[REVIEW] Open Setup Reviews

[mykonian]

More and more I think 7Ps just don't really work with a full night kill. It's too powerful.

I tend to agree - it's difficult to create a setup that is fair for the town that doesn't produce an optimal strategy or a boring/formulaic sequence of play. Town has no margin for error in 7p, so they do need a decent amount of power in catching scum and/or narrowing the lynch pool. This does limit creativity as far as role combinations go. It's easier to balance a 7p game when you dabble with mechanics instead of roles.

For the first sentence: in an open, there is a always an optimal strategy. Claims come ideally at a certain day. It's just easier to find it in a small game.

Scum doesn't have a margin of error in 7p either.

It's a lot easier. Using a poisoner mechanic for scum already weakens them by a lot. Scum didn't get a lot of shots anyway. No killing kinda ruins it and really limits the actions you can give town. I personally liked revenge mechanics (mafia lightning rod or as a game mechanic), because it gives you a little more space to make mistakes. I assume one could limit the amount of kills (even more) artificially... but I don't know how accepted that would be.

[mykonian]

I'm not infallible, so feel free to examine the provided code, but barring glaring errors being pointed out to me, town wins ~60% just by random lynching. That's a broken game.

That's a shitty definition of "a broken game". A broken game is a game where a strategy exists that gives town better chances to win without scumhunting. You have no idea if town could improve on that 60% by scumhunting. I'm pretty sure they can, because if they scumhunt poorly and actually gain no information, you are back at that 60%. It is an unbalanced game, that's true. But that's something different the broken.

Just compare this with texas justice. That's a known broken game. Everyone shoots in a circle, and if scum isn't next to each other in that circle, town has won. And since the chances of winning that way are much better then with lynching and no strategy, it pays to not play the game and use the strategy. That's broken.

[mykonian]

Uhh, 60% IS way better than average scumhunting, though. And it's easier.

Follow the Cop in the Original Newbie Setup wasn't 100% or even 80% effective either, but it made the game too damn easy.

If you scumhunt averagely, you get the random lynch expectation. Which is that 60%

You seem to assume that town hits scum 50% of the time regardless of the setup. That assumption is rather silly in an unbalanced game.

[mykonian]

That's a shitty definition of "a broken game". A broken game is a game where a strategy exists that gives town better chances to win without scumhunting. You have no idea if town could improve on that 60% by scumhunting. I'm pretty sure they can, because if they scumhunt poorly and actually gain no information, you are back at that 60%.

No. No, they're not.

If town scumhunts badly, they are hurting their chances, because the assassin will gain information regarding the king.

And scumhunting in AitP is rather hard, since there is only one scum player

and no night actions (except for the final kill).

blue: that's the game you are playing. If town scumhunts badly, yes they are better off random lynching.

Red: that has nothing to do with scumhunting. That has to do with your random odds. The more scum, the larger the chance you hit them. Scumhunting is how much you can improve on those odds by using information in your favor. The fact that you are more often correct the more scum there are doens't make you a good scumhunter.

[mykonian]

Would you have a link to the last paris mafia game?

[mykonian]

No lynching is as usual very determental to the town compared to the scum. The calculations done there are purely focussed on the threat of the mimes. The game is about the double threat for every team.

Mimes are the most fun role in that setup. They are hard to win with. And since people usually play this game for the first time, a lot of attention goes out to them. I don't believe that the no-lynch day 1 was really optimal, only anti-mime optimal. And regardless of all that, I doubt there will a lot of mimes who would complain they don't have an equal chance to win. Few people ever get to play a fun jester, and the mime is exactly that.

It's very hard to improve also, I think. The game is already pretty much stripped of all possible unnecessary bits.

[mykonian]

That sounds like it's worth an experiment.

[mykonian]

The base problem with regfan's plan is he's missing the lynch, I think. It's the only surefire kill people have, and plans of crosskilling tend to get ruined by scum commuting, leaving you with a 50% shot afterwards, which isn't a terrible improvement of a usual day, and runs the risk of hitting two townies at which point town gets set back majorly.

Lets see. If you extend regfans idea to three people shooting groups you get a 100% information return. If all die, too bad, they were townies who all shot. If someone commutes, that means two survive, of which one has to be a townie who didn't get shot because scum commuted, and the commuting scum.

You can't lynch day one and then make two circles of three, that wins the game for scum.

If you make one group, if all are town, you lost the game (20% chance).
If one is scum, you lose one townie there, and have a 100% lynch day 2 if scum chooses to commute (60% chance)
If there are two scum, same as the above, 100% lynch day 2. (20% again)

All other townies which aren't in that group of 3 commute, leaving day 2 with a 80% chance to lynch scum and 20% of insta lose. I think that's in town's favor, as losing by lynches is easier then that 20%, I think.

I could go on telling you in that case night two wouldn't be bad for town with just one scum kill there and 3 townies left, bringing the game to a 3p lylo with scum's commute still in and the one of the confirmed townie as well, but knowing all this from reading this post,

scum won't commute when he knows he's being shot

Which means in that same scheme, you lose in 20% of the cases, in about 60% of the cases three people die night 1, 1 scum and 2 townies and in about 20% of the cases one person dies and scum is trying to wifom you (hence the "about" 20% and the "about" 60%, as wifom tells you scum will have to choose suboptimal in the 60% scenario sometimes to make it in the 20% case, which would otherwise always lose them the game since you'd know both scum without wifom)

In the 60% case, which is all that changed because of scum's response, day 2 is a 3p lylo, with scum still having both one shots and town having two vig shots, meaning it has to be a straight up lynch as you can't beat the scum at night with any certainty. 33% win ratio is still better then doing nothing as town, as a mountainous 7p with a mislynch day one only gives you a 13% chance of winning.



If you did lynch scum day one, and you do the same, you have a 50% chance of victory, and a 50% chance of a 3p lylo (which adds to 66% for town). If you let scum shoot you once (which he could do at any night, which he wants to do the night when he wins, affected by WIFOM since town saves it commute for that night.) that leads to a best case scenario (where he shoots right away) of 60% and a worst case (where he shoots the night he wins) of 50%. If town manages to commute his shot you are back at 66%, making the breaking strategy slightly better then the strategy of simply not shooting as town.

You improve a little on the not shooting strategy by shooting controlledly as town at night, as a vigs chances are a little higher if the game was truely random (since he won't shoot himself), but tbh, I think that's a point where the assumption of randomness isn't good enough anymore to show a difference between strategies, and I don't know if it would pull up the chances of winning above 66% anyway.