Balance and fun games (why town should be favoured)

[callforjudgement]

I disagree with the OP, but I see the issue in quite a different way.

The fact is that there's only so much you can do to improve town's win rate without shutting down scumhunting as a strategy (because night madness works better). If I see a setup with a 90% town winrate, I assume that the reason is that town's getting too much information from night actions, and that little to no scumhunting is actually occurring; in other words, I wouldn't expect that to be a result of any player actually playing well, and in particular I wouldn't expect the quality of the players to have anything to do with the game's end result. If I see a setup with 10% town winrate, then I'd assume that the issue with the setup is that town didn't have enough information to work with, but I'd also expect the relative quality (scumhunting ability for town, being-read-as-town ability for both factions) of the players to have a large influence on the end result. As such, setups with a high town winrate are likely to be fundamentally broken, whereas setups with a low town winrate is more likely to be superficially broken.

However, it's not actually the case that towns are tragically bad in practice. For example, 6:2 nightless (theoretical EV 50%) is considered highly townsided on Mafiascum, which means that town can outperform their theoretical EV in practice. This means that there's scope to raise town's EV quite a way without cutting down on the amount of scumhunting required to do well.

There's also an observed pattern that as soon as you give Mafia a standard nightkill, town winrates go way down even if the setup has the same EV and otherwise works much the same way. As such, I suspect that what's happening here is that there's a large disparity in skill levels between players on Mafiascum (I don't think this is a statement that many people here would deny!), and that scum are more inclined to "policy kill" good scumhunters than town are to policy lynch weak scumhunters. (Most likely, this is because scum have much more freedom in their NK choice – they can aim it at any townie, the majority of players – than town do in their lynch choice, as they have to try to hit scum.)

As such, I think fixes to the problem of high scum winrates need to work on reducing the number of scum NKs or limiting scum NK selection, as opposed to adding town power roles. It's not a coincidence that many of my setups have restricted access to scum nightkills, in one way or another. I also don't think there's a need to "fix" unfun games, given that the majority of players seem to enjoy the majority of games (often even if they're unbalanced, so long as the player doesn't know the game is unbalanced at the time).

Here's a thought experiment for the OP: given a 7:2 setup, is "town win if they identify a townie D1, lose otherwise" or "town win if they identify a Mafia member D1, lose otherwise" a better setup? (I mean, they're both terrible setups, but…) In the former setup (78% town EV), I don't see scum really being able to do anything dramatic or grandstandy; their best choice is to keep their heads down, try to act as they have no information, and hope to be selected at random (and town may well just randomize). In the latter setup (22% town EV), there's a lot more incentive for town to scumhunt and scum to pretend to be scumhunt, and in fact it's not unheard of for scum to actually be lynched on D1 of a Micro as a consequence of scumhunting.

[callforjudgement]

I feel like you answered your own question at the end.

I did, but I was curious as to whether you agreed with me, and this isn't a Mafia game so I don't have to leave the answer unsaid as a reaction test. (The point is that the unbalanced-in-favour-of-town setup is less interesting than the unbalanced-in-favour-of-scum setup.)

For what it's worth, I argued in another thread (the one on Mini Normal statistics) that it's quite possible that games would be more fun if intentionally balanced as mildly scumsided (~30-40% town winrate) rather than as 50% town winrate (although the argument relies on some assumptions that I'm not sure are true in practice). The reasoning is based on the dayplay/nightplay theory, in which giving town PRs to strengthen them (needed for reasonable town winrates in a Normal setup, which is what the thread focused on) tends to increase swing (which makes the game less fun because one side or the other has a chance of getting slaughtered) and tends to increase the chance of the game being decided by something other than scumhunting and looking innocent.

As for the underlying argument about how good the town should be, I'll put things this way: it's my belief that the average player on mafiascum.net has a better scumgame than towngame (this says nothing about the absolute skill levels of the player, just which position they're better at playing). You can't necessarily expect a player to become better at playing town faster than they become better at playing scum, either. Whether or not we want to take that into account when balancing setups is something that I'm far from sure about (although in terms of NRG setup review duties, I do, because it's quite clear from the typical views in Mafia Discussion threads that players value having a town win rate close to 50%).

PEDIT: I'm suddenly getting a huge urge to run a 6:2 NIghtless. The theoretical EV of this is 50%. I suspect that it's highly likely that town will win, and scum will consider the setup imbalanced in the postgame discussion (although, of course, this isn't guaranteed).

[callforjudgement]

I still think there should be mechanics in place to prevent scum from systematically killing the best town players in the game. It feels very against the spirit of the game.

Agreed, but I also feel as though scum should have the ability to kill players who are reading well (not because they're good players, but because they happened to be running hot that game). As such, protection should probably be based on the player specifically.

My setup Training Hospital (untested, and probably buried by now; the idea is that each townie was a semi-Doctor and two semi-Doctor protections on the same player would negate the nightkill) was designed to allow town to protect strong players (so that only the weak ones could get nightkilled). I think that's the sort of direction it's probably worth going down.

Likewise, I think it would help to have mechanics in place to discourage scum from simply bussing their weak members (as opposed to bussing their weak and strong players equally often), A simple way to encourage that is to give the scum a power role that looks valuable but happens to be useless in the setup (e.g. 1-Shot Strongman when there are no town roles that can stop a kill).