I think the (very early) community median is surprisingly optimistic. Although when I've felt this in the past (for the US), I was wrong. However, at the moment there's apparently a 12% chance of new cases dropping below 50k *this month*. Granted, only 11 predictions have been made. I've assigned 10% chance of it dropping below 50k cases before July 2020. The last time we were at a three day rolling average of 50,000 was March 27th (46 days ago). We've sustained 80,000 new confirmed cases a day since the beginning of April, and it seems very unlikely ...
@(alexrjl) oh that's awesome, I somehow missed they were using the same questions! Thanks for pointing that out :) I may be wrong, but it seems like it uses the final forecast to compare scores. This seems to make a particularly big difference to continuous questions, where distributions can be narrowed over time if you update before the question's close. (For me, my Brier score is 0.075 while Metaculus' score is 0.100, but then my log score is smaller than Metaculus' for continuous questions — I don't think my relative judgement quality is significantl...

@alexrjl oh lol, thanks — being a bit slow!

I expanded my range to put further weight on the US' share being less than it is currently (28%). I know that Frontier is set to change things, but I don't think the possibility of delays are properly accounted for in the community's forecasts. It takes a long time for such things to come online, let alone be built.

Beyond the exascale projects, does anyone have any links that suggest the US' relative investment is expected to increase from previous years?

I'm not placing too much weight on experts' forecasts on how much cheaper it will get; experts were consistently too pessimistic about how the cost of renewables would decrease over time — economies of scale and experience curves are not particularly intuitive. That being said, I doubt the $/MT CO\(_2\) will change as quickly as the $/kWh for solar power; solar photovoltaics were easy to iterate the design of to make incremental improvements, and I'd imagine provides an upper bound for the rate of progress on this technology's figure of merit. The pric...

@Sergio Thanks, that makes a lot more sense — skim reading meant I attributed it to the un-augmented model

I'm fairly new to using Metaculus (despite creating an account a while ago). The thing I'd most like is a prediction slider that works on mobile! I realise that's an irritating request (and I'm fine without fully mobile-friendly website) but would like to be able to make or update predictions on the go.

Also @AABoyles' point on Dec 27, 2019 about discrete/category-based questions is something that I have noticed as well.

While I think Trump voters are more likely to die from Covid-19, I don't think Trump-voting states will show up at the top. Firstly, the 12 states with the highest proportion of people aged 65+ comprise of: - 6 Clinton-voting states (Maine, Vermont, Delaware, Hawaii, New Hampshire, New Mexico) - 6 Trump-voting states (Florida, West Virginia, Montana, Arizona, South Carolina, Pennsylvania) - In order (oldest first): Maine, Florida, West Virginia, Vermont, Delaware, Hawaii, Montana, New Hampshire, New Mexico, Arizona, South Carolina, Pennsylv...

I expect there'll be an error rate of around 5% to 6%; there's going to be diminishing returns for a bunch of these metrics, so I expect big jumps from a 7% error to be unlikely (even if possible). I also imagine the practical maximum scores are <100% for some of these metrics (but I haven't looked into it at all).

— edited by qassiov

With exaflop computers coming soon, I expect an increase of a couple of orders of magnitude, as that's the sort of change we've seen historically over a decade. It's not quite like Moore's law, as rates have slowed over time, but perhaps more incentives for building supercomputers will reappear.

I've guessing we'll see 30x-200x growth from the first exaflop computers - exciting!

My median guess assumes an improvement of 3 percentage points, with 1.5pp and 5pp falling in my 75% bounds. This is just based on the observation that this value has had annual improvements ranging from 3.5pp to 6.5pp in recent years, with growth slowing in recent years (presumably due to diminishing returns for some of the benchmarks).

However, I've got a reasonably long tail on this, since I think a >6pp improvement (~140 here) is more likely than a <0.5pp improvement (~118).

@Linch @ninjaedit I thought similarly about the data updates, also had 1.02M for my 95th percentile. I was more optimistic for my 50th percentile range though and put 900k-926k-964k.

@admins I calculate this to be 117.9, using the data from PapersWithCode. Weirdly though, the Stanford Cars SOTA decreased on PapersWithCode (as the original SOTA got replaced with a pretrained version of the model, I think). Also, many of the models in the spreadsheet for CUB200 were missing — I'm guessing this is because PapersWithCode might not be the most reliable source?