Sunday, 29 July 2012

Bike Racing and Cooperation...

I've been writing blogs on how the Prisoner's Dilemma can be seen in action in the banking crisis (It's Only Taken Three Years...) and in the housing market (Games Theory and the Estate Agent) - and yesterday we saw Games Theory ideas in action in the Olympic road race. 

It's Games Theory that drives the plot of my first novel, The Defector, and in particular a thing called the Prisoner's Dilemma (PD). If you haven't come across it before, I will point you at my own description in the foreword to The Defector, a suspense thriller in which it features as the central plot device. Or you can check out a much more technical take in the Stanford Encyclopaedia of Philosophy (SEP) entry.

If you're going to read on, please get to grips with the Prisoner's Dilemma first!

The peloton is a place where everyone has to decide whether to cooperate or defect. The co-operators take their turn at the front, while the defectors hide in the bunch, freewheeling in the slipstream and hoping to conserve their energy for the sprint at the end.

We can see this in PD terms - all the cooperators give themselves the same chance at winning if they all do even amounts of work at the front. But there's a big benefit to defecting when everyone else cooperates, as the energy conserved would give you a massive advantage in the final sprint.

If it was that simple, it would turn into a slow bike race pretty quickly, as everyone would defect and huddle into the centre like penguins in the Antarctic. What makes it more complex is the fact that you can defect in a different way, by trying for a break-away. If the peloton dawdles then one or more riders have the opportunity to sprint away from the group and build a lead that can't be broken down before the finish.

This is another form of defection. Instead of cooperating and riding together to break the back of the 150+ miles - and then seeing who's the strongest and fastest at the end - let's just see who's strongest by riding hard and trying to break the peloton up the whole way. Until we have a last man standing.

This scenario is made more complex because the riders are working in smaller teams, and those teams have different interests depending on the make-up of their team. The teams with the best sprinters have the biggest interest in the race finishing with everyone in a single bunch. So a race would normally develop with the teams with sprinters cooperating to try to control the peloton and keep them together, taking it in turns at the front of the peloton to haul in any breakaways.

Meanwhile, those teams lacking sprinting power will defect - not take any of the load, and do everything they can to get one of their teammates into a decent breakaway.

What happened yesterday was unusual, in that only one team was interested in the peloton finishing together in a mass-bunch sprint. And that was Team GBR. Everyone in the race knew that Mark Cavendish is the best sprinter in the world, and that he would almost certainly win a bunch sprint to take gold. They all figured that the normal reward for cooperation had evaporated - taking it in turns at the front was pointless, as Cavendish would win the resulting sprint.

And so the normal rules went out the window, they all defected, either tucking into the peloton to conserve energy and see what happened, or constantly trying to engineer a break-away, but... But perhaps this was actually a form of cooperation. The rest of the peloton shared an interest in breaking the normal race strategy - defection became cooperation, and vice versa.

The outcome was pretty inevitable - Cavendish had around him the strongest individual riders in the world. But faced with an entire peloton unwilling to cooperate in engineering a massed bunch sprint at the end, it was too much work. Eventually, one of those breakaways was going to work - and in the end, it did. It was followed by a couple more, and the group splintered until there was only one man in the final breakaway. And that was Alexandr Vinokurov, gold medallist and last man standing.

And that will definitely be it for this blog until September... I'll see you back here in the autumn.

Wednesday, 25 July 2012

Games Theory and the Estate Agent

Just before Christmas last year I wrote a blog called 'It's Only Taken Three Years...' about the banking crisis and how the Games Theory ideas that drive the plot of my first novel, The Defector, might be applied to find a solution. The blog was picked up by the crowd-sourced news website, and was featured on their home page.

One of the comments that the article attracted was that the solution I'd suggested was pretty impractical as it required a far more engaged population than we will ever have - something I'd tacitly admitted in the article. But it got me thinking about other ways that Games Theory might be used for social good. And the one that immediately sprung to mind - not least because I've just been buying a house - was the problem of shocking behaviour in property transactions.

We've all heard the horror stories, and late price rises, people dropping out of sales after the other side have spent hundreds of pounds on surveys and legal fees. In the UK, the last Labour government had a go at fixing this with their woeful and now abandoned housing information packs. But they were going about it the wrong way - it's not the house that we need more information about, it's the people on the other side of the transaction, the buyers or sellers.

If you haven't come across the Prisoner's Dilemma (PD) before I will point you at my own description in the foreword to The Defector, a suspense thriller in which it features as the central plot device. Or if you want something a bit more technical and meaty, take a look at the Stanford Encyclopaedia of Philosophy (SEP) entry.

We can draw a couple of conclusions about the Prisoner's Dilemma;  for the individual, rational, self-interested player in a one-off game of PD there’s only one real choice – Defection. However, things change in what's called an iterated version of the game, this is what the SEP entry has to say about it.

"Many of the situations that are alleged to have the structure of the PD, like defense appropriations of military rivals or price setting for duopolistic firms are better modelled by an iterated version of the game in which players play the PD repeatedly, retaining access at each round to the results of all previous rounds. In these iterated PDs (hence forth IPDs) players who defect in one round can be “punished” by defections in subsequent rounds and those who cooperate can be rewarded by cooperation. Thus the appropriate strategy for rationally self-interested players is no longer obvious."

This is where the impulse to cooperation comes from for the rational self-interested player; the knowledge that other players will judge you on your previous behaviour. So what does this mean for buying and selling houses? The problem with a property transaction is that it’s a one-time PD game. It’s very unlikely that you will ever conduct more than one residential property transaction with the same individual. And so you can behave as badly as you like to get the outcome that you want, with no consequences. The only people who will ever know that you gazumped your way to a better deal are the other parties to the contract, the solicitors and the estate agents. So long as you stay within the law, there are no consequences for bad behaviour, outside of your own conscience.

To clean up the housing market we need to turn each housing transaction into an iterated Prisoner's Dilemma, where there is a much stronger impulse to cooperation. What’s required is the knowledge that any poor behaviour will carry forward to the next transaction. If all estate agents or solicitors (or both) were compelled to record the behaviour of those involved in each property transaction on a publicly accessible website database, then the next party to a deal with any given individual would have a much better basis for deciding whether or not to proceed.

Imagine you’ve just dropped the asking price on your house, and it’s produced  a couple of offers. One is from someone who’s bought three houses, all of them in a perfectly straight-forward manner, and who was regarded as quick, efficient and easy to deal with by the solicitors concerned. The second is from someone who’s been involved in nine house purchases, gazumped the other party on three of them, forced a late price change in four, and dropped out of the other two purchases just before exchange of contracts.

It would be pretty clear which offer to accept wouldn’t it? As I said earlier, what the government needs to provide is not more information on the house, but more information on the other party to the transaction.  Once you have that, I suspect there would be a lot less bad behaviour in the housing market.

And that will be it for this blog until September, I'll be blogging for the two weeks of the London Games at the ISAF Olympic website -- and then I'm on holiday. I'll see you back here in the autumn.