Open 722: Red Flag [Endgame]

[BlueBloodedToffee]

Wow.

Why did you not hammer me BV?

[zMuffinMan]

Oh time for my crazy theory then... creature BV?

[zMuffinMan]

Or BBT is just trolling

[BlueBloodedToffee]

Nah, I don't troll.

Wasn't expecting that at all

[Green Crayons]

I'm going to guess Hiraki as scum #4, but cfj is a runner up just for lols.

[BlueBloodedToffee]

Yeah, it has to be Hiraki

[zMuffinMan]

Hiraki is possible I suppose (well he's one of the only 3 possibilities really) and that would explain his willingness to vote BBT and not BV, but i was kinda ruling him out because of the loltunneling

[zMuffinMan]

Makes I guess more sense than creature? Just got paranoid at creature's unwillingness to vote BV

[callforjudgement]

There is no 'benefit'. It's just a natural reaction from a townie who thinks they have lynched scum to win the game.

I was going to save this to tomorrow, but given that it looks like game over now:

As town, I'm basically always surprised when a lynch hits scum (unless I have a particularly strong read or there are theory reasons to think that the target is scum, e.g. a Cop investigation). It's very hard to actually get scum lynched unless they're bussing, so normally the gain you get from your lynch is that mislynch-bait flips town and gives you more information about the scum while it happens.

On a Day like toDay, when (until the end) there wasn't a clear wagon and clear counterwagon, I'd expect it to be quite hard to push scum over the lynch line, so I was expecting whoever ended up finally getting lynched to be town regardless of who it was (but for us to get decent information about the scum in the process).

I agree that Hiraki seems like the most likely choice for the fourth scum under the assumption that BBT is town but BV is scum. (Mathdino has a similar voting record but would have unvoted/shifted by, at the latest, #2309.)

[Mathdino]

Lol bbt tunneling me nonstop the whole way through

[Mathdino]

I was actually gearing up to lynch Hiraki tomorrow

The way he reacted to my townread was terrible
Especially the "thanks coach" stuff. He'd have been more aggressive as town I felt

And then his entire voting record since that townread was bad

I didn't want to broadcast that read because i was afraid scum would kill themselves

[BlueBloodedToffee]

Lol, Math badly reading me the entire game and town reading scum.

[BlackVoid]

Wow.

Why did you not hammer me BV?

I laid it out in the scum PT (Hiraki was my partner btw) but we both agreed that if I hammered you and killed myself, Hiraki wouldn't be able to survive three lynches given people had him as "leaning town" or "not so sure as their other townreads" and POE would eventually catch up with him. I didn't see hammering you as a plausible victory path. Our best bet was to get somebody else lynched. Given how strongly many of the town played and how interactions cleared them or their predecessors, that somebody else pretty much had to be Archwing so I was planning to set him up for a lynch from the moment I replaced in pretty much. What I didn't count on was zMuffinMan derailing it which pretty much made it so that we couldn't win.

[Mathdino]

Told you guys Fumuki was locktown haha

My meta on Hiraki was obviously trash but I don't think i ever projected tons of confidence there anyway lol

[Mathdino]

Lol, Math badly reading me the entire game and town reading scum.

Man I nullread you the entire game

I told you I was playing by PoE and PoE worked

Dont give me that

I lynched scum lol

[callforjudgement]

I didn't want to broadcast that read because i was afraid scum would kill themselves

I didn't mention it for obvious reasons, but I'd have been fairly happy with a scum suicide N3 from a win-the-game point of view. Among other things, it'd mean that all the players generally considered town would still be alive, so with two mislynches + lylo left, we'd likely be able to solve the game by PoE. Given how quickly we identified Hiraki as scum today knowing BBT town and BV scum, we'd probably have been able to do the same thing D4 after a BBT mislynch, or D5 after a MathDino mislynch followed by BBT mislynch.

I'd have been fairly sad with it from an "ugh, I actually have to still play this game" point of view, though; 6:1 is a really tedious setup to play, even when you have interactions from earlier days to fall back on. (Without them, it's not really Mafia.)

[callforjudgement]

Incidentally, this setup gave me the idea for a setup of my own, but just in case I fall foul of ongoing games rules (the game's technically still ongoing), I'll wait for NSG to call the game as over before I post it.

[Mathdino]

I still think this setup is scumsided using white flag as a benchmark

D1 scumlynch was avoidable

Avoiding that is basically all scum need to do

[BlackVoid]

I don't think the setup is scumsided. I thought it was pretty balanced. The fact that scum actually flip when they kill themselves gives off interactive tells.

[callforjudgement]

I also believe this setup is scumsided. It's probably balanced if you remove the scum self-kills but they're more powerful than expected, as although they're pretty bad for scum from an EV point of view, the extra flexibility gives them space to salvage some particular bad matchups. If we'd only identified two likely scum rather than three (out of the 5-7 players who were heavily scrutinised), the kill-and-sacrifice play N2 would have worked much better.

Most notably, though, I think this setup has really bad skew problems. I need to go write a post about skew because it's a balance property that isn't widely discussed, but is fairly important to keep to a reasonable level if you want setups to be enjoyable. (A very brief description is that skew is the tendency for one faction to be able to make massive gains if they play well or get lucky – e.g. lynching scum in Red Flag is really good for town – but the other faction doesn't have that ability, e.g. the best scum can do in Red Flag is to drive a mislynch and then kill someone who's widely considered town.) It's possible that it also has kurtosis problems but I haven't made my mind up as to whether or not that's a real balance property yet, or just some random piece of statistics that isn't relelvant here. (Kurtosis is the property of swing mostly coming from unlikely events, as opposed to mostly coming from likely events.)

[callforjudgement]

I will say, though, that the balance really heavily depends on D1 play. The more scum are heavily suspected D1, the better the town does.

[BlackVoid]

Why do you think it's scumsided if only town can make huge gains by lynching scum early but not the other way around. Just curious.

[BlackVoid]

I think the playerlist definitely made it seem as though it was balanced when in reality it could have been scumsided I admit. Trying to simultaneously play around you, BBT, and zMuffinMan was more challenging than an average playerlist would be I think.

[callforjudgement]

Why do you think it's scumsided if only town can make huge gains by lynching scum early but not the other way around. Just curious.

Because town's huge gains are fairly unlikely to occur. The best chance is D1; trying to get scum lynched in a 7:4 is a nightmare (you need six of the seven townies to agree, assuming no bussing), and by the time you reach 5:4 you'd better have everyone proxy their vote to someone who's generally agreed to be town.

[BlackVoid]

Good point. I didn't read D1 closely enough but I don't know why ceejay and Hiraki both bussed TheBrie.