Predictions
Profiles: Metaculus / Manifold Markets
Tags: akarlin.com / unz.com / @powerfultakes / @ethotaku
Last updated: Apr 15, 2023
Table of Contents
The worth of any given analyst should ideally distill down to how well he or she predicts the state of the world from the information and methods at their disposal. #SkinInTheGame
To this end, I have never shied from taking public bets on things that I think I have some confidence in. This page tallies the results in what I will try to progressively make as fair and impartial manner as possible. However, do note that prior calibration exercises were all done in a piecemeal and inconsistent manner – the closest time I ever got to doing this systemically, in the Scott Alexander fashion, was in the late 2010s, when I temporarily got into a habit of making start of year predictions and calibrating at the end (2016: Predictions/Calibration; 2017: Predictions/Calibration; 2018: Predictions/Calibration).
I only committed to updating this page rigorously in April 2023, including mirroring predictions I make on the blog, Twitter, or Metaculus. So subsequent predictions only becomes truly “calibrated” from that date.
Geopolitics
Jan 11, 2020: No Iranian Color Revolution 💯
Jan 30, 2021: No Belarus Maidan 💯
To add: Brexit; Afghanistan; Karabakh War II; Great Bifurcation thesis
Ukraine War
I am going to tally this up more methodically in a separate post, but for now, I will make the following remarks:
(0) Retrospectively correct in 2018 it would no longer be a “smooth cruise to the Dnieper”.
(1) I was correct on my strong insistence from early January that the war was going to happen.
(2) I drastically overestimated Russian military strength and political competence (as did RUGOV, US intelligence, and virtually all of Metaculus). Naturally, these Tweets became memes in NAFO Twitter and what I’m best known for as a predictor.
(3) I was correct on most all other aspects and specific predictions on the conflict to date, inc. on Kherson withdrawal, Ukrainian manpower problems, etc. See tally.
Politics
Apr 24, 2011: Predict rising prominence of Green, “pirate”, and Far Right ideologies relative to established SocDem/conservative orthodoxies. ✅Seems to have happened: Greta; Wikileaks; Alt Right.
US Elections
2020 US Elections
Nov 2, 2020: Biden wins 279/Trump 259, raw votes: 51.10% vs. Trump 47.07. 💯
Final results: Biden 206/Trump 232, raw votes: 51.3% vs. 46.8%.
Amusing tabulation of my correct predictions (meant to “own” me?) from Russian anti-vaxxer MAGA chud.
To add: Results in Kireev’s competitions. Other US elections. Russian elections.
2024 US Elections
Oct 9, 2021: Predict Biden vs. Trump in 2024, with a Biden victory at The Stark Truth. ❓coming along nicely so far
Tech & AI
Economics
Society
Jan 13, 2018: Poland to legalize gay marriage within 10 years. ❓
Demographics
To add: Third World fertility rates
Troll Predictions
To add: Russia to be blamed for Holocaust
Finance & Crypto
In some ways, trading is the “purest” form of prediction peddling, since you have to very literally put your money where you month is.
Apr 15, 2019/Nov 2021: My main real main accomplishment in this sphere is 10x’ing portfolio by calling the crypto bottom in April 2019 and selling the top in November 2021. 💯
Otherwise, it has been a hit and miss. I got #rekt for 2/3 of portfolio prematurely leveraging up in May 2021. Caught early $SAMO for a 20x (+30% portfolio). Got wiped (-10% portfolio) overstaying my welcome on $OHM ponzis.
Aug 24, 2022: Predicted FTX collapse and jail for SBF. 💯
Alameda ganbled on shitcoins is underwater and will pop like Celsius when BTC goes below $15k.
FTX goes under.
Interpol arrest warrant for SBF and numerous other crypto crooks.
— Anatoly Karlin (@powerfultakes) August 24, 2022
Trading Calls
Nov 24, 2021: Caught the pico-top of the metaverse/#GameFi bubble. 💯 Yes
There is a much more detailed tally from May 2020 in which I scored myself 68/80 on predicting the course of the Corona pandemic.
Jan 22/24, 2020: First raised the alarm, guesstimated IFT at ~1%, and speculated about 20 million global deaths. 💯
Jan 27, 2020: Suggested it was bioerror based in Bayesian reasoning about the BSL-4 lab in Wuhan. Treated as conspiracy theory for the next two years, before officials started coming round c.2022 as Cold War with China heated up. ✅looks likely
Feb 24, 2020: Predicted global spread and “millions” of deaths. 💯
From late Feb/March 2020: Advocated “go hard or go home” policies from the start in relation to lockdowns, shilling for masks when MSM was against them. 💯