Jan 02, 2016
ssc
Read on (unread)

2015 Predictions: Calibration Results

Scott Alexander evaluates the accuracy of his 2015 predictions, finding overall good calibration and considering it a successful year. Longer summary
Scott Alexander reviews his predictions for 2015, assessing their accuracy. He lists 35 predictions across world events and personal life, marking successful ones and crossing out failed ones. He then scores them based on confidence levels, presenting the results in a graph. Overall, Scott considers it a successful year for his predictions, with good calibration except at the 50% confidence level. He also comments on Scott Adams' reported prediction success for 2015, suggesting ways to verify the authenticity of such claims and expressing interest in seeing Adams make concrete predictions for 2016. Shorter summary

Many people are making predictions for the new year around now. I see fewer people going back and grading the accuracy of their predictions about last year, even though that obviously has a lot of relevance for how seriously we should take this year’s predictions.

In keeping with the SSC tradition, here are the results for my (late) predictions for 2015. Successful predictions are normal-looking, failed predictions are crossed out:

World Events
1. US will not get involved in any new major war with death toll of > 100 US soldiers: 70%
2. North Korea’s government will survive the year without large civil war/revolt: 95%
3. Greece will not announce it’s leaving the Euro: 60%
3. Neither Russia nor Qatar will lose their World Cups: 80%
4. Ebola will kill fewer people in second half of 2015 than the in first half: 95%
5. No terrorist attack in the USA will kill > 100 people: 90%
6. Assad will remain President of Syria: 70%
7. Israel will not get in a large-scale war (ie >100 Israeli deaths) with any Arab state: 90%
8. Syria’s civil war will not end this year: 80%
9. ISIS will control less territory than it does right now: 70%
10. ISIS will continue to exist: 80%
11. Iran will reach a deal with the West on nuclear weapons: 80%
12. No major civil war in Middle Eastern country not currently experiencing a major civil war: 90%
13. Iraq’s situation not to get any worse (eg gov’t collapse, new rebellion): 60%
14. Obamacare will survive the year mostly intact: 60%
15. Hillary Clinton will be the top-polling Democratic Presidential candidate: 95%
16. Jeb Bush will be the top-polling Republican candidate: 50%
17. Trans-Pacific Partnership to pass at least mostly intact: 60%
18. US official unemployment rate will be less than 7% in Dec 2015: 95%
19. Bitcoin will end the year higher than $200: 95%
20. Oil will end the year greater than $60 a barrel: 50%

Personal Life
21. SSC will remain active: 95%
22. SSC will get fewer hits in the second half of 2015 than the first half: 60%
23. At least one SSC post in the second half of 2015 will get > 100,000 hits: 70%
24. Shireroth will remain active: 90%
25. I will remain at my same job through the end of 2015: 95%
26. There will be no further ramifications or lawsuits from either side over the flooding of my house: 80%
27. I will reach my savings target: 90%
28. I will get a score at >95th percentile for my year on PRITE: 50% (unknown, haven’t gotten score back)
29. I will be involved in at least one published/accepted-to-publish research paper by the end of 2015: 60%
30. I will not break up with any of my current girlfriends: 80%
31. I will not get any new girlfriends: 50%
32. I will not finish [project]: 60%
33. I will attend NYC Solstice ritual: 80%
34. I will flake out of my plan to lead some kind of Solstice Ritual myself: 60%
35. I will be living in the house I’m currently trying to arrange to rent: 70%

Scoring
Of items I marked as 50% confident, 0 were right and 3 were wrong
Of items I marked as 60% confident, 4 were right and 4 were wrong
Of items I marked as 70% confident, 4 were right and 1 was wrong
Of items I marked as 80% confident, 5 were right and 1 was wrong
Of items I marked as 90% confident, 4 were right and 0 were wrong
Of items I marked as 95% confident, 7 were right and 0 were wrong

As usual, the dotted line represents perfect calibration; the closer my blue line comes to that, the better I’m doing.

The graph looks like there’s a massive failure at 50%, but this is just an artifact of very few questions at that level. If I’d gotten just one more right, I would be at 33%, ie as close to 50% as it’s possible to get with a set of three. Given that the difference between total success and total failure was just one question, I don’t feel too bad about total failure. Everything else looks pretty good. I’m prepared to call this another successful year.

A side note: Scott Adams has also graded his predictions from the past year, and reports incredible success: 9/9 correct despite going way out on a limb and saying things everyone else found really unlikely (like that Trump would stay Republican front-runner). The obvious way to accomplish like that is to make lots of things that vaguely sound like predictions, then only highlight and count the ones that end up correct; after a quick scan of Adams’ blog, there’s no sign that he’s doing this; his win seems pretty genuine. Another method might be to make vague predictions and grade them in your favor, and there is some sign of this – for example, someone going off Clinton’s poll numbers versus Sanders could say they are in fact still going up. Nevertheless, it’s obviously been a good year for Adams, and I’d be fascinated to see him make a list of official concrete predictions for 2016, all in one place, maybe even associated with confidence levels. [EDIT: Someone else’s more pessimistic analysis; this makes it more pressing that he do everything beforehand this time]

If you enjoy this fan website, you can support us over here. Thanks a lot!
Loading...
Enjoying this website? You can donate to support it! You can also check out my Book Translator tool.