I assume that many of you have seen that Good Judgement is running a September Superforecasting Workshop. Although I would love to attend, the tickets are $1250 which is so far beyond my reach it's not even worth considering.
So what I wanted to put to the community was the idea of running a cheaper workshop or even a conference. There are some really amazing predicters in the community so I honestly think it would be possible to match the quality of the official workshop. I am happy to do the legwork in organizing it if people are interested? Apologies if this is not the place to put this.
On a related note, has anyone been lucky enough to attend the official workshop? If so, what was it like, did you feel it was worth it? Would love to hear your thoughts!
@LuisEnrique Of course this is low priority but others can chime in on other topics to publish on the Discussions page. I personally like the idea of having unverifiable "Mystery Forecasts" such as whether or not there is a great white shark measuring more than 6 meters currently living in Earth's oceans. Or whether or not there is currently a member of Congress who has been given unlawful access to classified documents. Et cetera.
@James-B@LuisEnrique "I would add that I am about to start work on a project to provide statistics and analytics for forecasters such as calibration curves, similarity scores to other forecasters, performance across topics, histograms of initial forecasts to showcase bias"
>> You could also include heatmaps for certain statistical profile metrics. Highly visceral and intuitive visually, in my opinion. Also, this is my ultimate dream for Good Judgement at large but it would be awesome to see local news networks or even big networks like CNN consulting the "Good Judgement Forecast" on their shows like AccuWeather or The Weather Channel provide their forecasts on such networks. For instance, after the weather report and before the traffic report, the outlets display the GJ Forecast on the probability of a question that they have a related news segment on. They could even report the percentage change in forecast like how reporters report the change in stocks day-to-day. And like how they feature social media posts, they could showcase short rationales from forecasters that inform the GJ Forecast graphic. Perhaps even they could employ a simpler model by having the GJ Forecast probability indicator displayed near the relevant headline and have rolling rationales scroll at the bottom or side of the screen.
I got cocky. But is also because there really isn't any cost. Over 80% of my predictions are 100% or 0%. I know that's not optimal game theory.
Second: don't be afraid to change your mind. When facts change, so do your predictions.
I got burned by quite a few questions when I stopped posting for a few months.
Confirmation bias: you'll "hear" people that are with you more. Follow people you disagree with, listen to their arguments.
Identify your biases- which are revealed by past action: have you voted democrat in the last 20 elections- you're probably left.
Did you graduate Ivy League? Elitist?
my final comment/better advice: there is a definite bias among the forecasters here. You can score very well against that bias if you can identify it.
For example there is a strong bias in favor of global warming you can score points sample on hurricane questions by tempering your predictions against the bias
oh - really final advice. If you're sucking, choose someone(s) with a good record and copy their forecasts.
So what I wanted to put to the community was the idea of running a cheaper workshop or even a conference. There are some really amazing predicters in the community so I honestly think it would be possible to match the quality of the official workshop. I am happy to do the legwork in organizing it if people are interested? Apologies if this is not the place to put this.
On a related note, has anyone been lucky enough to attend the official workshop? If so, what was it like, did you feel it was worth it? Would love to hear your thoughts!
@James-B @LuisEnrique "I would add that I am about to start work on a project to provide statistics and analytics for forecasters such as calibration curves, similarity scores to other forecasters, performance across topics, histograms of initial forecasts to showcase bias"
>> You could also include heatmaps for certain statistical profile metrics. Highly visceral and intuitive visually, in my opinion. Also, this is my ultimate dream for Good Judgement at large but it would be awesome to see local news networks or even big networks like CNN consulting the "Good Judgement Forecast" on their shows like AccuWeather or The Weather Channel provide their forecasts on such networks. For instance, after the weather report and before the traffic report, the outlets display the GJ Forecast on the probability of a question that they have a related news segment on. They could even report the percentage change in forecast like how reporters report the change in stocks day-to-day. And like how they feature social media posts, they could showcase short rationales from forecasters that inform the GJ Forecast graphic. Perhaps even they could employ a simpler model by having the GJ Forecast probability indicator displayed near the relevant headline and have rolling rationales scroll at the bottom or side of the screen.
I don't know. Food for thought, I guess.
I had a .05 vs .4? Through 50 questions or so.
I got cocky. But is also because there really isn't any cost. Over 80% of my predictions are 100% or 0%. I know that's not optimal game theory.
Second: don't be afraid to change your mind. When facts change, so do your predictions.
I got burned by quite a few questions when I stopped posting for a few months.
Confirmation bias: you'll "hear" people that are with you more. Follow people you disagree with, listen to their arguments.
Identify your biases- which are revealed by past action: have you voted democrat in the last 20 elections- you're probably left.
Did you graduate Ivy League? Elitist?
my final comment/better advice: there is a definite bias among the forecasters here. You can score very well against that bias if you can identify it.
For example there is a strong bias in favor of global warming you can score points sample on hurricane questions by tempering your predictions against the bias
oh - really final advice. If you're sucking, choose someone(s) with a good record and copy their forecasts.