New Year: 2016 Prediction Calibration Results

In the spirit of #SkinInTheGame, Taleb’s idea that pundits should at least stake their reputations on the strength of their knowledge, last year I made some predictions about what has come to be known as The Current Year.

Like Scott Alexander, I am calibrating my predictions by comparing the percentage of predictions I got right at each probability level versus their probability (e.g., for predictions at the 70% confidence level, perfect calibration would represent getting 7/10 of them correct). Predictions with a probability rating of less than 50% are converted to their inverse.

Correct predictions are left as is, while wrong predictions are crossed out.


Conflict

(1) The Syrian government will control a larger proportion of territory in a year’s time relative to today: 80%. Gains in Latakia and the capture of Aleppo, but ironically, pushed back further in Palmyra than at the same period last year. Though the strategic value of Aleppo cancels out Palmyra tenfold, in technical terms this is still a failed prediction. My main area of uncertainty was regarding Turkish or Western intervention against Assad. In truth, the sadder and more banal reality is that outside a few elite units the SAA remains mostly worthless.

(2) A majority of these happen: (a) SAA liberates Deir Hafir; (b) Palmyra; (c) All of Latakia; (d) Links up with the Nubl pocket; (e) Maintains hold on Deir ez-Zor airport. 80%Deir Hafir is still under Islamic State, while small bits of Latakia are still controlled by the rebels. The Nubl pocket was linked up with, and Deir ez-Zor airport is still under Syrian control. Though recently recaptured, Palmyra was still liberated, so this is technically a correct prediction.

(3) Assad will remain President of Syria: 90%YES.

(4) The Iraqi government will control a larger proportion of territory in a year’s time relative to today: 90%YES.

(5) Islamic State will continue to lose ground in its heartlands and might end the year controlling little more than its capitals, but its overseas franchises – most notably in Libya – will expand further: 50%. Has been all but excised from Libya.

(6) The Houthis gain ground in Yemen: 60%. Comparing the maps between Dec 2015 and today, the Houthis seem to have lost ground, although very marginally. I should probably stop making predictions about wars and places I know very little about.

(7) The War in Donbass reignites: 30%INVERSE 70% it doesn’t: YES.

(8) Mariupol ends the year in DNR hands: 10%INVERSE 90% it doesn’t: YES.

(9) “Putinsliv” aka Putin abandons support for DNR/LNR and Ukraine recaptures them: 5%INVERSE 95% Putinsliv doesn’t happen: YES. That said, nobody really expected this apart from the more zrada-anticipating Russian nationalists.

(10) A new conflict in the former Soviet space: 20%INVERSE 80% no conflict: YES. Actually I outright said the most likely place for that would be Armenia vs. Azerbaijan in Nagorno-Karabakh, and it came to pass, though not at a large enough scale to quality as a conflict.

Russia/Eurasia

(1) Politics – The Russian Duma elections are slated for September 2016. United Russia will comfortably take a majority of the seats: 95%YES. UR took 343/450 = 76% of the seats (I predicted 80%).

(2) Politics – Electoral falsifications will be less than in the 2011 Duma elections: 70%YES. There was much more than I expected – I was expecting the introduction of a partial FTPT system to greatly reduce this problem – but fewer than in 2011 nonetheless.

(3) Economics – The recession will end in 2016: 80%YES. Multiple indicators suggest this has indeed happened in the second half of the year, so I am willing to call this as a win.

(4) Economics – There will be overall positive GDP growth in 2016: 60%. Nope.

(5) Ukraine – The recession will end in 2016: 70%. YES. Ultimately, its so depressed that there’s hardly any room to fall further.

(6) Ukraine – The Poroshenko regime remains in power: 80%YES.

(7) Demographics – Russia will see natural population growth: 40%. INVERSE 60% there will be no natural growth. I was wrong – according to preliminary figures, the Russian population increased by 18K to November, relative to 24K in the same period last year. The population also almost always grows in December.

(8) Demographics – Russia will see population growth: 95%YES.

(9) Demographics – Life expectancy will increase: 80%. The mortality rate continues falling, a modest 1.3%, and the population isn’t getting any younger, so that’s a YES. It will probably be around 72 years in 2016, just as I predicted.

(10) Demographics – TFR will increase: 50%This is currently very hard to assess. The number of births fell by 1.7%, but its well known that the number of women in their childbearing years is still falling, so overall the two effects will have almost perfectly canceled out. Therefore I am not yet in a position to rate this prediction. The total TFR for 2016 will certainly be in the TFR = 1.75-1.8 range, just as in the past two years.

World

(1) US/Allies will impose no fly zone (i.e. attack Assad) over Syria: 10%. INVERSE 90% will not. YES.

(2) US will not get involved in any new major war with death toll of > 100 US soldiers: 90%YES.

(3) An Islamic terrorist attack in Europe causing more than 100 deaths: 30%INVERSE 70% there will not be. YES. The Nice attacks killed 86 people.

(4) Brexit: 10%. INVERSE 90% no Brexit. Well that’s a fail, though at least I did up it to more than 50% a week before the referendum.

(5) The Euro is here to stay: 90%YES.

(6) China will not go into recession or have a hard landing: 90%YES.

(7) End of Western sanctions against Russia: 10%INVERSE 90% sanctions remain. YES.

(8) Israel will not get in a large-scale war (i.e. >100 Israeli deaths) with any Arab state: 90%YES.

(9) North Korea’s government will survive the year without large civil war/revolt: 95%YES.

(10) Oil prices will NOT end the year below $40: 70%YES. WTI Crude is currently at $53, which is about what I expected. Note that many analysts were predicting $20 oil.

(11) Will be hottest year on record thus far: 80%YES. Absolutely, and by a large margin. The Arctic is in absolute meltdown.

(12) No further large scale civil wars/revolts/revolutions in Middle East/North African countries not already so afflicted: 70%YES.

(13) No large scale civil wars/revolts/revolutions in USA: 99%YES.

(14) No large scale civil wars/revolts/revolutions in China: 99%YES.

(15) No large scale civil wars/revolts/revolutions in Russia: 95%YES.

(16) No large scale civil wars/revolts/revolutions in any EU country: 90%YES.

(17) China tops 2016 Olympics Gold medals table: 40%. INVERSE 60% it will not. YES.

(18) Germany will win UEFA Euro 2016: 30%. INVERSE 70% it will not. YES.

(19) Russia will predictably disappoint at UEFA Euro 2016 and will get knocked out at the group stage: 50%YES. “… But Russia fans are regularly schooled on the dangers of abandoning pessimism” – indeed!

(20) Mount & Blade II: Bannerlord is released: 80%. The Turk disappoints. It’s my solitaire goddamit, get a move on!

USA

(1) Donald Trump will secure the Republican nomination: 40%. INVERSE 60% Trump won’t get it. 

(2) Hillary Clinton will secure the Democratic nomination: 90%YES.

(3) Hillary Clinton becomes US President: 70%. Do note that I raised my assessment to 50% by June 2016 in a discussion with Razib Khan, and to 60-70% by September 2016 (no Internet record of it but ask Mike Johnson or Scott Jackisch), but then my nerve failed at the last moment (even though my final prediction of a 291-247 HRC win in the Electoral Collage was closer than that of most analysts, and I even got Michigan right). I may not be the God-Emperor’s psyker like the brilliant Scott Adams but being wrong was never sweeter.

(4) The US enters recession: 20%INVERSE 80% no recession. YES.

(5) Peak SJW?: No percentage due to inability to measure. My impression is that “peak SJW” has indeed passed, at least for now. Do you agree?

Myself

(1) I will write a record amount of blog posts: 70%. 127 to 130 last year. 

(2) I will author or coathor an academic paper: 60%. Currently collaborating on an S factor analysis of Russia with Emil Kirkegaard

(3) I will finish writing at least one book: 30%INVERSE 70% no booksYES.

(4) I will finally heed the advice of my detractors and fuck off back to Russia: 90%YES.

(5) I will end up being underconfident on these predictions: 50%Seems evenly calibrated. But the New Year is arriving in 15 minutes, so I don’t have time to calculate the exact calibration, so most fortuitously my two responses at the 50% confidence level remain exactly 50% correct. Wasn’t a great idea to have three questions at this confidence level!!


Here is my calibration graph:

predictions-calibration-2016

I got really unlucky on my 60% confidence level predictions.

Here is what happens where all the 60% and 70% confidence predictions are combined into one 65% confidence level:

predictions-calibration-2016-adjusted

Not to blow my own horn, but this is some impressive calibration.

Anatoly Karlin is a transhumanist interested in psychometrics, life extension, UBI, crypto/network states, X risks, and ushering in the Biosingularity.

 

Inventor of Idiot’s Limbo, the Katechon Hypothesis, and Elite Human Capital.

 

Apart from writing booksreviewstravel writing, and sundry blogging, I Tweet at @powerfultakes and run a Substack newsletter.

Comments

  1. “Not to blow my own horn, but this is some impressive calibration.”

    Very interesting – thanks.
    But I’d just point out that knowing the right odds for your bets (=calibration) is not a money making strategy in itself. To cash in you have to find someone else offering the wrong odds!

  2. In 1 “The Syrian government will control a larger proportion of territory in a year’s time relative to today” you are formally right. The prediction made before the start of the offensive the forces of Assad against ISIS. After this offensive, Assad has lost part of the reconquered territories , but of the part reconquered territories kept under control. On other fronts the territory under the control of Damascus has increased too.

  3. Thanks, but I looked at the maps of Syria in late December 2015/early January and today, and whereas the SAA were at the very gates of Palmyra back then they have now been driven 50km away, together with huge chunks of the adjoining desert, and have been almost encircled at Tiyas airbase.

    I did foolishly define success there in terms of territory, so technically speaking I lost, even though as I mentioned the capture of Aleppo is strategically far more important than the reversal in Palmyra. However, I can’t exactly go back and redefine things in my favor.

  4. Philip Owen says

    Clearly not yet suited as a writer for a British Redtop. Repeats of the same failed predictions is what the readers seem to want.

  5. Gains in Latakia and the capture of Aleppo, but ironically, pushed back further in Palmyra than at the same period last year.

    Palmyra will likely be back under SAA control in 2017. The SAA has already halted the ISIS advance and retaken some of the lost territory around the Tiyas airbase, and the major counter-offensive hasn’t started yet.

    Reportedly, when ISIS retook Palmyra with a force of 4000-5000 jihadis, it was barely defended, by only about 1000 pro-Assad fighters, almost all of whom were poorly trained and equipped NDF militia (Putin almost certainly could have prevented this embarrassment, by deploying a mere 3000-4000 Russian troops to help defend the place while the SAA was occupied in Aleppo).

    Demographics – TFR will increase: 50%. This is currently very hard to assess. The number of births fell by 1.7%, but its well known that the number of women in their childbearing years is still falling, so overall the two effects will have almost perfectly canceled out. Therefore I am not yet in a position to rate this prediction. The total TFR for 2016 will certainly be in the TFR = 1.75-1.8 range, just as in the past two years.

    TFR was 1.78 in 2015, so if the ceiling is for 2016 is only 1.8, it will likely decline. Which would be the first time that has happened since 2005. Although perhaps some decrease in fertility should be expected after two years of recession.

  6. I may not be the God-Emperor’s psyker like the brilliant Scott Adams but being wrong was never sweeter.

    You were more accurate than Scott Adams.

    People are forgetting the exact prediction he made. It wasn’t “Trump will win the presidency” it was “Trump wins 65% of the votes in the general election.” Read it and weep.

    http://blog.dilbert.com/post/131552504961/trumps-third-act-part-of-the-trump-persuasion

    He got the general direction of the election right. But so would someone predicting a Trump win by a thousand electoral college votes. He still ended up incredibly far from the truth.

  7. E. Harding says

    Apparently, I got Brexit and HRC more correct than you, Scott Alexander, and the markets, placing the probability of both at 50%:

    https://againstjebelallawz.wordpress.com/2016/01/01/110-predictions-for-december-31-2016/

    I also got the Yemen more correct than you. On November 8, I predicted Trump getting 264 EC votes (including NH) on the day of the vote, but he ended up winning my home state of Michigan instead, which I did not predict would occur.

  8. E. Harding says

    Also, I (like many other worthwhile men) was banned from the Slate Star Codex comments for supporting Trump a bit too fervently before the election, but it was all worth it in the end (he won the state Scott A. and I live in).

  9. Congrats.

    I posted about that on SA’s calibration thread.