Sunday 11th April: Knightmare Challenge?

[Mystara]

Many thanks to everyone who took part. We had four teams, which was great.
For those of you who didn't attend, the concept was loosely based on university challenge, with starters being entirely Knightmare based and worth 0 points. Thus, without some general knowledge and some knightmare knowledge, it was impossible to score any points.

I think that three problems were highlighted. None were major, but I'd like to try and resolve them all:

1) General knowledge questions were deemed too difficult.
Approximately 50% of the general knowledge questions were answered correctly. I will include some easier questions in the future.

2) The Knightmare questions were too hard for newbies.
Agreed. But frustratingly, they were too easy for Knightmare-regulars.

3) The process of 'buzzing in' is too problematic with lag. In many cases, as soon as the question was asked, someone had typed a '!' to indicate they knew the answer. Even a half-second lag made it impossible to volunteer an answer.
I agree with this too, which is unfortunate, because I think the fast-pace format made it fun.

The biggest problem is the third one. A few suggestions were made by participants, which I'll list here, along with my comments:

a) Forester suggested that rather than 'buzzing in' to volunteer an answer, each team takes it in turns to be offered a starter question and is given perhaps 5 seconds to answer it, before moving to the next team.

My problem with this approach is that for the Knightmare questions to be easy enough (which we want, in order to encourage non-hardcore fans), every team will always answer their starter question in the 5 seconds provided. In which case, if there are x teams, you have to wait for (x-1) teams to go through the bonus rounds in order for you to score some points. The pace becomes much slower. So, for example, given that there were 4 teams yesterday, you'd only get to answer 1/4 of all questions and you'd have to wait for all 3 teams to answer their questions before you could go again.

So the pace would be slowed.

b) KaM suggested alternating general knowledge and Knightmare questions for the starter and to make them harder.

This eases the need to provide Knightmare questions (which is hard to do). It also makes the quiz more newbie friendly. If the questions are made harder then lag is less of an issue. However, IMHO, it removes the focus of Knightmare.

c) My own suggestion was to change the format of the starter questions to being "nearest guess" questions with a 10 second timer. No penalty for getting it wrong. The bonus questions would always contain one Knightmare question and would be given to whoever got closest to the answer.

This would allow some amount of conferring for the starter questions. It also massively reduces the disadvantage of lag and makes it newbie friendly (without knowing any Knightmare, you CAN still score points).

My concern is that it still slightly reduces the focus of Knightmare. My suggested fix for that is that getting a KM question wrong costs you 5 points. If you're a GK expert, then you still have a slight advantage over someone with only KM knowledge because the starter questions are GK based, but a mixed team will always be best.

Another option is to give a bonus if you get all the bonus questions right, i.e:
1 question right: 5 points
2 questions right: 10 points
3 questions right: 20 points

This gives a slightly bigger advantage to GK experts. However, the difference is mostly perception based. You're no longer being penalised for not knowing Knightmare, which feels a bit bad.

I shall continue to tinker.

The quiz was sufficiently popular that I'm happy to do a full run, with prizes (though I have little money, so don't get too excited). I would hope that none of these proposed changes would deter anyone who showed up yesterday?

It will take me a few (3?) weeks to iron out the format and code up a bot. Watch this space.

[Darren]

I enjoyed the quiz, very big well done! I am pleased there was a good turnout, it also brought the chat to life. I had hours of fun discussion.

[Mystara]

It's embarrassing how long I've taken over this.

However, with that said, the bot script is written. I've also run a few tests (just myself) and it seems to work relatively okay. All it needs is a little polishing and refining. It also needs the front end adding on (i.e. the ability for people to add themselves to teams). However, I'm not so worried about these bits, as they're only a couple of hours work and are relatively straightforward.

I realise it's fairly short notice, but I was wondering if we might be able to do a quick (ish) run through on Sunday in Knightmare chat, just to see whether the script itself actually works? Assuming all is fine, I can refine any time limits and add the front end within a couple of days so that we can start in the new year.

I've even got a DVD of 100,000 questions to ask. Sadly, these need inputting into the database. They also don't cover Knightmare, which means I'm going to have to think up a bunch of questions myself

[Kieran]

Aha, I am glad there is some movement on this.

Has it really been so long since it was last played?

Someone has been stealing time in the dungeon...

[Mystara]

Thanks to those who helped test.
We eliminated a whole lot of bugs. But some changes are still required:

1) I think I'm going to change the command to provide an answer to simply "!", ie. !42
There doesn't seem to be any real value in continually typing "!answer".

2) I need to make sure there is a sufficient gap between the end of one question being asked and the next question being given. Otherwise, multiple !answers are treated as answers to different rounds

(Part of the problem here is that Eggdrop is limited to how quickly it can respond to commands. That's solved in the next version, but I haven't yet been brave enough to upgrade. Perhaps I shall do)

3) The channel should probably be moderated when a team is answering questions

4) I was able to google for answers within the time periods given for each question. I think it will therefore be necessary to revise the time limits down by about 5 seconds for the starter round. Especially with the modification of the command to supply an answer, I think this will make things fairer.

Another test will be required. In particular, it's important to see how the bot reacts to the channel talk being busy. Since the rounds are time sensitive, it has to respond within an adequate period of time.

I shall re-randomise and change a few questions for the next test. But it will be substantially the same question bank. I see no point in wasting good questions on a test. Especially since they take so long to write

I estimate the next test as being 16th January.
In the interests of speed, if people wishing to take part could post their team name (and members) below, it would be much appreciated. This way, I can pester you all when nobody responds

[Mystara]

Hmph.

Well, there wasn't much response. Can we get some testers for this?