HQ Trivia App Puts On-Demand Generation on a Strict Schedule | Hacker News Reader

HQ Trivia App Puts On-Demand Generation on a Strict Schedule

wsj.com

42 pointsmudil28 comments

I play this game semi-often and as the popularity has risen, the performance of the game has gone down hill.

Interestingly enough almost all the issues occur when starting the game - after the first question the game runs (somewhat) smoothly, even though there may be 400k+ viewers.

I wonder what technical hurdles their hitting that causes the app to lag at the start (an issue the commonly delays the game by 30-60 minutes).

Probably because most people leave as soon as they lose?

I thought so too, but from playing the game there could be as many as 300k+ participants by the 3rd question, and on Sunday games that number could be as high as 500k, nearing their all time high, and tens of thousands still watch all the way (for some reason). The issues, as I have noticed them, have always been around getting started. Today I noticed an issue where my sister was playing HQ, but didn't get any questions and everyone in her chat was complaining about the game being broken but not in mine. I then noticed her viewer count said 22k while mine said 300k - somehow she was in the wrong session (so maybe there's a CDN issue)?

or (gasp!) the viewer count is a complete bullshit number

I'm curious, would this be legal?

sure if they disclosed it and said, live view count is just an estimation or for entertainment purposes only.

The host often mentions the number of viewers, so it'd be weird for them to be having multiple different numbers. The numbers might not be real, but this would imply they should at least be consistent.

Why lie about the viewer count though at this stage? Thats an incredibly cynical view. In good fun, if someone asked how I would tackle HQ's scale problem my first response wouldn't be "the viewer count is bullshit."

Besides you can also see the amount of people who responded to each question up until the end. If you really wanted to catch the act you could go through the process of verifying the last couple winners yourself.

In any case thats besides point, weather the view count is 100k or 600k, starting the game has problems. Once the game has started, the participant count, weather 300k or 50k is smooth. I think chis is right, their question/answer logic has trouble delivering the questions to all their users as you can even notice the host has trouble loading the questions so he can read them out.

My theory is that they lag when logging people in and sending the questions over. Seems like streaming video is a solved problem, but they’re struggling to scale their custom logic governing the trivia.

They've only had a few weeks to scale up from a few thousand to a few hundred thousand concurrent websocket connections. It's weird that it works great sometimes, and other times it doesn't.

I'd love to read their engineering blog!

In my experience it's the video that's the problem.

For me, the questions and answers are always served but the video rarely shows. Which honestly works for me because I find the host pretty annoying.

Streaming video in the conventional case may be a solved problem, but HQ has the additional (and seemingly unique to me) challenge of needing the streams to be nearly perfectly in sync. It's less like a bunch of people streaming HBO and more like half a million folks viewing a live webinar or something.

I can imagine WebEx, a far more established and mature product, might cough a bit too.

Maybe just an authentication-based issue. Things like Bcryot are designed to be resource-intensive on purpose, with default difficulties usually causing hash rates of 10-20 per second per core.

I hope it inspires competitors, because it's a fun concept -- gathering together with a bunch of people at a certain time to play a game together. But the HQ Trivia game itself is... terrible.

I'm aware they're beholden to certain design constraints. Such as making something that's immediately understandable and fun to a wide range of people. But. They have not risen to the challenge, in my opinion.

I expect there will be numerous competitors in this space. And some will have much better game design chops. And one of those will do very well.

(And then the whole apps-like-TV-game-shows-everyone-plays-together market will totally collapse after four hundred start-ups launch and obliterate the novelty.)

Can you be more specific in what how they’re not rising to the challenge?

I think the live stream quality and synchronization is incredible. All my friends gather at lunch and the video and sound play together exactly.

The game design. I meant user experience design constraints.

For example:

- Most people will lose after fewer than half the questions have been asked and have little reason to stick around.

- The chance of having a meaningful "win" is so low as to be nearly pointless. I won a round that, like, 100-ish people out of 50,000 won. That's too low. Games can also have multiple flavors of winning.

- Money is not a useful incentiviser. In fact, it can ruin the experience. Again, I won a round (along with about 100-ish people). Won about $15. Which after all of the hooting about $5k or $10k in prizes felt pathetic. (Also I can't withdraw until I get $20, which felt sort of insulting or borderline scammy and tainted the experience.) Good games create their value without extrinsic prizes -- or, at least, the extrinsic prizes have to be valuable enough to add significant drama to the game.

- Straight multiple-choice trivia is the unflavored ice cream of game design.

I could probably come up with a few more things, but it's late and I've got this third glass of half egg nog/half whiskey blend that I need to finish before bed. (Not a Christmas thing, just my year-round nighttime ritual.)

I largely disagree with you.

It'd be nice if you could still win 'points' or have some other reason to stick around. But I still do feel fulfilled even dropping out early. I think it makes getting further feel even more special, because you're simply in the game longer than normal.

Money is a great incentivizer to get people excited / interested. It's already fun but the _potential_ of actual cash makes it feel more exciting. It's also a better word-of-mouth hook.

The simplicity of multiple choice makes it easier to play with multiple people and still select an answer. It gives framing (you know one is correct), but 3 choices is still enough drama after it starts to get hard.

I'm not saying there's no possible improvement, but I feel like they've really hit the nail on the head with this one, and their growth curve seems to agree.

> and their growth curve seems to agree

It's natural but probably wrong to assume that a product's success means it's getting nearly everything right. I agree with the original point: they've hit upon a really exciting concept in getting everyone around the world together at certain times to play a game, but the specific game is not very good.

I think you could even plug in traditional TV game show formats (something more like Jeopardy, for example) and do quite well. That would keep much more of the initial audience around for the whole game, and watching the inevitable forthcoming ads that will be required for the company to make any money from this growth.

How many people got threatened to be fired for this article?

I played for about two weeks starting early November then got bored and don’t play anymore. I wonder if this will be the case with most other players over time as well. Twice daily seems too much. Reminds me of when Who Wants To Be A Millionaire went from once a week to every night in the early 2000s. In that case viewership tanked and never recovered.

If you're hitting the paywall, you can get to the article by using the link from the author's twitter.

https://twitter.com/johnjurg/status/943894139657179142

That didn't work for me. This did, though: http://archive.is/Xb2k8

Coming to you live from the greatest city on earth, the city that never sleeps, Annandale Virginia.

Can't get past the paywall. Can we ban wsj?

Buy a subscription and the support the journalists who write these articles.

Sure, once I can get a subscription that covers what is actually linked here and not just one article.

Anyway I thought hn had a no paywall rule.

IMO that would be a bad overreaction. Journalism needs to fund itself and I'd much rather they do so by subscription than with more and more shitty ads. We shouldn't shut them out from the internet at large.

Reply on news.ycombinator.com