Some of the answers to these questions are highly subjective, leading them to not be great examples of how wrong someone is.

For instance, the % of budget spent on defence by the US question only includes (I assume) the DoD discretionary budget and the Overseas Contingency Budget. It leaves out the VA, the State Dept. Homeland Security, the Justice Dept. , the National Intelligence Program, and the NSF, all of which have defence components.

So in that case it's not so much that someone is wrong if they pick the incorrect answer, it's just that the definition of defence spending is nuanced.

My only issue with this is it needs more questions. Besides that love the idea/application/implementation.

The creators of the site have a particular political axe to grind, and because they grind it in every question, it subverts its own purpose. Rather than trying to simply ask, "What do I think this number is because of how much I know about the world I live in" it's "What do I think this number is, conditioned on the fact that this particular person included it as a question on the site and chose these particular buckets." Rather than making you better at estimating things, it will make you worse because the questions are systematically exaggerated for difficulty.

Because the goal of each question is to get people to select the wrong answer, the buckets are arbitrarily sized to make guessing difficult, the sources are chosen to be those that have a particular bias, and the questions are worded in ways to make the question misleading on first read.

The main reason this causes me to cast mild aspersions on it is because it's such a missed opportunity. I post this as someone sympathetic to the authors' views, but who is more sympathetic to accurate discourse. People don't know what they don't know, and it's in everyone's best interest to become well-calibrated to their uncertainty. Instead the site will just be ignored by people with opposing political views, and used as back-patting fodder for people who are in support of these political views. It would be much better if it asked a wider variety of questions, scored you based on how close you were rather than simply right or wrong, had ways of signalling and scoring your uncertainty for each question, and had reasonable buckets for each question. I understand that this may not serve the ends of the creators as well, however.

Some examples:

What was the average American college student's college-related debt in 2013? (The actual source is for college GRADUATES, not college students)

How old is the known universe, in years? (One answer is 7000 years, the other is 14 billion years)

What percentage of animal species are headed to extinction in the next century? (This is sourced to Democracy Now! and the exact quote of the source is "Scientists say we are now experiencing the sixth extinction, with up to 50 percent of all living species in danger of disappearing by the end of the century" which is a lot more hedging than the question implies. >10% of the questions on the site are related to percentages of animal extinction [Like what I did there?].)

What percent of U.S. prison inmates are functionally illiterate? (The definition of functionally illiterate is very counterintuitive to the average person, so this question doesn't mean what you think it means)

What percentage of inmates in U.S. federal prisons have mental health problems? (This refers to people who have reported symptoms of mental health problems, not those who were diagnosed as having mental health problems)

These are just the first random questions I clicked on, and by no means exhaustive.