Open 120 - Rebels in the Palace - Game Over! before 756

[Trumpet of Doom]

Sorry all. Had a competition I had to be at over the weekend and couldn't get online. And I have a concert this afternoon.

For my random/jokevote,

vote: Empking.

Come on, guys, it's in his freaking name!

Will try to post more tonight.

[Trumpet of Doom]

...come on, guys, give me something I can work with...

I can't tell if Shin's post about random lynching is a scumtell or just a newbtell; personally, I agree with Corvuus 67 that discussion is better than no discussion. I think an

IGMEOY: Shin

is appropriate here, but I'll leave my vote on Empking for now.

[Trumpet of Doom]

Yeah, I want to see what reasoning Shin has (had, I guess, due to this impending replacement) for thinking random lynches are our best course of action. If he/she/it had run the numbers and knew the percentage chances of win/loss/happily-ever-after, that'd be one thing, but if not, that's something else entirely.

Shin, I looked at the Texas Justice page you had originally linked, and the main dissimilarity is that that game can be ended in one night. This one will take significantly longer. I'm trying to work the numbers on it, but I'm doing it by hand, so it'll take forever. If anyone wants to do it faster than I can, they're more than welcome to.

And what's the deal with the self-vote? I was going to vote you before you self-voted, but I'm having second thoughts now that it's L-2... in the only game I've been lynched in and turned up town, I was lynched on page 5, so I'm not going to lynch anyone else this early.

I will, however,

unvote: Empking

and get myself out of the RVS.

[Trumpet of Doom]

ToD: This can be over day 1.

Sure, but I'm not expecting it to be (there's a one-in-twelve chance if we random lynch, people, that's less than 10%!) without some serious discussion and people figuring out who is more/less likely to be king/guard/rebel so that we really know who to vote for to lynch the king.

Also, give reasons with you're vote.

...you mean, my vote that's not on anyone right now? When I place it, I will.

If, on the other hand, you meant the vote I was going to place on Shin, it was mainly going to be for the "random lynching is obviously best" train of thought.

[Trumpet of Doom]

You have to admit that game breaking is a viable strategy in this game, just not a particularly fun one.

Don't you be the one to tell me that trying to break this setup isn't particularly fun. I have to go through and change some of my numbers because I realized something last night that I hadn't counted on when I was setting up the probability tree. And I'll see how viable it is once I get done doing it.

But since we seem to be agreed not to random lynch (my preference anyway), I'll set the math aside for now.

No Random Lynch.

[Trumpet of Doom]

There you all go making me do work again. I'm not going to definitely support random lynches until I know our chances of success with them... I'm on page 9 of a slightly-over-15-page Word document, though, so it might take a while.

[Trumpet of Doom]

There was a somewhat similar discussion in one of my other games, where someone brought up the point that it's hard to do any meaningful discussion if the lynch is a foregone conclusion. So even if we hit a guard or the king in the next four days, if we choose the lynches now, we won't be able to glean much from what's happened because there won't be much to discuss. And now, back to number-crunching.

(For the record, though, I checked the probability. [(4/12) + (8/12 * 4/11) + (8/12 * 7/11 * 4/10) + (8/12 * 7/11 * 6/10 * 5/9)] x 100% = 85.86% chance of hitting some form of scum if we random lynch for the next four days, which is better than a 6 in 7 chance (85.71%). I'm merely against the idea that hitting scum would be more beneficial in terms of discussion potential if we random lynch than if we discuss and build real, supported wagons on people who are doing scummy things.)

Speaking of people who are doing scummy things:

Vote Wall-e

Mod announces #14 will be replaced for lurking (in fairness, this game does have unusually strict activity requirements). He gets on and just posts this. No voiced opinions either way, just a vote that, to me seems like "here, let me skate through this by posting as little as possible and being on whatever bandwagon looks most likely to lead to a lynch"... I don't like it. Therefore,

Vote: Numberfourteen.

Also, welcome to the game, Wall-E! And orangepenguin, too... sorry I missed you when sekinj announced it the first time.

[Trumpet of Doom]

I've finally finished the calculations.

Without considering the possibility of a random lynch hitting the king when town exactly equals scum (since it won't work - a full majority can't happen), otherwise random lynches have rebels winning ~

65.63%

of the time, guards/king winning ~

26.20%

of the time, and the game coming down to the king and one rebel ~

8.63%

of the time (happily ever after). Yes, I realize the percentages don't equal 100%; blame it on rounding.

I'll probably get everything up on the wiki later tonight or tomorrow at some point, so if people want to look at my process and check it, they can. In the meantime, since random lynches have a >65% probability of winning,

Unvote; Vote: Wall-E.

[Trumpet of Doom]

Numbers are posted to the wiki. For those who want to look, the page name is (without quotes) "Assassin in the Palace/Rebels in the Palace Numbers".

[Trumpet of Doom]

Trumpet of Doom, could you post the full maths behind those statistics so that I can make sure they aren't made up please

Check post 172 again. I didn't want to copy the entire thing into a post because it takes up over 15 pages in Word, and our mod won't let us link to the wiki, so I did the next best thing.

[Trumpet of Doom]

Two days per guard.

If we lynch a rebel D1, the odds go from 66%, 26% and 9% to 60%, 33% and 9%, respectively. If we lynch a guard D1, they become 70%, 17% and 12%.

[Trumpet of Doom]

grumble grumble making me redo the numbers...

I don't care whether we roll dice all at once or roll dice for each new day, it'll have the same effect.

[Trumpet of Doom]

Go with the four? Okay.

Vote: Lynx The Antithesis

Once I finish redoing the numbers, I'll get them posted again.

[Trumpet of Doom]

Well, I have good news and bad news.

The good news is that at the start of the game, the chances of winning via random lynch don't drop much if we factor in that guards and king only need to outnumber rebels to win - ~60%
win, ~40% loss.

However,

(and this is the bad news) if we lynch a rebel D1 (which we did), our chances drop to ~53%.

[Trumpet of Doom]

Next up: orangepenguin.

Vote: orangepenguin

I would point out, though, that if OP's rebel, we're in LyLo tomorrow.

[Trumpet of Doom]

That's the second person in three days we've speedrandomlynched without giving them a chance to comment. Just throwing that out there.

[Trumpet of Doom]

Zwet, I don't quite see where you're getting that he has to be a guard, or even that there's a "very good chance" that he is. I see it as that there's a 5 in 9 chance he's a rebel, and that we're in lylo. If we're going to go with probability, I'd rather go with one with a >=50% chance of success.

Or, you know, we could try to scumhunt and see what happens.

[Trumpet of Doom]

What? No we don't.

Or if we do, you show me how you're drawing that conclusion.

[Trumpet of Doom]

Not more than, though, which is what you just said in 247.

[Trumpet of Doom]

I get:

3/8 + (5/8 * 3/7)

= 21/56 + 15/56

= 36/56

= 9/14, which is almost 65%. Not quite the >80% that Emp's 45/56 is (how'd he get that number, anyway?), but still good enough for me to support random.

Naturally, I'd also support analysis, though.

[Trumpet of Doom]

I get:

3/8 + (5/8 * 3/7)

= 21/56 + 15/56

= 36/56

= 9/14, which is almost 65%. Not quite the >80% that Emp's 45/56 is (how'd he get that number, anyway?), but still good enough for me to support random.

Naturally, I'd also support analysis, though.

Your number doesn't add us hitting two scum does it?

Yes, it does. Let's review: We're trying to catch at least one scum in two days. There are three ways this can happen:

We hit scum today (3/8) AND we hit scum tomorrow ( * 2/7)

OR ( + )

We hit scum today (3/8) AND we hit a rebel tomorrow ( * 5/7)

OR ( + )

We hit a rebel today (5/8) AND we hit scum tomorrow ( * 3/7).

Thus, P_{≥1 scum} = (3/8 * 2/7) + (3/8 * 5/7) + (5/8 * 3/7).

Now, watch carefully:

(3/8 * 2/7) + (3/8 * 5/7) = 3/8 * (2/7 + 5/7)
= 3/8 * 7/7
= 3/8 * 1
= 3/8

Substituting back into the original formula, we get P_{≥1 scum} = 3/8 + (5/8 + 3/7), which is what I put in the post you quoted. Make sense?

You've been pulling arbitrary numbers out of your ass all game long.

...um, I hate to burst your bubble here, but except for today, I've been the main one pulling numbers out of

anywhere

...

Do we want to random lynch or not? If we do, I'm fine with doing the dice roll, I'd just want to know if I should do tomorrow's as well as today's.

[Trumpet of Doom]

...zwet, help me out here, because I'm having trouble understanding your last post. How was TonyMontana supposed to say something about a game he hadn't joined yet?

Corvuus: You want us to post two people we think are probably scum and one we think is town (other than ourselves)? Fine by me.

Scum 1: Hohum. Again, #14 is an albatross on this guy's back, and I'm not terribly fond of ragging on someone for their predecessor's mistakes (I replaced into my first two games and three of my first four, the last of which should be finished soon)... but he's entered with the deck stacked against him. However, I want to see him try to get out of that shadow. (Wow, that's one mixed extended metaphor.)
Scum 2: zwet. At first I thought he was town, but over these last couple of days there have been things that have thrown me for a loop about his play. Most recently, his little exchange with TM.

Most likely town: This is not an easy question. I'm gonna have to go with Corvuus, although I agree that Emp probably isn't king.

[Trumpet of Doom]

I just noticed:

his stance/meta is similar/close to mine.

Just curious: Where are you getting my meta from? My only two completed games have been with Trumpetscum, and the only other game I'm dead in really isn't a great one to meta me from.

Now, a few thoughts.
If Empking is king, neither zwet nor hohum can be guards. (Similarly, if Emp is guard, neither zwet nor hohum can be king.)
If zwet is king, neither Emp nor Tony can be guards. (Similarly, if zwet is guard, neither Emp nor Tony can be king.)

I can see zwet as a guard, given Tony's list of things zwet said in 347, which rules out the possibility of Emp or Tony as king. While I will switch my vote if it becomes obvious who the king is, for now I'm comfortable with a

Vote: zwetschenwasser.

[Trumpet of Doom]

/prodded. As of now, I'm almost sure zwet's some flavor of scum (yes, you're trying really hard to sound like you don't know who the king is, I get it, dude), but at present, that's all I'm seeing. Whatever ortolan's claiming to have, it had better be good.

/ninja'd. Dammit, veerus, couldn't you have waited until ortolan said whatever he was going to say?

[Trumpet of Doom]

veerus deserves at

least

an

FoS

for that hammer - deadline was in two flippin' days, dude, you could have waited a bit.

ortolan's case on Corvuus/Qanqan... there's not much I can argue with there, though I must say I felt they were both doing good town impressions. I think I'll have to

Vote: Corvuus

for L-1 unless I see a really good reason he's town.

[Trumpet of Doom]

TBH, I'm sort of inclined to take ortolan as being a little scummy - while Qanqan's protesting that he's being unfairly singled out for the "Corvuus is town" comments, I'd probably complain about something like that regardless of alignment, so it's a nulltell for him. But missing the amount Corvuus seemed to be tying himself to me (or not taking it seriously), that's... if I were a third party, I'd almost think ort's the last guard and I'm king.

Vote: zwetschenwasser

- I'm still up for a zwet lynch for previously stated reasons, but if an ortolan wagon gets going, I'd be willing to jump on that instead.

[Trumpet of Doom]

TBH, I'm sort of inclined to take ortolan as being a little scummy - while Qanqan's protesting that he's being unfairly singled out for the "Corvuus is town" comments, I'd probably complain about something like that regardless of alignment, so it's a nulltell for him. But missing the amount Corvuus seemed to be tying himself to me (or not taking it seriously), that's... if I were a third party, I'd almost think ort's the last guard and I'm king.

So...I'm scummy because... um... nah I see nothing supporting that comment, all you comment on is Qanqan

Plus are you actually saying here "ortolan is scummy assuming I'm the king?" LOL

I am interested in TM's case on zwet and intrigued as to why ToD is playing like a jester.

No. Let's review:

Known guard Corvuus said I was the person he thought was most town, but when pressured for reasons, his were particularly unimpressive, as you yourself noted. Qanqan and I both said we thought Corvuus was most town... and you go after Qanqan instead of Corvuus? Were I a neutral third party, it would look to me either like you're a guard and I'm king, or like you're a guard trying to tie yourself to me so that if/when we lynch you and find out you're a guard, you'll take me down with you.

Thus, I think the guard is ortolan.

As such, Qanqan can't be king. I can see a guard leading a wagon on another guard. I can't see a guard wagoning the king. Of the other three players, I think the king is probably zwet, at least partly for these posts, in which he tries very hard to look like he has no idea who the king is. I think you're overdoing it a bit, zwet.

To get them both in a convenient place:
King: zwetschenwasser
Guard: ortolan

As such, my vote is staying where it is.