorts of situations wherein you are uncertain about the state of one particular function. For example, you could apply it to the weather raining or not raining. Then your choices are wear rain appropriate clothes, or not. Ideally you want to only wear rain appropriate clothes when it's raining, but it comes down to what will bother you more, wearing rain clothes when you don't need them, or getting soaked in your normal clothes.

The other situation where they are interesting is when you're relatively sure of one of the states, but not certain, but the payoff for being correct is small or insignifcant compared to being wrong. For example, funding some unlikely to work research but for a defense purpose. Funding the research and having it pay off is huge, Funding the research for no pay off is not ideal, but on the other side, not funding the research and not having it pay off achieves nothing, but not funding and having it succeed is disasterous. Other situations turn up in social contexts as well.

I'll expand on this more later when I can actually draw some in here, it's too awkward to talk about them without

We spend a good deal of time learning how to perform certain tasks, so our brain can pretty much complete them on autopilot, freeing up our brain to do other tasks. The problem is that in doing so, our brain stops forming memories of that task unless they are significant, eg a car crash. So we repeat tasks and form habits and routines, so our brain can be lazy.

The paradox is this. Do we spend our time in the same routines, and remembering on highlights of our existence, but live efficiently. Or do we force ourselves to not use the easy pathways formed in our brain, so we're constantly learning and relearning and remember more in detail, but live less efficiently.

It's a tough concept for me to resolve.

http://www.wired.com/underwire/2013/03/nielsen-family-is-dead/

So I read this article tonight and it covers a number of highly relevant points. I'm not going to cover everything, but I do want to talk about the first point they bring up, that is, how network executives measure the popularity of a show is shifting. Rather than surveying the habits of random households, some companies are now just sampling the available data on twitter. It means going from a sample size of 25,000 with a good deal of selection bias to an entire, worldwide demographic of millions of people, also with some selection bias.

It speaks to the complacency of companies when they aren't constantly trying to innovate and improve, only maintain share price that they didn't start doing this sooner. What is really interesting though, is whilst it's starting with TV shows, there's no reason the traffic on twitter can't be used measure a whole range of social phenomena. Journalists are already tweeting details of events, court cases live and generating interest and discussion. As more people become involved, the information that can be gleamed from twitter analysis will become more reliable cross sections of public opinion.

There is definitely self selection bias though, the demographic of older people or those without reliable access to the internet are unable to make full use of the power of social media. It shouldn't be solely relied upon just yet, but for now it's large enough that institutions can no longer ignore it. I'd like to think that in 30 years time, when the older generation has passed away, that the entirety of the population could be reliably sampled through the internet, there'd be no need to rely on phone polls, face to face polls. Polling companies themselves would have shifted to big data analysis on trends from twitter or facebook. Public opinion on legislation could be reliably sampled by simple search query, or a tweet asking for opinions from a popular politician or celebrity.

I'm excited about the efficiency of the future, I can't wait for it to get here.

As part of writing, and being a social creature in general I'm often thinking about communication and how it actually works. There's a number of categories and ways to categorise communication. Do we split it up between verbal and nonverbal? Or do we sort it by the number and roles of the participants, one to one, one to many, many to many, many to one. Or can we sort it by the time delay between the sending and receiving of messages, from symmetric real time to asymmetric with a long delay. We can even split it up by ephemeral or non-ephemeral

I'll start with the one to one communications as they are the type that most people have, and can comfortably understand already. Within this classification, we can have several sub classifcations

• In person - verbal and non verbal
• Verbal only (phone or skype)
• Written only (texting, chatting)

We can also add the non realtime symmetric communications

• Verbal only (voice mail)
• Written only (email, letters)
• Video messages (contains verbal and nonverbal information)

Lastly, we have the non realtime, asymmetric communications

• Written only (Books, blog posts, documents)
• Verbal only (audio instructions, audio books)
• Video only (movies, tv shows, instructional videos)

Now I'm a bit skeptical of including movies and tv shows, as they wouldn't solely fall under a definition of a one to one conversation. They also can be one to many. It's hard to call them a conversation, but it can be classed as a one to one communication, except one party is delivering information, and the other party is receiving it. Anyway, there's far too many to cover in one post, but I'll focus on the first one, in person communication.

One to one conversations in person are probably the most important thing to get right, but the way we are taught in schools to communicate represents an idealised way we should communicate compared with the way we actually communicate. Typically, we spend a good deal of time learning about nouns, verbs, adjectives, sentence structure. The purpose of this is to develop mastery of the language. Now I can only write about Australian English, but I imagine most of what I'm about to discuss is relevant as well. We are taught to communicate using our words, since if everyone can agree on their definitions, then all ambiguity should be removed.

For the most part, this is correct, communication with words is what we should aim for, but where things get complicated is in all the non lexical communication. The tone, the pacing, the emphasis on certain syllables all delivers additional information. Additionally, the body language communicates information about the state people are in, opening up, making yourself appear larger communicates confidence, where as crossing arms, crouching, or making yourself smaller communicates insecurity and fear. The distance placed between people, the angle between them, the amount of eye contact all communicate information about how the two people perceive each other.

Now all this information is a lot to consciously consider, for the vast majority of people, it's mostly subconscious. If you had to actively manage your body language, your tone, your posture as well as choosing the correct words to say to effectively communicate your message, your brain would be fairly overloaded. This is where the ambiguity comes into one to one conversations in person. The words someone is saying might be what you want to hear, but the rest of what they communicate adds significant weight about whether or not you can trust those words. It is especially difficult when the words and the body language are communicating different messages, which one are you to believe? Is the person actively controlling their body language or is a result of their subconscious.

Over the years I've been doing a lot of reading about communication, body language, posture, non verbal cues and various other topics. One of the things I try to do now, is to simply make conscious observations during conversations about these things. What is my posture, what is theirs? Am I fidgeting or am I engaged with what they are saying? I view it as trying to peak into my subconscious to see what it is I like or dislike about someone. I also view as a challenge to peak into their subconscious/conscious to see how they feel about me and what I'm saying.

I'll definitely come back to this topic at some point, but that's all for now.

Well 2 days in a row followed by a long break isn't too bad. At least I remembered enough to feel guilty about not doing it.

Can't think of anything great to discuss at the moment. I'm finishing off my next paper to be published. It's a different kind of writing, significantly more restricted and a whole lot of editing to ensure the correct meaning comes across. I guess it's the same as any writing and communication though, you may have the ideas in your head, but unless you can communicate them clearly, efficiently they are worth nothing. Scientific writing requires one to be explicitly precise and at the same time, concise. This limitation poses significant challenges. Typically one could do the following for example.

The machine was left on overnight and there was a spill. It was fixed by turning it off in the morning.

With a sense of context, it is obvious the spill was not turned off. It, is referring to the machine. We could make this sentence more ambiguous by removing a part

The machine was left on overnight and there was a spill. It was fixed in the morning.

Here, we are left unsure if the machine being left on is the problem, or the spill is the problem, or both. Obviously avoiding ambiguity is what good writing should always aim to achieve. However, it can quickly become very verbose, especially when describing unnamed objects.

The brown machine was left on overnight and there was a spill that affected the black machine. The black machine was fixed, the brown machine was turned off and the spill was cleaned up in the morning.

In scientific writing however, items don't have simple or small names such as the brown machine or the black machine, they often have long, awkward names. This is partly due to the nomenclature used to describe molecules. It is insufficient to simply write glucose, there are two isomers, so we write D-glucose, but even then, there are two forms, so we write α-D-glucose. In regular writing, it would be sufficient to simply introduce it as α-D-glucose and then refer to it as glucose from that point on, only adding clarification if you later introduced a different form. As science papers are rarely read by scientists in order, it is a necessity to ensure that if a reader picks up the paper and starts reading from a given figure, that it remains unambiguous.

It is taking me a while to get used to writing in a style where each sentence I write must be able to stand on its own, such that a reader with some scientific knowledge can make sense of it without being confused by attempts to reduce the number of words on a page. Basically, nomenclature is really important. I'm definitely looking forward to being able to write freely