Newbie 1973 | Mountain | Over


Forum rules
User avatar
Ph0enix
Ph0enix
He/Him
Goon
User avatar
User avatar
Ph0enix
He/Him
Goon
Goon
Posts: 723
Joined: December 5, 2019
Pronoun: He/Him

Post Post #725 (ISO) » Thu Jan 02, 2020 12:04 am

Post by Ph0enix »

In post 720, PenguinPower wrote:
In post 718, Ph0enix wrote:
In post 711, Aloratom wrote: @Ph0enix What brought you around?
Well, I found out I apparently don't know what "bring around" means. I guess you are referring to why I ended up voting PP. To be honest, my job would have been far more difficult (cause there were stuff that bugged me about you both) had it not been for PP's opening vote. I didn't see a reason for it whatsoever (provided PP was town, that is; I explained in a previous post why I find the vote only makes sense for PP to make if he's scum).
I would have made the same vote as town in that situation (with the same read strength).
Hm. Interesting.
User avatar
Plotinus
Plotinus
Kitten Caboodle
User avatar
User avatar
Plotinus
Kitten Caboodle
Kitten Caboodle
Posts: 7611
Joined: March 13, 2015
Location: UTC+1
Contact:

Post Post #726 (ISO) » Thu Jan 02, 2020 7:25 pm

Post by Plotinus »

I've released the mafia PT: https://forum.mafiascum.net/viewtopic.php?f=90&t=81550

It's a good read.
The failure mode of clever is asshole.

Modding checklists | Sequencer is in Game 5 | Space II is in Day 4
User avatar
Plotinus
Plotinus
Kitten Caboodle
User avatar
User avatar
Plotinus
Kitten Caboodle
Kitten Caboodle
Posts: 7611
Joined: March 13, 2015
Location: UTC+1
Contact:

Post Post #727 (ISO) » Thu Jan 02, 2020 7:36 pm

Post by Plotinus »

Also I've been keeping stats for the last few games on what is the alignment of the person who is next due for a prod. Here are the stats for this game:

Day 1: scum 2/16, town 14/16
Day 2: scum 3/7, town 4/7
Day 3: town 1/1

It's commonly thought that scum lurk more than town does, but since I started paying attention my data (about 4 or 5 games worth now) shows this isn't usually true.

Method: When the dayphase begins, I put a note in my calendar for about 36 hours later to "prod everybody". The next time I check the game, I edit this note to have the name of the person who is next due for a prod and i note down a t or s. The next time I check the game, if that person has since posted, I note down the name of who i next due for a prod with a t or s and change the time, otherwise I do nothing until I prod them, then I mark down who is next due for a prod with the t or s. At the end of the dayphase I sum these.

In this game, on day 1, lurkers were much more likely than chance to be town. On day 2, lurkers were more likely than chance to be mafia. On day 3 the only person who didn't pst in the very short dayphase was town.

In previous games there was also 1 dayphase when lurkers were more likely to be mafia, but overall they were more likely to be town.

My sample size is pretty small still but I think overall the data will show that a "lynch the lurkers" strategy will do no better than random chance unless combined with a more sophisticated definition of lurkers that requires them to contribute more meaningfully than the low bar I have set for prodding. i can't raise the bar very much as a mod because it'd be too much mod intervention, for example requiring that contributions to the game be insightful or eloquent or contain at least 3 bulletpoints would be overstepping my role.
The failure mode of clever is asshole.

Modding checklists | Sequencer is in Game 5 | Space II is in Day 4
User avatar
the worst
the worst
Snuggly Duckling
User avatar
User avatar
the worst
Snuggly Duckling
Snuggly Duckling
Posts: 34628
Joined: November 7, 2015
Location: pond

Post Post #728 (ISO) » Sun Jan 05, 2020 12:37 am

Post by the worst »

That's a really good analysis!
I don't closely keep tabs on the alignments of replace-outs. I'd rather not monitor/publish the stats because I don't think it's really necessary to the game. But from keeping an eye on the replacement queue, I found a similar trend: there didn't seem to be a meaningfully high rate of any particular alignment replacing out. I don't have the data compiled as cleanly for newbie games (this would be very difficult!) so it's really cool to see a moderator keeping an eye on prod timers by alignment. It looks like our findings are pretty similar. If you can call my cursory observations findings. :P




Now while I'm here... I just wanted to take a moment to plug the Get to Know a Scummer sub-forum... :shifty:

Hello! As you may or may not know, I help run the
Get to Know a Scummer
feature on this website!

GTKAS is a subforum where you get to talk about yourself and learn neat stuff about other users!

If you haven't joined yet, you can gain permission to view the subforum (as well as the Speakeasy) by applying here!

All you need to do is:
  • Have at least 100 posts in mafia games (this will be displayed as your total posts on your profile).
  • Find the Speakeasy Inhabitants group, Select it with the button on the right hand side, then press submit "Join Selected" below!
  • One of the site's Discussion Mods (deep blue colour) will approve your request. This can take a couple of days.
Once you have access, you may join the queue here to get in line to receive a GTKAS thread!

Your GTKAS thread is where you'll be asked questions by myself and other users on site!

If you're interested, please apply for access and join the queue!
User avatar
PenguinPower
PenguinPower
He/Him
.peng
User avatar
User avatar
PenguinPower
He/Him
.peng
.peng
Posts: 22435
Joined: June 15, 2016
Pronoun: He/Him

Post Post #729 (ISO) » Wed Jan 22, 2020 10:09 am

Post by PenguinPower »

Image
<(") | (")>
Locked