In another thread, yeotaJMU said:

And I understand where he's coming from. NIT champs, countries that became sovereign, Mr. Olympia--these are all subjects that aren't common knowledge.

So I had an idea that combines the leveling aspect with the trivia aspect. It takes Yeota's multiple choice sheep and kicks it up a notch. It maintains the trivia aspect but lessens the number of incorrect answers.

The questions would be in the form of "name a number between 1 and [trivia question]. People who don't know will just guess 1 or 2. People who do know have the whole gamut of answers to choose from.

Sample questions (I don't know the answers to all of these, so feel free to steal):

Name a number between 1 and:

1) the number NBA scoring titles won by Michael Jordan
2) the number of pontiffs who used the name Pope John
3) the length of the reign of King Edward VIII (in months)
4) the number of U.S. Presidents to serve two full terms.
5) the number of faces on a regulation craps die

These questions don't have to be numbers. For example:

6) Name the first letter of the last name of someone who's been UK Prime Minister since 1945.

The benefit is that people who don't follow Papal history can still guess the number of Pope Johns.

The downfall is that it doesn't reward specific knowledge. E.g., if you know exactly how many Pope Johns there have been, saying that number could be bad, because everyone who knows that number will be tempted to say that number. But that's where the leveling aspect comes into play.

I honestly have no idea if this is a good idea or not. So I'm hoping for some good comments/criticisms/suggestions here.

best bet is just to run one and see how it goes...

trouble is, you don't reward excellent knowledge that scores 'unique'.

I think we have sheep and we have peehs.

If it's the levveling contest you want, I would suggest running a regular sheep, as opposed to a reverse.

OTOH, if you really wanted to go the way you describe, I would offer multiple choice for each answer, but you do not have to take one of the options on offer. So who Dueted with wycleff Jean?

[ ] someone who dueted with wycleffe jean.
[ ] someone else who dueted with wycleffe jean.

But you may choose someone not on this list.

Or you could even go -

[ ] someone who dueted with wycleffe jean.
[ ] someone else who dueted with wycleffe jean.
[ ] someone else dueted with wycleffe jean.
[ ] someone who DID NOT duet with wycleffe jean.

Just make the reverse sheep questions easier- essentially sheep questions with a finite range (omg that came from nowhere and might not mean anything).

Example of a sheep question that wouldn't work for reverse sheep: Name a Candy Bar. (because there are too many obscure candy bars).

Example of a sheep question that would work for reverse sheep that people have done before: Name a planet in our solar system. (limited number of answers, not trivia, becomes leveling game, etc.

name a war crime dude: bad reverse sheep question because who knows more than a couple and there are so many. will result in many incorrect answers, some clustering on a right answer and some 1 pointers for people with KNOWLEDGE. reverse sheep shouldn't really go for knowledge first and foremost imo.

name a current model of toyota vechicle in production: good reverse sheep since people know many answers for this, won't be incorrect answers. leveling comes into play, if there is bunching up then it's due to bad leveling, not due to lack of knowledge holding people back.

Manu, when I ran the first reverse sheep a few months ago, I used those type of questions, people complained that it was just rand(). I mean name a colour of the rainbow for example.

toyota question is better.

But if you want questions to be about levveling, then sheep is the game to play imo.

I actually like the war crimes question for reverse sheep. Rewards in depth knowledge, everyone can probably guess at one or two, unless you are completely ignorant of history in which case you will get a high score.

that, my friend, is "trivia"

it's a different form of trivia

instead of knowing the 1 right answer in trivia (who was the 27th president), you have to know one of the multiple right answers (name a president who was previously a vice president)...and knowing a more obscure one may get you a better score while knowing one that everyone knows may get you a worse score (altho then leveling comes into play and some obvious answers are 1 or 2 pointers)

the main complaint is when <90% of the players know at least 1 correct answer for the question...those are 'trivia' questions...if 99% of the players can give at least 1 correct answer, then it becomes a leveling contest with knowledge helping (don't have to level if you are the only one that knows an answer)

however, sometimes that knowledge comes back to bite you because half the field knew that 'obscure' answer and u score a high score...ideally most people know a few 'obscure' answers for a question and then it's a leveling (ie guessing) exercise to determine which answer to go with

also, the fact that so many questions have an 'obvious' answer scoring 1-2 points means that, in practice, peehs has been a leveling exercise so far

but yes, some questions half the field get incorrect, and those questions are the ones that cause these complaints...the less of those questions we have, prolly the better for getting less complaints

this

really, there's no way the game would have got so popular if all the questions were "pick a number between 1-10" or "name a colour of the rainbow".

the best questions are ones with 10-15 right answers, with no possible ambiguous answers, where everyone knows 2-5 correct answers but virtually no one knows all of them.

Like "a US president who served two full terms" or "a river over 1000 miles long". Sadly it's hard to get questions like this so there are always a few lower quality questions too.

an alternative scoring method would be to punish a wrong answer based on how many people got the question wrong

so for virginia woolf, half the people got it wrong, so they aren't penalized as harshly while all but 1 gets 'name a president who served 2 full terms' so that person gets punished more

i haven't thought much about alternative scoring mechanics, but maybe i will at some point

The only problem is when the questions get too obscure to where 80% of the group takes an incorrect answer. The war crimes question is fine, almost everyone should and does have a correct answer.

yeah, for those questions, u get penalized less because most don't get it right if we altered the scoring mechanics

u wanna reward being the only one to get a specific answer and punish those who get a common answer and punish those who get it wrong, but u want to keep those differences in check...u don't want 1 person to get 1 point while the rest of the players get 39 points because nobody knew the answer, but u also don't want the rest of the players to get 3 points because u are doing the 'most common answer + 2 for that question' scoring...but u also don't want them to arbitrarily get 15 points because another unrelated question happened to have 14 people say the same answer


so u want to make it a scoring mechanism that scores each question differently based on the results for that question...take into consideration the range of answers and how grouped together they are (15/3/2/1/1 vs 10/8/3/1 vs 5/5/4/4/3/1)

i know it is hard to come up with questions.

to me, ideal questions are ones where everyone knows 1-3 answers, some people know 4-6 answers, and few people know more answers than that.

another aspect is the ability to make a good guess if you dont know the answer. for example, i personally dont know an answer for the wycleffe jean question. i have heard that name, but i dont know anything about him. there is virtually no way for me to make a guess on that question.

a better question would be, imo, name a top duet that ranks in the top 20 sales of duets. i certainly have never looked at this list, nor do i even know if a list exists. i also am not well versed on duets. but would everyone be able to make some kind of reasonable guess on this question.

i also think good questions are ones where everyone knows 1 or 2 "safe" answers, but many people know other possible answers. this involves risk and leveling at the same time.

i get what hitch is saying about multiple choice. they would get boring after awhile. the trivia aspect needs to be involved, but not exclusively.

let me say again that i realize making questions is hard, and that i appreciate the mods running these things. it took forever for me just to do the multiple choice so i cant imagine how long it takes you guys to tally this stuff.

my two cents.

i'm gonna come up with a scoring mechanism and then run one to test it imo

I definitely agree that the scoring for an individual question should determine the score that a wrong answer gets for THAT question. I don't think a DIFFERENT question with a super popular answer should jack over somebody for the rest of the game.

The scoring system I am using for "potpourri" has worked out fairly well in my opinion (see thread if you don't know what I am talking about). However, it severely punishes you for missing a question that the field deems easy through a combination of a really popular answer and a bunch of different correct answers. I think the scoring system I used is fine actually but some may not like it.

I really enjoyed doing this reverse sheep (although it took much longer than I thought it would) and plan to do another one soon. For that one, I am planning to score as follows:
Score like the potpourri sheep
Throw out 2 worst scores for each player (assuming 15 or 16 total questions)

Throwing out 2 worst scores for each player will add suspense because you will still have a chance even if you bomb two questions. On the other hand, it will take a bit of skill out of the game. Comments?

eh i wouldnt throw out answers. if you get em wrong you gotta pay

Hey I agree completely. I really don't mind the bad penalties for missing. However I thought there was a sentiment that people were frustrated with questions that they had absolutely no clue on, causing them to totally brick an answer that ensured they would not win or finish highly.

Maybe an alternative would be to put up say 18 questions and say you may pick 16 of these to answer...give people a chance to avoid 1 or 2 questions on a subject they are TOTALLY clueless about.

simpler way to do this and also make sure each question also has the same number of guesses is to let everyone drop their 2 worst scores.

That was my original thought, discussed in a post a couple above this one.

However, I don't think they're totally the same, because when you let people pick, they can avoid questions they have no clue about, but they can still miss questions by going for obscure answers or putting answers that they thought were right.

If you drop worst 2 scores it will always help the score majorly, regardless of the reason, not knowing anything about the subject or stretching too far for an answer.

Also deciding which questions to drop can be a strategic choice in itself.



I will meditate on these things.

I like the idea of taking best (x-2) scores.

I do think 18 questions is generally too many, 10-15 would be better.

Sweet or sour took you forever??

I think the idea of putting, say, 16 questions out, but everyone chooses 14 to answer is an interesting one. It adds a sheep within a sheep factor: because now you're also trying to figure out which questions people will answer.

if there are 30 players but you think 20 will skip over a particular (tough) question, you might want to try it. because with only 10 respondents, chances are any correct score won't be too high.

Yes, exactly what I was thinking. I don't think it's giving too much away to say that half the field probably would have preferred skipping the NIT question, but if you take away half the respondents:

1) Any right answers will get better scores than they otherwise would have.
2) If you get it wrong the punishment won't be as severe.