ex machina

If I don’t get tempted back into finance, I’d like to look at mathematically modelling aspects of human behaviour.  The point here isn’t to predict what individuals will do, or even necessarily to get accurate quantitative predictions about aggregate behaviour; more to develop models that capture qualitative features of human and political behaviour and to get some idea about the stability of those features.  Are asset price bubbles a stable attractor?  How well is the number of political parties in a system predicted by the electoral system?

This is kind of what economists do already, although I expect that one can probably throw a lot more mathematical sophistication at the problem. I’m not the first person to think about attempting this and, if I was a better researcher, I’d probably be reading the report from this conference, rather than randomly pontificating about my own ideas.  But I’m not, so here are some of them …

After listening to some public health people talking about BSE and CJD I began to think about whether one could get an idea for when cattle ranchers would comply with BSE reporting rules and when they would “shoot, shovel and shut up“?  What are the effects of the similar incentives for bureaucrats / political decision makers to cover things up?  Can one use these models to estimate the maximum possible levels of under-reporting of BSE and does that effect predictions about the level of a possible CJD epidemic in Canada?

Mathematically, we expect a “phase transition” – the first farmer who reports a mad cow in a province will probably never farm again.  Regardless of the level of financial compensation available, this is a pretty high cost.  But the thousandth farmer to report will probably not make the local newspaper, and so can cull, and rebuild, their herd with little stigma.  So once reporting hits a certain level, one would expect a fairly sudden switch from a “tendency not to report” to a “generally content to report” state.  So two regimes with the same levels of epidemic could, in fact, have wildly differing reported levels of epidemic depending on whether the phase transition had taken place.  The point at which this happens will be parameterised by not only the level of an epidemic and the amount of compensation available (and level of punishment for non-reporting) but also by farmers’ perceptions of the future willingness of politicians and bureaucrats to tolerate non-reporting (and the interest of the media in exposing it) and their perception of the future likelihood of other farmers failing to report.  These two last factors will be affected by the current decisions made by farmers, so we get an interesting interaction between current events and uncertain future events.  We can capture this kind of thing using similar maths to that used to model derivative financial instruments (I know that’s not a great advertisement right at the moment, but they are still useful tools!)

Similar dynamics might be exhibited in the behaviour of politicians and regulators. When perceptions of the levels of an epidemic (or any nasty and growing event – housing bubble?) are low, the politician who advocates thorough monitoring and regulation will likely be looked on unfavourably.  But at higher perceived levels – as the interested parties expand beyond the farmers to the general population, thoroughness and action are likely to be rewarded.  The politician who acts too soon will likely be ruined before they are vindicated; the politician who acts too late will be blamed for not acting soon enough.  Politically, I guess that the right time to move is probably just before everyone else does, which, mathematically, is a pretty unstable thing to predict.  However, one might be able to say something about the statistical distribution of behaviours.  I wonder if one could validate this kind of prediction against the behaviours exhibited by different governments in response to swine flu?

Lockhart’s Lament: a very readable essay about the difference between mathematics and most mathematics education.  Along the way, explains why mathematicians look a bit pained when people tell us that they weren’t very good at maths at school.  Really, it’s very good.  Thoroughly worth a read, even for (especially for) non-mathematicians.

A bit of context

Oh, and I do.

One of the (many) new things that I’ve learned since Arthur arrived is how many songs I don’t really know properly (and how tedious Fly Me To the Moon can get after you sing it five times in a row).

I struck a new vein yesterday when I remembered Woodcraft songs.  Happily, the district fellows (16-20 year old Woodcrafters) have a song list on their website.

Here’s Arthur enjoying Bog down in the Valley O.

From the Economist tribute to John Maddox, rescuer of Nature:

Writing any serious tribute to the science journalist and editor, Sir John Maddox, can only really start a long time after deadline, with the assistance of a cigarette and a glass of wine.  Thus he famously began work on his editorials. Some were delivered so late that their first lines were being typeset before the last had been composed.

Finished a draft of a paper today.  Sprog better get moving if it’s going to win the race to delivery.

According to some of the commenters at BoingBoing, their service is uniformly shitty.  But still …

B’nai Brith published their 2008 audit of anti-semitic incidents in Canada yesterday and were touting a year-on-year increase in anti-semitism in Canada.  B’nai Brith is the self-appointed independent voice of the Jewish community in Canada and is consistently supportive and uncritical of Israel.  The actions of Israel as a nation state are perceived by many people as being criminal and unjust and, as such, B’nai Brith’s efforts to conflate the interests of Israel with those of all Jews probably don’t do a great deal to help counter anti-semitism.

Here are a couple of non entirely satisfactorily connected comments on the audit:

Anti-semitic incidents in Canada 1999-2008

The consistent year on year increases in harassment year on year are startling.  Has anti-semitism in Canada quadrupled since 2001, or is there a less scary reason?  The clue is here.  Web-based hate accounts for 405 incidents of harassment.  Looking back through previous reports, we see that in 2007, there were 310 incidents, 2006: 254, 2005: 164, 2004: 47, 2003: 32.  Prior to 2002, web-based incidents are not broken out.

Lets break them out:

Anti-semitic harassment broken down by web and non-web based

We can do the same thing for total incidents:

Anti-semitic incidents broken down as web and non-web based

In both cases, excluding the web-based harassment makes the total number of incidents appear to have been stable since 2004, not increasing.

The searchability of the internet makes it easy for people to actively seek out offensive material and report it.  People can now come across material incidentally that previously they would never have noticed.  Improved communication methods like e-mail and message boards means that it’s more likely that people will encounter offensive material after being referred to it in the context of “this is bad, we should do something about it”. In 2002, B’nai Brith reported that “Many more sites and chat rooms are monitored on an ongoing basis”.  In short, improved communication makes it more likely for anti-semitism to be exposed.  An increase in reported harassment does not necessarily mean that there are more people doing the harassing.

This is not to say that web-based harassment is insignificant.  People are spending more and more time online and anti-semitism is unpleasant and dangerous in any form.  However, any conclusions about trends based on figures which include web-based harassment probably are insignificant.

Any survey methodology that relies on passive reporting will be subject to reporting bias – for example, an increase in fearfulness could cause extra vigilance in reporting.  The reports make no attempt to control for this.  However, plausible reasons for real increases occurring from 2002-2004 are presented. For instance, from 2004:

The total number of incidents for March and April combined – 275 – amounted to close to one-third of the incidents for 2004. This timeframe as a whole corresponds to a period of heightened activity by Hamas and the resulting counter-terrorism operations by Israel, including the assassination of Yassin and, subsequently on April 19, 2004, of the leader who replaced him, Abdel Aziz Rantisi. The spike in March, and the continuation of heightened antisemitic activity well into April, could well be explained by a dynamic that saw elements sympathetic to Hamas – which has been outlawed as a terrorist organization by Canada – acting out their anger against individual Jews or community targets.

It’s worth noting that the reports are conscious of the difference between anti-semitism and anti-Zionism.  From 2008:

Incidents that involve or reflect an increasing worldwide trend toward virulent anti-Zionist rhetoric have been included only if there is a clear anti-Jewish component.

So, surprisingly, are some of their spokespeople:

Mr. Kurz … added that while not everyone who is critical of Israeli policy is necessarily an anti-Semite, that criticism “feeds an atmosphere in which anti-Semitism has a great deal of fresh air”

Which begs the question: who is responsible for that?  Canadians don’t hold Americans (or Brits) responsible for the catastrophe of Iraq, they hold Bush, Cheney, Blair, et al. responsible.  Canadians don’t hold Catholics responsible for the idiocies of the pope.  So, if there weren’t groups of noisy people claiming to represent all Jews and at the same time defending Israel against any criticism, don’t you think that Canadians might manage to tell the difference there too?

The problem with conventional cars is that they are engineered to be able to travel 1000s of kilometres at high speed, but most  journeys are just a couple of dozen kilometres, at low speed.  The problem with electric cars is that they aren’t that great for taking on a canoe trip, or driving from Toronto to Montreal; this is partly because they don’t have great range, and partly because recharging them takes a long time.  The problem with hydrogen cars is that they don’t exist, but even when they do, there’s going to be a problem driving them very far from a hydrogen fueling station.

Most people can’t afford to have multiple cars to deal with all these different use cases, so they opt for the lowest common denominator, which, for the foreseeable future is going to be a conventional petrol fueled engine, or maybe a hybrid.

As a solution, I propose: modular engines.

The idea is that each car comes as an electric vehicle with a battery pack and an all-electric drive train (i.e. electric motors directly driving 2 or 4 of the wheels) with regenerative braking.  It then has one or more “bays” in which additional power modules can be installed.  Each bay has electrical contacts and a connection to cooling and exhaust systems.

When you’re using the car as a commuting vehicle, you stick a battery module into the bay(s).  When you go on a long distance trip, you install a diesel, or petrol, generator module.  If a government or a car manufacturer chooses your region to trial a hydrogen fuel distribution network, then you can install a fuel cell module. For drag racing at the weekend, install an ultra-capacitor (and make sure that your electric motors can handle the current).

The point here is that you avoid both the compromise problem – your car is all electric when it can be and hybrid when it needs to be and the chicken and egg problem – achieving the critical mass of hydrogen-powerable vehicles necessary to sustain a hydrogen delivery network is a lot easier when the vehicles can also use an alternative power when they are outside the range of the network. With the right infrastructure in place, the vehicles will be very flexible; without that infrastructure, the vehicles will be just like other vehicles.

Some more details:

• The weight of the modules would be at the “two strong people to lift it” point.  The idea is that you won’t be swapping modules frequently, but at the same time, you won’t need specialist equipment.  Of course, this means that a diesel generator is going to be significantly smaller than a typical motor in a car engine.  Part of the effect of this will be offset by the fact that a diesel generator, running at a fixed rpm, will be a lot more efficient, and have a higher power output, than a motor designed to provide torque over a range of speeds.
• You wouldn’t necessarily own your modules – you might rent them.  Particularly the ones that you won’t use very often.  This makes servicing easier too.  Car manufacturers or third parties could rent out and maintain “fleets” of modules.  They would guarantee certain performance characteristics for the module and your car would monitor that.  When the module needs servicing, your car tells you, you take it back and get a new one.
• On a shorter timescale, fuel stations could keep a stock of charged battery-modules, so rather than filling up with fuel, you swap out the empty batteries and put in new ones.  This would require “module handling” technology at the fuel station, because you won’t necessarily have 2 strong people on hand to help, but any technology like that is likely to be a lot cheaper than existing fuel station technology for the storage of fuel.
• Larger cars could have more bays for more power.
• Modules will need to be “smart”.  Different cars are going to be able to provide them with different levels of cooling; with multiple modules, it will be necessary to optimise how they are used.  None of this will be necessarily any more difficult than existing engine management.

Problems:

• As I mentioned above, the power output per module is a problem.  I don’t know what kind of power output you can get from a diesel generator weighing 80kg.
• Fuel storage.  Should the fuel tank be part of the module? Pros: makes module insertion / removal easier (fewer things to connect).  Cons: increased weight, less space for other things, proximity of fuel to (hot) combustion engine.  If you had the fuel tank part of the base car, you’d need to engineer some kind of universal fuel delivery system.
• Cooling – is air cooling going to be good enough?  Like the fuel, liquid interchange with modules would add engineering complexity.
• Exhaust – the exhaust system in conventional cars affects the performance of the engine.  Maybe with generators this is less significant?

Assuming that I am not charged with an indictable offence between now and the ceremony, I will become a Canadian citizen tomorrow morning.

Should I be proud to become a Canadian?  There are things about Canada that I like a lot, but there are also things that I dislike and, just like most of the aspects of Britain that I like or dislike: I’m not responsible for any of them.  I’m not responsible for the BBC and I’m not responsible for the invention of concentration camps.  Canada has become a successful multi-cultural society without any help from me, but it also screwed over a lot of aboriginal people without my help too.  I can’t see how I can accept responsibility for the latter; in which case, isn’t it disingenous to take personal pride in the former?

I applaud Canada for its relatively progressive immigration policy, and I’m personally grateful that there have been citizens of Canada who have voted for that (or at least failed to vote against it), but a surfeit of gratitude is dangerously close to complacency.  Canadians’ thankfulness that their health service is not as screwed up as that to the south means that they forgive their own politicians’ inability to deliver a truly universal and comprehensive service at home.  Does the thankfulness that many Canadians have for having been able to escape oppressive regimes lead to their tolerance of the less oppressive but fiercely unaccountable policing here in Canada?

I hope that I will be a good citizen, but I don’t intend to be a good Canadian citizen by wearing a flag pin or pretending that the mutual conflation of Canadian identity and a chain of coffee shops founded by a dead hockey player is of benefit to anyone other than said coffee shops’ marketing department.  Being a good citizen is about making things in your society better.  To do that, you first have to identify and acknowledge where there is room for improvement.

According to Peter Kent, a junior Canadian foreign affairs minister, Hamas

… has made a habit of using civilians and civilian infrastructure as shields for their terrorist activities

and

… behaves as if they are trying to have more of their people killed to make a terrible terrorist point.

Is this true?  Does Hamas “bear the full responsibility for the deepening humanitarian tragedy”?

The claim that civilians were being used as human shields is often made by the Israeli government and military to explain their deaths.  The shelling of civilians resulting in 106 deaths sheltering at a UN compound in Qana is a good example of this.  There were Hizbollah fighters firing mortars from a site a few hundred metres from the compound, but the UN major general who investigated the site 2 days later found that

“The pattern of impacts is inconsistent with a normal overshooting of the declared target (the mortar site) by a few rounds, as suggested by the Israeli forces.”

In similar form, the bombing of a UN school in Gaza on Tuesday was initially blamed on Hamas militants firing mortars from the location.  Yet today the Israeli army admitted that the Hamas mortar fire came from outside the school compound and the school was hit when one of the mortars fired by the Israeli Defense Forces in response went astray. Two of the mortars fired in response hit their target, which means that the soldiers on the ground must have been fully aware at the time that the original fire did not come from inside the school.  The initial denial of that by the IDF looks like a deliberate lie or at least a deliberate policy of making no attempt to verify statements before making them.

In the Qana case, the killing of the civilians could have been deliberate; in the case of the UN school, it seems most likely that it was a mistake.  In both cases, the initial reaction of the IDF and the Israeli authorities was to lie and to accuse their enemy of using civilian human shields.  A bit of googling reveals plenty of other examples of accusations of Israel lying in order to divert blame for its killing of civilians (and, to be fair, it’s not just Israel.  When countries go to war, they kill civilians and, understandably, tend not to want to own up to it).

That there are many well-documented examples of this suggests that that one should treat the Israeli army and government’s excuses for civilian deaths with caution.  A pattern of Israelis talking about Hamas using human shields does not imply a pattern of Hamas using human shields.

[This was going to be a longer post that looked at more than just the idiotic statements of Canwest-arse-licking politicians, but attempting to write fairly about this stuff is tiring.  Here's an interview with Robert Fisk that's too good not to link to.  In particular the second response, third paragraph.]

David’s dad, Cecil, was on The Current today, providing follow up after Ken Silverstein was on talking about his expose of Washington lobbyists.  I thought that he did a good, if a little conservative, job.

He also got to pimp his book .

Part of the purpose of this blog is to provide an alternative outlet for rants that I would otherwise send to the CBC written in green ink, but having ones father-in-law-in-law on the show is a special occasion and I was unable to resist.

Ken Silverstein’s investigation into the dirty side of Washington lobbyists clearly embarrassed them into attempting to divert attention away from his findings and onto his methods.

It’s a bit disappointing, then, that you spent about 5 minutes on the subject of Silverstein’s work and then 20 minutes (5 with Silverstein, 15 with Cecil Rosner) on journalistic integrity.

The lobbyists must be doubly pleased.  Your participation in their demonstration of how to deflect negative publicity by distracting the media meant not only that you spent little time on the embarrassing expose, but also provided a beautiful marketing example for them to pitch to future clients.

And all for a non-story.  Silverstein didn’t break the law and, maybe more importantly, he didn’t break any commitments of journalistic confidentiality.  The lobbyists were unaware they were talking to a journalist, so he will not have compromised future journalists’ ability to obtain off the record information.

The only substantive criticism was that Silverstein’s deception in order to get a story is somehow morally wrong.  But if a trade off must be made, shouldn’t a journalist’s primary concern be the amount of truth in a story rather than the amount of untruth employed to obtain it?