is forecasting actually effective/altruistic or am I just getting nerd-sniped?
I thought after getting good at forecasting it would end up being more useful...
In 2020, when everyone was making sourdough, I got really into forecasting.
I spent dozens of hours forecasting the answers to question like, “How many viewers will watch the finale of the Great British Bake Off?” I researched statistics and took into account unknown unknowns. I even won a prize!
But I don’t think very many of the questions I’ve forecast the answers to have actually been decision-relevant, for me or anyone else.
Most of the time, if I’m using a forecasting platform, it’s just for fun. Even if I’m forecasting geopolitics or something else important, I’m just doing it to show off - I’m not using the information for anything.
But sometimes other people do. This comment on Metaculus has always stuck with me:
In my personal life, there’s one forecast I can think of that’s been useful to me: I forecast what would happen to house prices before I bought my house. I predicted they would fall after we bought our house but was pretty confident they’d be back above the price we bought at within 5 years (I was right about the first part; let’s hope I’m right about that second part!).
I also forecast my odds of getting a promotion. The odds for each application were depressingly low, but if I applied for ten promotions, the odds were still low but not negligible, I thought. I did not apply for ten promotions, and I did not get promoted, so I guess my forecast was… correct?
But most of the time my forecasts don’t feel this relevant to the real world or my own life. They mostly feel like looking for my keys under the streetlight. I usually forecast questions where I have historical statistics available, which means I get consistently good scores, but the answer just doesn’t end up being very useful. My friend Denise Melchin, who forecasts questions about her personal life like whether she’ll change jobs or have another child over the next year, says forecasting helps her embrace living in a world of uncertainty. I guess that’s true, but I’m not sure that’s enough value for me to regularly actually forecast things!
Do you use forecasts to help you make decisions? I’d love to hear about it
Yes! Similar to Denise I forecast a load of personal life things. Most of the value has come from handling anxiety. I think it's often more about the process than the actual number though. Being very explicit about creating coherent narratives around my fears allows me to do some combo of accepting of the bad outcome and challenging of the bad narrative.
(I wrote about this here:
https://seekingtobejolly.substack.com/p/i-am-using-forecasting-to-worry-and)
The connection between EA and forecasting is that if you're doing "long-termism" you have to be able to predict the future.
I'm sceptical anyone can actually do this to any reasonable degree. With short-termism you can make use of feedback loops - try something, if it works, continue doing it, if it doesn't, change track.
With long-termism you can head down the wrong road for your entire lifespan - many people's lifespans - and never know it until it's too late. If forecasting accurately is possible in the short-term, it opens up at least the possibility of doing it long-term.
Sadly psychophistory is not real and I think it's probably fundamentally impossible to make projections that far into the future.