Ten Year Anniversary Game: Living Thread

[Macalaure]

Grimoire thread, post #190

Chances of lynching wolves, based on the first 31 games played:
Day 1: 29% (sample size: 34)
Day 2: 31% (36)
Day 3: 43% (40)
Day 4: 35% (34)
Day 5: 34% (29)
Day 6: 29% (17)
Day 7: 55% (11)
Day 8: 0% (3)

All better than the random 25% we have, and it looks even better considering that most villages have a smaller wolf ratio than ours.

Now I'm itching to turn these into confidence intervals. But the sample sizes are small, so the result would probably not be very good. Maybe if I compiled it for all games played, but do I really want to do that?

[Rikae]

Originally Posted by Macalaure

Grimoire thread, post #190

Chances of lynching wolves, based on the first 31 games played:
Day 1: 29% (sample size: 34)
Day 2: 31% (36)
Day 3: 43% (40)
Day 4: 35% (34)
Day 5: 34% (29)
Day 6: 29% (17)
Day 7: 55% (11)
Day 8: 0% (3)

All better than the random 25% we have, and it looks even better considering that most villages have a smaller wolf ratio than ours.

Now I'm itching to turn these into confidence intervals. But the sample sizes are small, so the result would probably not be very good. Maybe if I compiled it for all games played, but do I really want to do that?

Actually I think ours might even be the usual ratio. The average game I remember is something like 16 players, 4 wolves or 12 players, 3 wolves.

Still, slightly better than chance. It is known!

[Formendacil]

Originally Posted by Rikae

Actually I think ours might even be the usual ratio. The average game I remember is something like 16 players, 4 wolves or 12 players, 3 wolves.

Still, slightly better than chance. It is known!

Even so, we can't calculate OUR odds of lynching a wolf off what has actually happened in the past, any more than you can say the odds of flipping a coin Heads or Tails on a given toss is anything other than 50/50--even if the ACTUAL flipping has come up 52/48 in favour of Heads. Although the Day 1 lynch rate for wolves has certainly been historically better than utter chance would predict, it is nonetheless close to what chance predicts.

So even if past games had a bearing on the chances of this game, we're still in a situation where past frequency predicts that we're better-than-70% likely to lynch an Ordo on Day, leaving me firmly in the camp of preferring that we not lynch anyone at all on Day 1.

The only exception to that would be if someone could correlate the frequency of lynching a Wolf on Day 1 to the likelihood of lynching a wolf on Day 2 (and 3, 4, etc). If, for example, lynching a wolf on Day 1 meant a 50% or greater chance of lynching a wolf on Day 2 vs. lynching an ordo on Day 1 meaning a less-than-50% chance of lynching a wolf on day 2, then there would probably be arguments to be made--but my own training in Statistics reaches no further than the ability to say that 25% chance of getting a wolf is the same as a 75% chance of getting an ordo.

Now, even with this preference for not lynching anyone, I realise that lynching someone today, at the very least, gives us SOME useful information to analyse tomorrow. What I'm not sure about yet is whether or not it's more useful to have that information or more useful to have that extra Innocent vote.

EDIT: X-ed with all posts since Morm #95.

[Nogrod]

So Form, do you suggest we'd let the baddies kill each other first and not try lynching any of them? Just sit back as sacrificial lambs and wait whether they kill us all or whether they happen to kill each other first, or whether our gifteds save the day or tell us what to do?

Playing this game kind of means we villagers play it as well - and not only those who have roles will play it.

[Rune Son of Bjarne]

Originally Posted by Nogrod

So Form, do you suggest we'd let the baddies kill each other first and not try lynching any of them? Just sit back as sacrificial lambs and wait whether they kill us all or whether they happen to kill each other first, or whether our gifteds save the day or tell us what to do?

Playing this game kind of means we villagers play it as well - and not only those who have roles will play it.

Pathos much?

[Rune Son of Bjarne]

I apologise for the successive posts.

I do really have to leave, and so will have to vote.

I can either vote Formendacil, since he volunteered, or I could vote for Greenie, since I do not care for the buddying up her and Nog is doing (also she mentioned Kant).

++Formendacil

Yup... I don't want to see Form gone, and I would be deeply saddened if this lead to his demise. Though I spoke ill about Greenie, I would hardly say that her actions merits the label "suspicious", "annoying" would be more fitting.

[Kath]

I have read through the thread and what I have mostly come away with is a vague feeling of horrible nostalgia for my dissertation results section from years back thanks to all these statistics. I am with Greenie, English please!

I am unconvinced about our ability to communicate with the Dead Thread at all. Indeed apart from those roles with the ability to come back and therefore report about what is going on in there, I can't see how we would get any information between the two threads. But I also hadn't realised what aganzir said about Dead Thread voting not starting til Night 3 anyway, which means the fewer people sent there the better really, excepting wolves.

Therefore I am happy to throw my two pence in for this tied lynch toDay. Deadline is late on a school night so this is it for me. Am on mobile so can't check and see how the votes are going but Form volunteered for the position so:

++Formendacil

Hoping that highlights ...

EDIT: It worked. Also fixed an accidental capital letter.

[Nilpaurion Felagund]

Well, at least the 'nap' worked.

I wasn't planning to post at all, was just reading up on my mobile while trying to recapture sleepiness (bad idea), but some people haven't really caught up to the true implication of the three-party dynamic here. Maybe because I spent the first few days after joining pondering on the metagame of this specific village (before being captured by work, my other work, my other-other work, and Kousaka Reina, you beauty, you. Oh, and some piano/violin cartoon thing.)

Have only caught up to Form's (Hi! ) post 88, so if this has already been addressed... you'd ignore or misrepresent me, anyway, so who cares. ¯\_(ツ)_/¯

So...

Even if the Dead thread does gain a Baddie majority, it is highly likely it's a disparate majority. Again, there are two Wolf Packs. They don't like each other. They don't know each other, except through the mechanism of the Dead thread's NIGHT vote or Seer dreams. They're as interested as us Goodies to find out about Wolves, because all the information they have access to is the knowledge of the roles of two other people (their Packmates).

If one Wolf Pack does gain a majority in the Dead thread, then we'd be well on our way (if not already there) to getting just one kill per NIGHT. I think that's a good thing, yes?

That is all.

[Nilpaurion Felagund]

So mistress Rikae has already addressed the exact same points I have in post 90, just in a more succinct and understandable manner.

I should be going back to sleep, I guess.

[Formendacil]

Originally Posted by Nogrod

So Form, do you suggest we'd let the baddies kill each other first and not try lynching any of them? Just sit back as sacrificial lambs and wait whether they kill us all or whether they happen to kill each other first, or whether our gifteds save the day or tell us what to do?

Playing this game kind of means we villagers play it as well - and not only those who have roles will play it.

I suppose I can't expect anyone but myself to remember it, but I have a long-standing dislike against Days 1, and the main reason for that boils down to this very sort of situation. I'm not saying--well, I not trying to say--that we should be docile and let either wolves OR gifteds boss everyone else around and determine the course of the game.

What I'm trying to say is that, unlike the villagers, the wolves HAVE to kill someone toNight. In the normal course of events, where the village lynches someone on Day 1, you could say that the village makes the first move (and 71% of the time, or thereabouts, we get it wrong). By causing Day 1 to be a no-lynch day, we force the wolves to make the first move.

And, in a way, this is the way the game SHOULD be played: the narrative of each game uses the Moderator as a first victim of the wolves, to whose death the village reacts by trying to lynch the werewolves. In PRACTICE, however, the actual participation of the players starts with Day 1--in a real world, if werewolves started killing people in a closed community, we'd have a history of interaction with our neighbours on which to base our lynchings; in the game world, we have nothing but a Day 1 that precedes any actual choices made by the Wolves. The death of the moderator is presented as a fait accompli and we have to try and analyze the actions of people who never actually made any decision to kill that person.

I suppose this is sounding more like a defence for my dislike of Days 1 than anything else, but I think there's a nugget of truth here. The record bears out that Day 1 lynchings get it wrong significantly more than they get it right (better than 2 innocent deaths on Day 1 for every successful wolf-kill). Restoring the natural order of things, where the wolves HAVE to act first means that tomorrow, in what would then be a REAL Day 1-after-wolf-killings, would remove the red herrings of trying to ferret out whose was an ordo, who was a gifted, and who was a wolf in the original lynching.

After all, on Day 1 it could be entirely a list of ordos who lynches an ordo. Or it could be the Seer, Ranger, Hunter, and Loves who cast those votes. My point is that we don't KNOW who will have anyone killed today, but we absolutely will know that the wolves kill whoever is killed at night toNight.

[Macalaure]

Originally Posted by Formendacil

For another, although 1-in-4 is a reasonably good chance of hitting on a wolf, it's still not as good as the 3-in-4 chance of missing one altogether.

You know, the only time we have a higher random chance to lynch a wolf than to lynch an innocent is when the wolves win.

Originally Posted by Form

Even so, we can't calculate OUR odds of lynching a wolf off what has actually happened in the past, any more than you can say the odds of flipping a coin Heads or Tails on a given toss is anything other than 50/50--even if the ACTUAL flipping has come up 52/48 in favour of Heads. Although the Day 1 lynch rate for wolves has certainly been historically better than utter chance would predict, it is nonetheless close to what chance predicts.

You're not leaving me a choice, so here we go. According to math, the interval from 14% to 44% has a 95% chance to contain the population probability for the success of Day 1 lynches.
Yeeah, I expected the interval to be terrible, but not this terrible. Form has a point.

Regardless. The day is a time for lynching, void of scruple or remorse. I will not abstain.

Since Immanuel Kant? ...ahem.

Kant is annoying, don't ever mention him again.

Immanuel Kant was a real ****ant, who was very rarely stable...

Sorry, had to.


Ok, enough of this. The next time I post I will actually put down something substantial. Or at least I'll try.


edit: shaking my head at the forum's auto-censoring

[Eomer of the Rohirrim]

Originally Posted by Macalaure

Regardless. The day is a time for lynching, void of scruple or remorse. I will not abstain.

Hear, hear! Even if no-one is looking particularly wolvish at the moment, this is the first day of werewolf for AGES, and by gum we oughtta string someone up to satiate our long-frustrated hunger.

So enough with all this numerical babbling for a moment - who do we like the look of?

Personally, I'm always inclined to vote for very many of you, just because - certainly Lommy, Shasta, Nilp and Boro are never wasted lynches, if you ask me.

[A Little Green]

Originally Posted by Form

The record bears out that Day 1 lynchings get it wrong significantly more than they get it right (better than 2 innocent deaths on Day 1 for every successful wolf-kill).

The same holds true for the rest of the Days too, though. According to Mac's statistics, a wolf was lynched in 29 to 35% of the cases on all Days but, for some reason, Day 3 (43%, so still more often a miss than a hit) and the very end of the game (Days 7 and 8). So yes, we're statistically more likely to lynch an innocent than a wolf, but that is sadly not a feature of Day 1 but rather of the entire game.


EDIT: x-ed with Nilp and Mac

[A Little Green]

Originally Posted by Mac

You're not leaving me a choice, so here we go. According to math, the interval from 14% to 44% has a 95% chance to contain the population probability for the success of Day 1 lynches.

Erm... I'm sorry can somebody translate this into English please?