RadiantCowbells's 8p Pre-Designed Setup


User avatar
callforjudgement
callforjudgement
Microprocessor
User avatar
User avatar
callforjudgement
Microprocessor
Microprocessor
Posts: 3972
Joined: September 1, 2011

Post Post #2 (isolation #0) » Wed May 09, 2018 11:30 pm

Post by callforjudgement »

Well, Micro 801 was something of a curbstomp for town (although town played much better than scum that game), and even ignoring the way the roles rolled, the setup is close to balanced EV-wise. Town is much better off here, so the question is, is the difference between a Closed and Open game large enough to offset that?

In this case, I don't think it is. The Vigilante can easily save themself by claiming. If they're dead, the Doctor can too (most players are going to guess there's a Doctor in the setup after a Macho flip, so scum aren't going to dare to claim it in case they get counterclaimed and thus a claim from town will be believed). So we effectively have two power roles likely to save themselves via the claim. On top of that, we have three role interactions that can give town a noticeable advantage: the vig killing twice; the vig killing someone and the doctor stopping the scum kill; the vig and doctor realising the setup and both aiming for the same target. There's no guarantee that any of these occur, but they each give town an overpowering advantage and the odds are in favour of at least one of the scenarios happening.

Another way to think about it is that the EV is almost certainly well over 50%, so even if town have no idea what's going on, they're fairly likely to win just playing randomly. That's normally considered undesirable.

I do like the idea behind the town roles here; it's just that the scum don't have enough power to compensate. Probably the easiest fix for the setup would be to remove a VT (taking the setup down to 8p); that has basically no impact in most balanced scenarios, but helps to prevent the town running away with the game if they get lucky early on. You might want to give town an extra, very weak power role in that situation so that setup speculation is more useful in an early 3:2 lylo, though.

As for balancing the setup with roles on the scum side, Roleblocker is probably too powerful (you'd want to remove the Macho from the Vigilante in that case, but it's still probably scumsided?), and Rolecop isn't fast enough for what's effectively a two-day setup. Perhaps giving town a Miller would introduce enough doubt about the setup to balance it? I'm not sure on that.

Out of interest, what happened the last time you ran the setup?
scum
· scam · seam · team · term · tern · torn ·
town
User avatar
callforjudgement
callforjudgement
Microprocessor
User avatar
User avatar
callforjudgement
Microprocessor
Microprocessor
Posts: 3972
Joined: September 1, 2011

Post Post #6 (isolation #1) » Thu May 10, 2018 5:36 am

Post by callforjudgement »

That Innocent Child is one of the most WIFOMy things ever. I don't think we want that in a Normal. (You could probably get the same balance effect in a less bizarre way via the use of a Miller.)

In the previous run of the setup, the Doctor died D1, which would swing the setup back round to a fairly normal balance. So scum got lucky and even so they had to survive both D2 and a 3p lylo. (The vig shot on scum isn't too surprising in a setup like this one, as lynches and vig kills should have similar accuracy.)

I don't think the Neapolitan helps that much as it won't actually allow scum to stop the Vigilante in any way (it'll be the morning of D3 by the time they actually locate and kill the vig using the role, which is already too late). The main advantage scum get from it is that it increases the chance that they shoot the Doctor N2, and thus reduces the chance that they shoot someone the Doctor is protecting.


I like the use of Ascetic rather than Macho on the Vigilante; they're equivalent in this setup but it makes setup speculation harder for town, which is useful in a townsided setup like this. Actually, if you simply take the modifier off the Doctor (making it a full Doctor), that might help the balance somewhat too. Now the vig and doctor are interfering rather than cooperating, which might get the balance into the place we want.

Btw, the version with a Loyal Doctor also has a breaking strategy. Suppose that three VTs die over the D1/N1 cycle (a bad but not terrible result for town). The Loyal Doctor can claim and ask town to vote for a vig target, then protect the voted-for target. Scum have two realistic options: shooting the vig target (if town) or shooting the Doctor. (They could also shoot the vig if they'd been identified by that point, but it comes to the much same thing as shooting the Doctor) In the former case, we end up in a 3:2 with two confirmed townies (the Vig and the Doctor). In the latter case, we end up in a 3:2 or 3:1, still with two confirmed townies (the vig and the vig target). So town have effectively had four attempts to hit scum, failed
all
of them, and are still in a 3:2 with 2 confirmed townies; that's much better off than they should be in that situation.

So I'd suggest something like this:

Ascetic Vigilante
Doctor
VT or Miller
VT
VT
VT
VT
Mafia Goon
Mafia Goon or Mafia Neapolitan

That avoids any attempts to break the setup, doesn't confirm townies based on setup spec nearly as much, and comes to much the same thing as your setup when players are playing for the dayplay rather than the nightplay.
scum
· scam · seam · team · term · tern · torn ·
town
User avatar
callforjudgement
callforjudgement
Microprocessor
User avatar
User avatar
callforjudgement
Microprocessor
Microprocessor
Posts: 3972
Joined: September 1, 2011

Post Post #9 (isolation #2) » Thu May 10, 2018 1:15 pm

Post by callforjudgement »

Let's analyse Macho Vig + Loyal Doctor + 4 VT vs 2 Goon (i.e. the 8p version of the setup).

D1 the setup goes down to 7p; N1 probably to 5p (possibly 6p, but that'll typically have one confirmed townie). That leaves us in a 3:2 worst case. However, at this point the vig can't really shoot (this is the main difference from the 9p setup) so we basically have a 3:2 lylo with 1-3 confirmed townies (0 is marginally possible, but requires the vig to shoot the doctor and scum to shoot the vig, so isn't worth worrying about).

This setup might actually be possible to calculate an EV for:

1/28 chance of a town win N1
12/28 chance of one scum dying D1/N1
15/28 chance of no scum dying D1/N1

I'm assuming crosskills won't happen because they really shouldn't. If the Doctor stops a kill (1/12 chance, taking into account the chance that the doctor is run up D1 and the chance that the Doctor tries to protect the Vig), then there will be 6 players alive on D2, at least two of which (the Doctor target, and any surviving town power role) are confirmed. Three confirmations is a forcible town win (this happens if nobody shot the Doctor and scum didn't shoot the Vig). Town is 11/18 to win if only two players are confirmed. It's possible but fiddly to calculate an exact value for this branch so I'm going to approximate it and say that town wins 7/8 of the time the Doctor stops a kill (this takes into account things like town lynching scum).

More normally, the Doctor will miss N1. It's a bit better than 50:50 that both town power roles will be alive in this situation (again, I haven't calculated the exact value). If they are, we'll be in a 3:2 or 4:1 with two innocents (with EVs of 1/3 and 2/3 respectively; more in the 4:1 case if the town power roles think of focusing fire on a single target). If one is dead, there's only one innocent, so the EVs are now 1/4 and 1/2 respectively. So the EV for this branch comes out at 45% or so.

The 8p setup's actually pretty close to 50:50 in EV; probably marginally below 50%. I don't like to go too far above 50% because if we habitually put EVs that high, lynching randomly will become correct play. The relationship between EV and town win rates depends on how much opportunity scum have to control their nightkill (in a Nightless town habitually wins even at very low EVs), and in this setup, scum don't have that much opportunity to kill because the game ends after 2-3 Days. So "approximately 50%" is likely to be a balanced EV here.

For what it's worth, mith's analysis shows that adding extra townies to an otherwise vanilla vig setup has only a very small EV impact (maybe 2-3%), so although he didn't analyse 2:8 specifically, the Mafia Universe setup probably has around a 52% EV. That's pretty comparable to the 8p setup with the Loyal Doctor (and Macho Vigilante to confirm that there's likely a Doctor in there). Setup spec should largely cancel out between town somehow not being able to figure out that the Doctor claim is town despite all the clues, and town realising that they can focus-fire the Vigilante and Doctor on the same target.
scum
· scam · seam · team · term · tern · torn ·
town
User avatar
callforjudgement
callforjudgement
Microprocessor
User avatar
User avatar
callforjudgement
Microprocessor
Microprocessor
Posts: 3972
Joined: September 1, 2011

Post Post #12 (isolation #3) » Fri May 11, 2018 4:59 am

Post by callforjudgement »

Yes, good to pass the 8p variant. VT×4, Macho Vigilante, Loyal Doctor, Mafia Goon×2.

/approve

I have no idea what to do next because this is a new process. I guess I'll PM implosion because if that's wrong, he'll be able to tell me what I should be doing instead, and it doesn't divulge any information it shouldn't.
scum
· scam · seam · team · term · tern · torn ·
town

Return to “Completed Game Reviews”