For the last three years, I've made predictions about how the coming year would go. In 2019, I made sixty (60) predictions, and now it's time to see how well I did.
Predictions left unmarked, while those struck through were incorrect. I am not throwing any out this year, which would be marked in italics. There are a few borderline cases, such as Boeing's Orbital Flight Test, which did not dock at the International Space Station due to a software problem, but otherwise appears to have been a successful spacecraft demonstration. I am also counting my time playing around with Python as learning a new programming language, because I had basically forgotten everything I learned about it from the last time I used it.
I will read 15 books this year: 95%
....20 books: 80%
I will read three technical/science/engineering books this year: 80%
....five books: 60%
I will write an average of one blog post per month: 90%
I will still be using Facebook, Twitter, and Tumblr: 80%
I will remain single this year: 60%
I will not be hospitalized this year: 95%
I will compete in NaNoWriMo: 80%
...and succeed: 70%
I will not begin learning Russian this year: 70%
I will begin learning a new programming language this year: 70%
I will either find an engineering job or enroll in grad school: 90%
I will still be participating in the Less Wrong/SSC community: 90%
No successful human clones announced: 80%
KIC 8462852 will not be satisfactorily explained: 80%
No strong evidence of extra-terrestrial life found: 95%
The Mars InSight lander will still be operational: 90%
Both Voyager spacecraft will still be operational: 80%
Hubble Space Telescope will still be operational: 90%
JWST scheduled to launch on or before December 31, 2021: 70%
SpaceX conducts Demonstration Mission 1 successfully: 90%
Boeing conducts the Orbital Test Flight successfully: 90%
Falcon Heavy will fly at least once this year: 80%
SLS scheduled to launch on or before December 31, 2020: 70%
Virgin Galactic tourist flights will not begin this year: 60%
No Chinese crewed space flights this year: 80%
No deaths in space this year: 90%
No country leaves the ISS Treaty: 90%
Sci-Hub will remain easily accessible from the United States: 80%
POLITICS/WORLD
Donald Trump will still be President of the United States: 95%
Mike Pence will still be Vice President of the United States: 95%
Donald Trump will still be planning to run for re-election: 90%
More than 5 major Democrats (current or former Governor, Senator, or Representative) announce Presidential bids: 90%
....more than 10: 70%
At least one major Republican announces a Presidential primary challenge: 70%
United States will not go to war with a nuclear power: 90%
United States will not enter a major new war (>100 US casualties): 70%
No additional countries opt to leave the European Union: 90%
North Korean government will not be overthrown or displaced: 95%
North Korean government will not begin major liberal reforms: 95%
US Nominal GDP will finish ahead of the EU and China: 80%
Bitcoin will end the year valued $500 or above: 80%
....$1000 or above: 60%
META
I will remember to score these predictions: 95%
I will still be overconfident in these predictions: 80%
I will be less overconfident than in 2018: 60%
The last two predictions required me to do some analysis before scoring, and serve as my tie-breaker questions. This year, I almost had to invoke them for that purpose.
60% predictions: 5 right and 3 wrong, for a score of 63%
70% predictions: 8 right and 6 wrong, for a score of 57%
80% predictions: 13 right and 3 wrong, for a score of 81%
90% predictions: 13 right and 1 wrong, for a score of 93%
95% predictions: 8 right and 0 wrong, for a score of 100%
Graphed, my results look like this:
This is much better calibration than either 2017 or 2018. Deviation from the orange line indicates the degree of under- or over-confidence. I had some small under-confidence at the very high ends, and some over-confidence at that 70% level, but overall my result hug the perfect calibration curve much more closely than the previous years I've done this exercise. I'm still counting this as over-confident overall, but just barely.
I did about evenly-well in each major category. Personal/Professional and Politics/World were tied at 74% accuracy, while Science/Space was 84% accurate. My overall accuracy was 78% i.e., I got 47 out of 60 questions correct.
Unlike previous years, there aren't any major, obvious failure areas. I still had a degree of launch schedule optimism, which contributed towards the over-confidence at the 70% mark, but not to nearly the same degree as in past years. I also missed both questions about UK politics. Finally, the other categorical error was expected fewer cases of polio this year than actually occurred. This could be a case of anchoring: there were very few cases in the years before I began making calibrated predictions, but polio eradication effort has run into some serious stumbling blocks the last couple of years, with 2019 being particularly bad.
My overall take-away is that my calibration has improved, with the obvious caveat at 70% confidence. Besides adjusting for this, there does not seem to be a major, obvious lesson for my 2020 predictions besides more of the same.
