30 December 2016

2017 Predictions

In the style of Superforecasting and the rationalists, some predictions for 2017, along with notes:

  • Donald Trump not removed from office by impeachment, assassination, or other means before 1 Jan 2018: 90%
    • The prior for removal from office is pretty low (9 of 44 presidents, ~20%; four of old age, four assassinated, and Nixon), so the survival rate for a 4-year term is at least 80%, and I'm only predicting for a single year.
    • Dems I know view Pence as a downgrade; assuming this is representative, Democratic party seems unlikely to initiate action.
    • I've been hearing rumors about the formation of leftwing militias ("antifascist squads"), but going from formation to assassinating a president in a single year would be remarkable.
    • Establishment Republicans may fear backlash from Trump electorate if they take public action to remove him.  I think old-school Republicans are a relatively likely source of impeachment proceedings, particularly if Trump's popularity ratings drop too far.
    • It's possible he could die of old age in office, but actuarial data give a 70-year old American man a 1.6% chance of dying this year (if I'm reading the table correctly), and I expect his healthcare to be better than average.
  • All thirteen of Trump's cabinet picks are approved by the Senate: 75%
    • Historically only 8 cabinet nominees have been blocked by the Senate, most recently by a Democratic Senate against Republican Bush I in 1988.  Assuming 13 nominees per previous 44 presidents, that's a base rate of 1.4% that a cabinet appointee will be rejected.  Granted, that's not conditioned on having a same-party Senate, but I'm not sure establishment Republicans totally count as friendly.
    • Since there are 13 nominees, the base rate that at least one will be rejected is (1 - 0.014) ^ 13, which is around 83% (again, not conditioned on friendly Senate).
    • Trump's nominees seem substantially more controversial than usual, and have drawn some criticism from establishment Republican senators.
  • Scalia's seat on the Supreme Court filled by 1 Jan 2018: 85%
  • Trump's average approval rating according to Gallup from 1 Jan 2017 to 1 Jan 2018 < 40% (currently 43%): 60%
  • DOW or NASDAQ down at least 20% in at least one single month period in 2017: 30%
    • We had a crash in 2000-2002, and then another in 2007-2009.  We're due for another one; the question as I see it is whether it'll be this year, next year, or the year after, which gives me a 33% base probability.
    • Plausible causes: 
      • The Economist tells me that banking reforms following 2008 crisis were inadequate; I don't grok the details.
      • Trade war with China or other Trump administration actions.
      • Terrorism
      • Reevalution of tech unicorns (dot com crash 2.0)
  • Deployment of nuclear weapons against armed forces or civilians by any nation-state actor: 2.5%
    • Our prior is one year in the 70 since the invention of the nuke, which is about 1.4%, but things seem a little tenser than usual, and that prior doesn't take into account nuclear proliferation to eg North Korea, Israel, Pakistan, possibly Iran...
    • Still, I'd be very surprised.  That would be a big shift in the current order.
  • Britain leaves the EU: 70%
    • May expects to trigger article 50 in the spring, but will then have two years to conclude negotiations with the EU.
    • I expect Belgium to try to give Britain the shittiest possible terms for leaving in order to set a precedent for other countries considering leaving.  The British Leavers might stall in order to strengthen their negotiating position, which could delay the exit into 2018 or 2019, but at the risk of losing power if they're perceived by the British public as dragging their feet.
    • In short, I think a 2017 Brexit is probable but not at all a sure thing.
  • France elects a right-wing government (Les Republicans or Front National): 60%
    • Seems to be the way of things recently
    • Hollande not seeking reelection is not a great sign for the Socialists
    • Historically France has elected more right-wing than left-wing presidents (in this iteration of French government, anyway)
    • I expect Fillon to take a beating from the public / government employee voting bloc, but still might happen
    • I absolutely expect further Russian interference in support of Le Pen.  Fillon is also relatively friendly with Russia and could receive support.
  • I know jack about German politics so I'm going to leave that alone.
  • Autonomous car kills a pedestrian or cyclist: 65%
    • I expect to see a big increase in deployment this year barring government interference.
  • US Navy conducts >= 4 Freedom of Navigation operations in the South China Sea: 85%
    • They conducted at least four this year that I could find easily
    • I do not expect Trump to deescalate
  • No US servicemen killed or wounded in South China Sea / China region: 83%
  • "International incident" with US bodycount in South China Sea / China region: 10%
  • Limited military intervention against Chinese assets in SCS: 5%
  • Large-scale war (>=1000 US casualties) with China, including proxy war: 2%
  • At least one Islamist terrorist attack in EU or Britain with killed+wounded >= 500: 80%
    • There was one in 2015, one in 2016
    • ISIS is kind of crumbling - this could be good or bad, introduces uncertainty.
    • Expected emergence of right-wing/nationalist governments in Europe seems likely to fan resentment, aid recruitment
  • At least one Islamist terrorist attack in EU, Britain, or USA with killed+wounded >= 1000: 10%
    • 9/11 and the 2004 Madrid bombing are our only datapoints; 2 in 15 years -> ~15% base
    • Trend is definitely downwards, though
I'm not planning to take bets on these; this is my calibration / training year.

29 December 2016

Hwoongth?

All the good domain names were taken, so I wrote a script to generate syllables in accordance with english phonotactics.  It generated "hwoongth", which made me laugh.

The function of this blog, then, is the playful pursuit of intellectual interests.

Topics in scope (an inexhaustive list):

  • Software "engineering"
  • Computer insecurity
  • Strategy
  • Forecasting
  • Politics (raised libertarian, now doubting.  I expect to disparage everyone)
  • History
  • Sociology
  • Gnosis and other brainbugs