Cuevano ~ Jorge Aranda

Twilight Struggle and Go

For the past few years, I have been playing two boardgames online quite regularly: Twilight Struggle and Go. I usually play them asynchronously—Go in the Dragon Go Server, and Twilight Struggle through Steam (there is also an iPad app linking to the same server). I love making a move or two when I need a break, and though games tend to last weeks or months in both cases, I like how their arcs intersect with my daily life.

Frankly, I am still quite bad at Go, but it has been on my mind a lot recently due to the AlphaGo Zero developments. To beat Go masters, previous versions of AlphaGo (Google's AI Go player) used a combination of machine learning techniques, recorded historical games, and human expert input. But AlphaGo Zero started from scratch, with no reliance on recorded games. It reached a level high enough to beat the previous champions a few weeks after starting training.

It is by all accounts both the most advanced Go player we've seen, and the most foreign; not merely a better execution of known strategies, but a new and (in my case, impossibly) hard to understand approach to the game.

Now, I don't care so much that we've found a way to program computers to beat us at it (though that is very cool), and I don't care for the silly singularity fears and ecstasies that Artificial Intelligence triggers in some folks (except to the extent that this distracts their efforts from more important issues). But I do care about the idea that a boardgame that has perhaps been with us for four millenia has had this infusion of freshness and creativity in our lifetimes. It's as if aliens had independently discovered the game, too, and played it better than us, and, though we can't understand each other, they are letting us watch over their shoulders and learn.

Twilight Struggle, on the other hand, is not likely to continue being played four millenia from now: it is too historically specific, and its ruleset too fiddly to travel easily across generations. But these qualities do not affect the game in the here and now. It's a great game, complex and knotty, and since it's less well known than Go, I'd like to describe it briefly.

Twilight Struggle is an asymmetrical game about the Cold War, for two players. One plays the United States, the other, the Soviet Union. The Cold War in the game is driven by three main factors: first, the Domino Theory of ideological contagion and regime change, second, the inevitability of world events, and third, the spectre of nuclear annihilation.

Both powers attempt to maintain their presence around the globe, blocking their opponent from achieving an overwhelming dominance in any region. They have a glimpse at some of the events bubbling up (revolts in Vietnam, charismatic Communist or Capitalist regional leaders angling for power, the Cuban Missile Crisis), and must position themselves to take advantage of these events, kick them down the road, or at least try to mitigate them. And though, by the sick logic of Great Empires, they must assert their authority everywhere, sponsoring coups, meddling in local governments, and so on (if they don't, their opponent will), the more aggressively they act, the closer they bring the world to a nuclear holocaust.

This morbid dynamic of whether to push things right to the brink of destruction in an already chaotic world, just to deny your opponent any room to maneuver, is the stroke of genius of the game. Novice or reckless players end blowing up the world a lot—a gloomy realization these days—, and though with experience (and a good guide) one improves humanity's chances of survival, the spectre of annihilation gives the game a long-running tension.

There is randomness in Twilight Struggle, and between players with similar skill a couple of bad hands or a string of unlucky dice rolls may doom the less fortunate player's chances. But provided one knows this coming in, this randomness colours, rather than spoils, the game, as it adds to its sense of certain but limited power. Rather than luck, and just like Go, Twilight Struggle rewards study, patience, and prudence. I find it amazing that, given the asymmetry, the large amount of events in the game, and the complexity of the board and rules, the game remains well-balanced, tight, and engaging.

Both of these are "just games," of course, but a good game captures our imagination. There is a joy in discovering this kind of strong, sharp, robust design, in exploring it, grasping it, however imperfectly, and slowly making it part of our lives.

Horgan is the new BC Premier

A couple of months ago, worried about the agenda of the BC Greens leader and his likelihood to collaborate with the BC NDP, I wrote:

Andrew Weaver, please prove me wrong! Agree to a solid, stable collaboration with the BC NDP, one that will bring forth electoral reform, enact environmentally strong policies, and reverse the damage done by the current government. Many of us would be happy to give you our support in future elections if you do so.

Well, after a roller-coaster ride, John Horgan, the BC NDP leader, is the new Premier; Weaver proved me wrong; and I am extremely happy about all this.

Weaver's (and his party's) support for the new BC NDP minority government seems strong, and the Greens appear to have forced the NDP hand somewhat: the new government commits to an electoral reform referendum, to re-explore the viability of the Site C dam project, to stop oil pipelines construction projects in BC, and to a number of additional sensible policies.

This path to Horgan's premiership was absolutely unlikely, but the outcome is equally promising.

Unlikely, because it relied on very slim margins of victory in several ridings, as well as on the support of an opponent with which, from the outside at least, he seemed to hold a bitter rivalry.

And promising, not just because the policies of the new government are big steps in the right direction, nor because the Greens will help keep the NDP honest and on target, but because, in a time with a generally dark political outlook, the NDP/Green agreement shows the way towards inter-party collaboration in Canada. I have always been puzzled at the lack of collaboration under the parliamentary system here, which seems to have much less interplay than similar systems in Europe. I believe Canada suffers from some fear or aversion to rise to a collaborative pattern, and to the extent that the BC NDP/Green agreeement is successful, we will gradually learn to think of collaborative solutions by default.

(One additional source of hope: the arrival of Sonia Furstenau, a teacher and environmental activist turned politician, to the Legislature as the MLA from Cowichan and Deputy Green Leader. Best of luck to her, to Horgan, to Weaver, and to everyone in the new government.)

Scientific computing in 10 years

Greg Wilson, being the gloomy jaded pessimist he is, thinks leading-edge programmers will be doing scientific computing in JavaScript in 10 years. I disagree—I don't know what these programmers will be using then, but to the extent that they will be building on currently mainstream technologies I think it will continue being R and Python, rather than JavaScript. We placed a bet on it; this post is here for reference.

Going for him, Greg has the most popular language today, which, from a humble and rushed start in our browsers, continues to shift shape and burrow in the most unlikely places, and to prove all of its naysayers wrong. Against that, I have technologies that have ingrained themselves in the scientific community, which have fast, usable, and sophisticated modules that took many years to build and get right, and which offer newcomers a different career path and paradigm, spiraling out of data and formulas (rather than out of the browser) into the rest of the world.

Whoever wins the bet, it should be a fun dinner in 2027, and I'm looking forward to it.

Prove me wrong, Andrew Weaver!

After two weeks of waiting, the final results from the BC Provincial Election should be released today. The election was a nail-biter, with the BC Liberals nine votes away from a majority government and the BC Greens in a kingmaker role, that is, with enough votes to prop up a Liberal or NDP government, provided that the early results held, after absentee ballots and a final recount in the closest ridings.

So far, it looks like the early results will hold, which is good news—at the very least, we won't have a BC Liberal majority. At the federal level, as a Green, this would have been my dream scenario, but I have previously expressed reservations about Weaver's progressive values and willingness to work with the BC NDP. I felt that, if the BC Greens landed in this scenario, Weaver would rather collaborate with the centre-right party, perhaps extracting a couple of small concessions (spun like magnificent deals) for his support. I would love to be wrong. There is broad support for a left-wing coalition, and rejecting calls for cooperation would also damage the BC Greens' reputation as a progressive party, and the damage would be long-lasting and extend to the Federal Greens as well (one of my main fears and misgivings about Andrew Weaver).

So: Andrew Weaver, please prove me wrong! Agree to a solid, stable collaboration with the BC NDP, one that will bring forth electoral reform, enact environmentally strong policies, and reverse the damage done by the current government. Many of us would be happy to give you our support in future elections if you do so.

Site Migration

You may notice things are a bit different here: I have migrated from Wordpress to Jekyll, partly to minimize exposure to advertisers, partly to reduce hosting costs (the site is now hosted on S3), and partly to scratch the itch I've had for a while to explore static site generators and serverless hosting.

The one downside to this is that the new site does not have comments, which I used to enjoy quite a bit. I thought about adding Disqus, but this would mean allowing one more third party to track your browsing habits, and I would either have to pay for the service or allow them to display ads, which I did not want to do. So: if you want to comment on a post, I would welcome your thoughts directly on Twitter, Facebook, or over email or coffee!