Werewolf 3 Game Thread

[Meiz]

Why? Whats your reasoning?

I had not done the math, and assumed we can afford more than 3 wrong lynches. Now that you pointed it out, I'm in favor of revealing masons at the end of this night.

With the help of that, I can also either go against novice tomorrow, or scrap the idea if the revealed roles don't support my theory. More regarding this after night ends.

[Rowain]

3 wrong lynches and it's over, right?

Yes thanks to the cull-kill we are down to 2 wrong lynches. If the baner manages to block a nightkill we get another wrong lynch .

[Roland]

I'll go back and read it. Don't take it personally please.

Nothing personal, and of course likewise. I know you put a great deal of thought and effort into targeting me, but that's honestly the only thing I remember. Of course, it's been noted by several people that I've been, at times, "unusually quiet", "riding the fence", "hedging my bets", etc. so it's not like I have a whole list of personal accomplishments to point to - fire&ice is basically it.

Yes thanks to the cull-kill we are down to 2 wrong lynches. If the baner manages to block a nightkill we get another wrong lynch .

We also gain an extra day if we lynch a 'Wolf, unless I'm wrong on the math. Correct?

[Serdoa]

So no, it is not helping, it is stictly worse. It only keeps the masons alive longer in specifically only the case where the wolves decide that they benefit by not targeting the masons. That's not a profit for us that's a loss.

Irgy, that is not true. It is a profit for us if the wolves don't try to kill the Masons because they could be protected by the Baner.

1. We have villagers whom we can trust. No, they do not know more but I know for sure that they are not trying to sew chaos and so I can trust what they state. If they state "I feel novice is a wolf for those reasons" I can trust that it is not a wolf trying to discredit a villager but there are really (good or bad) reasons to distrust a particular player. THAT is a big gain already but

2. the biggest gain: We lose a villager each night, no matter what we do. As everyone could be a wolf, everyone must be suspected. If the Masons reveal themselves AND the wolves don't try to kill them, they do kill one suspect. That makes it actually easier to hunt down wolves. So this play ups our chances that the Baner can either protect the right target OR that the wolves do help us by removing possible suspects.

On top of that we force the wolves to make a decision. A decision which tells us stuff about their thinking. Right now, no one here has a clue why they have targetted Sandover for example. Maybe because they target completely randomly just because no one has any clue at all anyway. We force them instead of letting us be forced.

Btw: I have explained why we have better chances to catch a wolf when our power roles reveal themselves. I have calculated it - not just gut feeling, hard numbers which I presented. Care to comment?

[Rowain]

We gain a day because we have a day were we don't lynch a villager but no we don't get more wrong lynches for killing wolfes.

[Roland]

We gain a day because we have a day were we don't lynch a villager but no we don't get more wrong lynches for killing wolfes.

Are you sure?

We start Day 5 with 10 v. 4. If we lynch a 'Wolf, that becomes 10 v. 3. Day 6 opens up with 9 v. 3. Assuming we lynch a Villager every day after:

Day 5: 10 v. 4
Day 6: 9 v. 3
Day 7: 7 v. 3
Day 8: 5 v. 3
Day 9: 3 v. 3 - Game Over.

Assuming we lynch only Villagers from here on out:
Day 5: 10 v. 4
Day 6: 8 v. 4
Day 7: 6 v. 4
Day 8: 4 v. 4 - Game Over.

That looks like a gained day to me. Where am I wrong in the math?

[Roland]

That looks like a gained day to me. Where am I wrong in the math?

Sorry. I misread your statement. I see what you're saying - yes, we gain a day, but not an extra lynch. Gotcha. That's what I get for trying to post from work - too many distractions, as usual. :neenernee

So, next question:
If we get back-to-back 'Wolf kills, do we then gain an extra lynch? Is that in the same vein as two Baner saves?

[Mardoc]

On the subject of reveals - they will certainly work to establish trusted villagers at a low cost in lynchings. Maybe not a low cost in nightkills. But from a lynching standpoint, we have two options: power-role reveals, wolf counter-claims. As a worst case, assume we pick one to lynch randomly. 50% of the time, we get the wolf, and a trusted villager, immediately. 50% of the time we kill the villager, but are still guaranteed the wolf the next day. So we're trading half a villager for a guaranteed wolf. It's even better for the masons, because it's half a villager for 2 guaranteed wolves. Because the loss ratio here is so bad, I strongly doubt that we'd have counterclaims in the first place.

I can see Irgy's point about having this benefit whenever you reveal, though, except for a crucial detail he's missing.

Irgy's probably asleep by now, but I'm going to argue against him anyway, on a detail. Yes, revealing the power roles early is a bad idea, mathematically, if you assume that odds of catching a wolf are equal and random.

This is the assumption that people are challenging, and I think, rightly. Back in WW1, in the lurker thread, someone (I think sunrise) kept an updated model of the odds of villager victory based on that assumption - and it always came out to be a horrible blow out, wolves winning by a landslide. Which of course didn't happen, because after a certain point it was no longer equal odds to catch a wolf or a villager, it was better than that.

Now, if you want to solve the problem mathematically, you have to make the simplifying assumption that the chance of getting a wolf is random and equal, or else spend research-paper levels of analysis on the true probabilities. Since no one has the time or energy to do that, they default to choice A, but realistically that's not the actual situation.

Why people think this is a good idea is because we're pretty much stuck on getting a kill ratio better than average, and they figure getting some hard information into the thread will give us enough information to bend the probability toward the wolves in the first place. Once we get enough information to catch one wolf, that will improve the odds of the next one (which under that simplifying assumption makes no sense, but is still true).

I don't like the idea of revealing between the night closed and 'so-and-so died' for the pure reason of that it seems like cheating. I think on balance, though, the masons at least and probably the fool should reveal tomorrow, IF AND ONLY IF it's clear that no one has any useful ideas, or of course if you end up getting almost lynched. Because at the moment, we need information, or else we'll die anyway. As long as we keep lynching people randomly, we'll lose. Once we get enough information to target wolves preferentially, we'll probably win.

The downside is that this benefit will be temporary, since the wolves will certainly try to kill actual trusted villagers. And the benefit does get bigger the longer we wait, if the special roles survive that long. It's just that, at the moment, the situation is starting to feel desperate.

[Mardoc]

If we get back-to-back 'Wolf kills, do we then gain an extra lynch? Is that in the same vein as two Baner saves?

No, because the whole question is whether we run out of villagers or wolves first. Every day that a villager dies and a wolf dies, we maintain the status quo. Every day that two villagers die, we slip closer to being outnumbered. All that matters is the difference between the wolves and the villagers.

What the Baner can do is give a day where a wolf dies and no villager dies, or only one villager dies. Those improve the odds.

That said, back to back wolf kills would still suggest we're on the right track and probably actually have enough info to kill wolves later on. But that helps by avoiding mislynches, not by anything else.

[Mardoc]

But from a lynching standpoint, we have two options: power-role reveals, wolf counter-claims.

Of course the other option is power-role reveals, no one counterclaims, gain a trusted villager immediately.