[SETUP] Investigative Reporting

Mathdino · Post Post #4 (isolation #0) » Sun May 06, 2018 12:37 pm

Mafia Watcher can't work in a setup with 2 PRs. I'd even hesitate to include a rolecop, tbh.

Wrote a program for this setup without a mafia watcher.

Assumptions:
- Cop investigates randomly among non-clears
- Vig shoots randomly every night among non-clears
- Mandatory lynches/kills
- Mafia shoots confirmed townies whenever possible
- Mafia claims VT when run up (BAD ASSUMPTION)

The town EV came out to be 44%.
This is going to be much worse considering that scum is highly likely to claim cop or vig when run up in order to buy them an extra phase or two.

Mathdino · Post Post #9 (isolation #1) » Sun May 06, 2018 1:00 pm

If mafia shoots randoms that aren't confirmed, the town EV goes up to 55%. That's why I tried both.

Plus there's the argument that if scum never shot confirmeds, it becomes optimal for cop to just confirm themselves so they don't get shot.

And you can cc cop just because it'll buy you a phase. This will drop town EV significantly considering it'll take like 2 days to get any scum lynch.

Mathdino · Post Post #11 (isolation #2) » Sun May 06, 2018 1:07 pm

Got it.

If mafia tries to kill only people who aren't confirmed town (to try to find the cop), town EV jumps up to 60%. Killing PR claims --> ICs --> randoms is always better.

Mathdino · Post Post #14 (isolation #3) » Sun May 06, 2018 1:35 pm

That was accounted for, yeah.

Edit: It's low but it will demonstrably alter EV.

Mathdino · Post Post #17 (isolation #4) » Sun May 06, 2018 1:45 pm

One VT or one random? If vig starts confirmed innocent it's pretty bad.

I suspect the main reason that killing ICs is so +EV for scum is because you run a high likelihood of NKing the vig. I haven't accounted for a potential strategy of not killing ICs once the vig is dead.

Mathdino · Post Post #20 (isolation #5) » Sun May 06, 2018 1:52 pm

Turning one VT in the game into an IC (or having NSG guaranteed to be in the game and is a VT) causes an EV of 41%...

There's something about optimal strategy that I'm leaving out.

Mathdino · Post Post #22 (isolation #6) » Sun May 06, 2018 1:54 pm

Oh my bad I forgot to actually add the ICs into the program.

Cop head start brings town EV up to 48%.

Mathdino · Post Post #29 (isolation #7) » Sat May 12, 2018 7:07 pm

Oh shit okay that's interesting.

Mafia Publishing Cop gets to fake results on people. The 3 results are randomized and displayed on daystart.

Mathdino · Post Post #31 (isolation #8) » Sun May 13, 2018 4:14 am

I slept on it and decided BBMolla's proposal would work better with 2 Mafia Publishing Cops and a Mafia Roleblocker.

My current issues with NSG's proposal:
- It breaks the rule of not designing Opens that basically depend on one person not being an idiot.
- The high rate of running up PRs on this site mean that it's not unlikely town just gets fucked on D1, ruining the ability to test this setup and creating an apathetic mostly-vanilla Large.
- A lot of it revolves around optimising strategies that aren't necessarily immediately apparent. While this is a gold mine for players like me, it does create a power imbalance dependent on mechanical know-how.
- Claiming strategies are shit because while it does incentivise VTs to fakeclaim PR, allowing this into the meta now allows scum to fakeclaim PR, making the setup even more scumsided.
- It's scumsided, yeah. More so in the current meta.

I think I'd be good with it if it were 3:10.