Back in January, like usual, I made 100 predictions about hold 2023 would unfold. I had no idea at the time just how much my life would be disrupted, and when I thought back about them as the year progressed, I generally assumed that this would turn out to be one of my least-accurate years to date. Now, on New Years Eve, it's time to see whether that was true.
Predictions which I got right are left unmarked, while predictions I got wrong are struck-through. I decided against throwing out very many predictions this year as ambiguous or inconclusive. The only one I through out was the Venusian phosphine prediction, because a new paper by the same author found the same results as before, but I haven't seen that replicated by other researchers. Since I don't feel comfortable resolving it either way, that prediction is marked in italics.
PERSONAL/PROFESSIONAL/ACADEMIC
I will read 12 books this year: 90%...15 books: 80%...20 books: 60%
I will read more fiction than nonfiction this year: 70%
I will write an average of one blog post per month: 90%
I will not experience a major political/religious/philosophical conversion: 90%
I will still be using Facebook, Twitter, and Tumblr: 80%
I will remain single this year: 60%
I will not have my wisdom teeth removed: 80%
I will not be hospitalized this year: 95%
I will not experience a confirmed case of COVID-19 this year: 80%
No one in my close extended family dies of COVID-19: 80%I will not receive any additional booster doses: 70%
I will compete in either NaNoWriMo proper or the Summer Writing Program: 90%...and succeed: 70%I will compete in both: 60%I will either finish my MS or abandon the program entirely: 95%
I will still be working for Barrios Technology: 95%
...on the gCAMMP team: 90%
I will still be living in Texas: 95%
I visit Kansas City at least once this year: 90%
...more than once: 60%
I will still be involved in the meatspace LW/ACX/EA community: 90%I will get involved in the local Mensa group: 70%
I will attend some sort of Solstice Event: 60%
No successful human clones announced: 95%
No artificial wombs for humans demonstrated: 95%
Venusian phosphine continues to not replicate: 95%
No other strong evidence of extra-terrestrial life found: 95%X-59 successfully demonstrated this year: 90%NIF experiments will exceed 4 MJ energy release: 70%
...5 MJ: 60%
No truly self-sustaining fusion reactors demonstrated this year: 90%
GPT-4 released this year: 70%
The popular press will still be using "AI" to describe linear algebra: 95%
No unambiguous artificial general intelligence released: 95%
No additional COVID-19 vaccines approved by the FDA: 70%The FDA does not approve a fifth booster dose: 70%
The FDA does not approve an mRNA vaccine for another disease: 60%
Sci-Hub will remain easily accessible from the United States: 80%
Sci-Hub will not resume adding new papers: 70%
Sci-Hub will remain operational to some extent: 90%
SPACEFLIGHT
Both Voyager spacecraft will still be operational: 90%
Hubble Space Telescope will still be operational: 90%
MSL Curiosity will still be operational: 90%
Perseverance will still be operational: 95%
Ingenuity will still be operational: 70%
International Space Station will still be operational: 95%Axiom Segment still scheduled to begin construction in 2025: 60%
Tiangong station will still be operational: 90%
Orbital Reef, Voyager, and Northrop Grumman space station projects all still in progress: 80%Boeing conducts the Crewed Test Flight: 90%...successfully: 80%
Europa Clipper still scheduled to launch in 2024: 80%
JUICE launches this year: 90%
Psyche launches this year: 90%
Artemis 2 still scheduled to launch on or before 31 December 2025: 80%
Artemis 3 still scheduled to launch on or before 31 December 2026: 80%
Lunar Gateway scheduled to begin construction in 2025: 70%Starship completes the orbital test flight this year: 90%...successfully: 70%
New Glenn does not complete an orbital test flight this year: 60%
Virgin Galactic begins tourist flights this year: 70%
United States will have more successful orbital launches than China this year: 80%
China will have more successful orbital launches than Russia: 90%
No deaths in space this year: 90%
No country leaves the ISS Treaty: 80%
POLITICS/WORLD
Joe Biden will still be President of the United States: 95%
Kamala Harris will still be Vice-President of the United States: 95%
Chuck Schumer will still be Senate Majority Leader: 95%
No form of federal carbon tax will be passed in the US: 95%
Donald Trump will still be running for President: 90%
At least one other major Republican (current or former Governor, Senator, or Representative) announces a Presidential bid: 60%
Biden approval rating above 35% per FiveThirtyEight: 80%...above 40%: 70%
Puerto Rico will not be admitted as a state this year: 90%
United States will not experience an obvious recession: 70%
US YoY Inflation will be lower in 2023 than in 2022 (November-to-November figures): 80%
US Unemployment remains below 5%: 70%
US Prime-Age Employment-to-Population Ratio remains above 78%: 60%
United States does not go to war with a nuclear power: 90%
United States will not enter a major new war (>100 US casualties): 80%
Rishi Sunak will still be Prime Minister of the United Kingdom: 80%
Justin Trudeau will still be Prime Minister of Canada: 90%
Emmanuel Macron will still be President of France: 95%
Vladimir Putin will still be President of Russia: 80%
Volodymyr Zelenskyy will still be President of Ukraine: 90%
No additional countries opt to leave the European Union: 95%
No full-scale negotiated ceasefire in the Russo-Ukrainian War: 70%
Ukraine will remain independent: 80%
Ukraine will not regain its pre-2022 borders: 90%
Ukraine will not regain its pre-2014 borders: 95%
China does not invade Taiwan: 90%
North Korean government will not be overthrown or displaced: 95%
North Korean government will not begin major liberal reforms: 95%
Fewer than 150 cases of wild polio this year: 70%
Fewer than 1000 cases of vaccine-derived polio this year: 80%
META
I will remember to score these predictions: 95%I will be overconfident in these predictions: 80%
I will be less overconfident than in 2022: 60%
How does that add up? Spoiler: surprisingly well!
Breaking it down by confidence level:
60% predictions: 8 right and 4 wrong, for a score of 67%
70% predictions: 11 right and 7 wrong, for a score of 61%
80% predictions: 18 right and 3 wrong, for a score of 86%
90% predictions: 24 right and 3 wrong, for a score of 89%
95% predictions: 20 right and 1 wrong, for a score of 95%
Graphed, the results look like this:
Despite managing to preserve my poor calibration in the 70% band from previous years, I think this is actually the closest I've ever managed to hug that calibration curve. After what an unexpectedly weird year this was for me and my family, I certainly wasn't anticipating that result.
In fact, for perfect calibration with the number of questions I had in each band, I would have needed to get 81 questions right in total for an overall accuracy of 82%. This is, in fact, the number of questions I got right overall! I got too many questions right in the 60% and 80% bands, and not enough right in the 70% band (and the 90% and 95% bands were spot-on). I don't think my errors have ever managed to cancel out that cleanly before.
If we look at categories, Meta was the worst, because I turned out not to be noticeably overconfident and thus one of the three predictions was wrong, for an accuracy of only 67%. That's not enough of a sample size to discuss meaningfully, though, so of the "real" categories my worst was Personal at 72%. This isn't really surprising: dealing with my father's unexpected hospitalizations and illness threw me off track. I didn't end up going above and beyond in reading or writing, and my academic ambitions were basically shelved after mid-March. Hopefully 2024 will go better in that regard.
Breaking out Spaceflight from Science/Technology didn't noticeably make a difference, because both came in at 80%. Those numbers are roughly comparable to previous years. Launch schedule optimism was, however, less of a contributing factor than in the past, so I seem to finally be making some headway there. I suppose my feigned pessimism is finally becoming convincing.
As usual, Politics/World did this best, with 97% accuracy. I think this was mostly achieved by predicting stable things and not making predictions where there was a lot of risk. In this category, I tend to play it safe and end up being quite underconfident. This is not the expectation I had when scoring this category and making the predictions for this year, which is interesting.
Overall, I'm not sure what takeaways to have here. My life took several unexpected turns in 2023, and despite this, the predictions I made in January were some of the most accurate yet. One interpretation is that staying to high-level outcomes is more robust than predicting more specific aspects that are more likely to be falsified. However, such predictions provide less value than more specific predictions (assuming accuracy, of course). Perhaps the most useful interpretation, then, is one of humility, and that we much accept our ignorance if we wish to to overcome it. There is no shortcut to knowledge and enlightenment.
