From a966e8d6a2a3585b8b3390a560dea89638ad1253 Mon Sep 17 00:00:00 2001 From: Nuno Sempere Date: Tue, 29 Sep 2020 16:59:10 +0200 Subject: [PATCH] Added list of forecasting projects --- ea/Forecasting/ListOfForecastingProjects.md | 26 +++++++++++++++++++++ 1 file changed, 26 insertions(+) create mode 100644 ea/Forecasting/ListOfForecastingProjects.md diff --git a/ea/Forecasting/ListOfForecastingProjects.md b/ea/Forecasting/ListOfForecastingProjects.md new file mode 100644 index 0000000..cf17dc7 --- /dev/null +++ b/ea/Forecasting/ListOfForecastingProjects.md @@ -0,0 +1,26 @@ +# List of forecasting projects: + +- Have forecasters read Toby Ord's book, and generate their own probabilities of x-risks +- Feed forecasted x-risk probabilities, in addition to some other inputs which could also be forecasted, to Trammel's model on patient philantropists, to find out what the optimal spending conditions for a patient philantropist would be. Right now, some of the inputs to Trammel's model are very much up in the air. +- Forecast whether some big companies will keep to their chicken broiler / animal suffering commitments; this would be useful for some orgs to prioritize their outreach. +- Forecast the value of altruistic projects; I (Nuño) have a small demo on this. +- Throw some optimization power at the Metaculus Ragnarok questions +- Look into the relationship between price or time spent and forecast accuracy +- Look into rewarding forecasters according to their Shapley values +- Predict which EA Fund applications will be funded +- Add interesting and nontrivial markets to Augur. +- Rapid Response EA Forecasting Hotline +- Forecast incubator ideas' likelihood of success +- Create a calibration hall of fame/shame for researchers' past forecasts +- Case studies / lessons learned on forecasting +- Pay and support forecasters to make forecasts on existing platforms +- Get a visible forecasting win for EA/QURI/EpiFor +- Participate in an OSINT project +- Suggest questions to Good Judgement Open with an altruistic bent. It is not unusual for GJOpen to accept question suggestions. +- Talk with CSET; they seem a little bit confused/lost. +- Recruit people for CSET; they need more forecasters to be able to produce useful policy feedback +- Forecasting for EA policy advocates +- Estimate the value of the forecasted variables/questions on Metaculus and GJP +- On-demand forecasting to research teams +- Write down lessons learnt from EpidemicForecasting +- Foretold consulting/training