I've been playing a lot more as of late so I've been able to see kind of how the rounds play out. I think it's just variability. Some rounds are harder than other rounds and there just isn't any guarantee that you will know some answer to some question on each round. Also, I think in the past I've felt that if I got two in a row where I scored 8s/9s that this was some evidence but I actually now tend to think it's just clustering. The chance on any individual hour that you don't get 10 correct is the same every hour, and occasionally it's just going to happen that they come two hours in a row, or more even. So I think at this time I take back what I said about the Easy game and think that whatever has been done to it, a balance has been struck. I still would like to get 10s a little more often but we can't have everything we want in life. Just this hour I made a 983 so such scores are still more than possible, just not as frequent as they used to be.

Oh, I'm dumb! That was from the Hard quiz! I knew that one was way too hard!

Categories are completely random. The only thing the algorithm ensures is:

- It's not the same topic as currently running in Whos the Smartest
- It's not the same topic as last hour

I'd like to reiterate the concern that somebody brought up previously about question length.

I think that, in a game where speed is a factor for determining who wins (which is practically everything except quizzes), there needs to be a character/word limit. There are plenty of questions that are definitely easy, but so diluted with unnecessary words that when a person is randomly assigned one or several of those questions, they're at a disadvantage.

I don't recall if I crabbed about this in another forum, but I recall having some questions that took up 1/3 of my screen (which I have set to 1920x1080).

Since quizzes are not written just to be fodder for the timed games,they may well contain questions that are longer than desirable for speed - but which contribute to the enjoyment of players taking the quiz in its entirety. It would be a shame to not allow them - especially for those of use who care more about the quizzes than the games!

There is a mechanism in place that allows specific questions to be removed from timed games if they are too long, but it can only be used if players who encounter such questions submit a correction note saying that the question is fine but too long for timed games. If the editor who handles the correction agrees, it can be removed. (Of course, they may feel that it is long but not too long - there is some judgment of that grey area.) If it's taking up 1/3 of a screen, it's too long for a timed game, however valid it may be in its original context. And, of course, the single questions, which do have a character limit imposed on both the question and the answers, are being used in games more and more.

Glad to hear there's a character limit on the new questions/answers. I had a gut sense that some of those were still what I'd consider too long, but I'm honestly not sure. And as you pointed out, what feels long is subjective at this point.

From a programming perspective, I'm not sure how some questions wind up being selected from the quizzes to be eligible for the timed games, but I'd think that at that point and on the existing question pool itself the code or a script could be run to identify questions that exceed a character length.

And to emphasize your point about quizzes not being written for timed game fodder, I absolutely agree that quizzes shouldn't be required to have a character limit per question. Some longer questions can be funny, include extra interesting tidbits within them, or build up to an answer, to cite a few valid/good things. But, they may perhaps not make the cut for a timed game question when the magic behind the scenes decides what goes into the timed game pool.

I think the 'ideal' scenario is we measure recall time and normalize the read time. That it takes someone longer to read a question doesn't have any actual bearing on their knowledge or speed in selecting the answer. It just has to do with how fast a person can decode text, and I don't think that is the primary aim of what we want to do.

Of course that can't be done but what we can do is ensure the questions are direct and short as possible so that it has roughly the same effect as if everyone was only allowed to answer within a set time frame. The closer the question gets to a single letter or word the more that part of the time can be normalized.

I agree to Juggernaut's point that there is a kind of virtue in long questions, but I don't think that such questions in timed games, and especially sentences with additional clauses, advanced use of punctuation and wordiness which actually can obscure what the question is asking or the information it is wanting, is virtuous in the timed game. Ideally there would be a superselective set of questions following close parameters that get selected for timed games, or the reverse which would be to exclude from inclusion those questions which aren't.

In short, I don't think exclusion on these grounds is not virtuous, and I base this on the fact that I myself have selected the wrong answer just because the length of the question, determined by the number of words and clauses between the who/what/when/where/how/why and the ? makes it unclear what is even being asked, which in time games means slowing down considerably to take in the question or going with what you think is being said. This doesn't 'test' actual knowledge, it tests reading comprehension.

My understanding has been that these questions are drawn from the question pool of questions submitted by users. They are not drawn from quizzes.

There is an interesting warning on the page to submit questions which says:

"Please try to ask questions in a creative manner. A question like "Who wrote 'Satanic Verses'?" is boring. A far better version is "In 1989, Ayatollah Khomeini issued a call for all Muslims to kill the man who authored the novel 'The Satanic Verses'. Who was this author?". As you can see, this second version is far more interesting. We certainly don't want the question too long, but we do want it interesting. It's a tricky balance!"

If this is true, it is by design that the questions have gradually over the years become more unwieldy and rather long, the system requires them to do so. But I agree, Who wrote Satanic Verses, is quite boring. They could have at least put the article in there to read Who wrote THE Satanic Verses. The article alone makes the question seem more appealing.