Live Action Mafia

A game of sneakiness and paranoia
It is currently Sat Oct 16, 2021 2:29 am

All times are UTC - 5 hours




Post new topic Reply to topic  [ 32 posts ]  Go to page 1, 2  Next
Author Message
 Post subject: python's thoughts
PostPosted: Wed Dec 27, 2017 1:53 pm 
Day 1: I clue 'BANGLADESH' with the intent of cluing 'BANG' as well as 'PORT' (Bangladesh is a coastal country) and 'EMBASSY' (and a couple other words very weakly that I didn't expect people to figure out). My team figured out 'BANG' meaning 'SHOT', 'MUG', and 'HEART', but proceeded to only shoot at countries. Countries were the known bystanders from this clue (plus the 'BANG' bystanders); I probably should have clued 'BANGKOK' instead but this would have given up 'EMBASSY', and hit BERLIN. BANGLADESH also had the advantage of cluing jamb and messing up blue's coordination. The only word shot successfully was blue shooting HEART.

Day 2: I clue 'TREES', incurring the bystanders BARK, ROOT, and TRUNK while cluing MAPLE, STICK, RAY, EAGLE, GRASS, and NUT. I clued 'TREES' rather than 'TREE' because EAGLEs fly above the TREES (not so much above a single tree though they would nest there), and more importantly TREES don't have a single TRUNK, and TREES BARK isn't a thing like TREE BARK is. ksedlar noticed the extra 'S' there but will team green figure out the reason? Meanwhile ksedlar is advising the team to not fire at SHOT which would actually be the first correct guess from the day 1 clue. At least a queue is building with ~40% correct clues, though I can't really use it yet since team green isn't consolidated yet.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Wed Dec 27, 2017 2:59 pm 
Ksedlar: "I'm pretty sure I'm the main contact of Brunnerj...so I can probably instruct them to shoot at dumb things." Ksedlar and brunnerj are both green but are both pretending to be red and trying to fake each other into thinking they're red.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Thu Dec 28, 2017 5:20 pm 
Day 3: My team gave me instructions with 10 minutes left in the previous day, and then changed their queue with 1 minute left, so I didn't have that much time to think up a good clue. Was going to do SOARING, 3 but their current queue would have resulted in them hitting a bunch of bystanders, so I decided to fix their queue with EXPONENT, 0. This ensures they get 3+ words correct this day, but doesn't add any new things to the queue. So the queue entering Day 4 currently has 1 correct word on it: yyao1's GRASS. So probably a large clue is needed.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Fri Dec 29, 2017 12:50 pm 
Day 4: Clue UPLIFT, 7. My team figured out EAGLE, SOUL, and ALPS, but missed BELT and SHOE (which literally LIFT UP pants/feet off the ground). Also RAY still isn't on their queue despite being clued by every one of my nonzero clues:

1) BANGLADESH -> RELIGION -> RAY (ok admittedly this is pretty weak. It was supposed to show up on their queue near the bottom though, and then I'd clue it to the top with my next clue...)
2) TREES live on RAYS of light (they knew pravi was green at this point so light wasn't a bystander)
4) RAYS of light are uplifting

Maybe AVOGADRO'S NUMBER -> VECTOR -> RAY will work? Or possibly STING cluing RAY and BELT (and maybe MOLE??) I really want to out ixa and adamyedidia though (so random guesses get more effective if needed) so AVOGADRO'S NUMBER looks best. Hopefully I don't have too many annoying parity constraints.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Fri Dec 29, 2017 12:52 pm 
ksedlar: "Okay, maybe let's say odd corwind is red, even cmcclena is red, and exactly 7 means neither/both of them are red."

If team green sticks with their queue, there will only be 6 red people alive at the end of this day, so this last clue would be illegal.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Fri Dec 29, 2017 3:22 pm 
aok: "the only downside, is the scenario where cmcclena is red and corwind is blue"

This is exactly the case. (jasonye's HP LOVECRAFT -> HEART clue has confused my team into thinking cmcclena is blue, despite corwind firing on green every day and cmcclena never firing on green)


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Fri Dec 29, 2017 3:22 pm 
aok: 'python is thinking "you idiots, corwind has been shooting at you every day and cmcclena hasn't; figure it out. cmcclena's words are just hard to clue."'

Sounds about right :P


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Fri Dec 29, 2017 5:09 pm 
the two sets that I can parity-switch between for True Names:

SET X: ANGEL, SHOP (cmcclena), HELICOPTER, HAWK (corwind)
SET Y: TIE, GRACE (cmcclena), BACK, LOCK (corwind)

Problem: Neither set contains a True Name.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Fri Dec 29, 2017 6:14 pm 
Arrrrgh why did ksedlar have to change the parity of the cmmclena/corwind thing? There are only 6 true names left, and their queue doesn't have a single one of them on it (after firings today). So I ideally want to give a large-ish clue to have them throw stuff out. But how do I get a clue for 4 out of DICE, RAY, SHOE, BELT, FLUTE, and MOLE? I want to clue AVOGADRO'S, 3 though they'd probably still miss the NUMBER -> VECTOR -> RAY connection, but maybe RAY would at least end up on the queue? Problem is this makes them think corwind is red...

They brought up SHOE and MOLE in discussions about UPLIFT but then forgot about them. So it looks like the parity check will force me to do 2/2/2 instead of 3/3, which isn't that bad assuming they guess all the 2's perfectly. With 3/3 I'd probably have to clue for 1 or 2 on the last day, but it's still a good deal better parity-wise imo.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 1:59 pm 
Ok so a win in 2 days seems impossible now, with my team only hitting 1 True Name today, they'll have to hit 5 tomorrow. They do have 3 of them in their queue, but they split their queue into three sub-queues, and I pick one of them. I can pick the queue with 2 true names in it, but one of them (MOLE) is at the bottom. So I'd basically have to clue BELT, DICE, FLUTE, MOLE for 4, and have them guess *all* of them correctly, while shooting RAY from clue-information.

SHAPE, 4 or GEOMETRY, 4 seem like the best in that BELT and FLUTE are both shapes (and DICE are perfect cubes) but the presence of PIE as a bystander is scary. Also, a bunch of other stuff has shapes like SCREENs, NOTEs, etc, and these clues also don't hit MOLE very well. The best-case scenario with one of these clues is my team might guess 4 correct names, leaving them with 1, while the worst-case scenario is they guess only 2 true names, leaving them with 3. In either case the best case is winning in 3 days.

So if I'm going for a win in 3 days, I should probably just be conservative. RANDOMNESS, 1 kills adamyedidia and has the right parity for ksedlar's queue to result in my team getting RAY and MOLE, which kill cmcclena and ixa. My team will hopefully use some of the leftover guesses on dgrazian, and then on day 3 we have 6 guesses for 6 dgrazian True Names, an auto-win even if I don't clue anything. The only problem with this is if my team decides not to shoot MOLE despite it being on their queue, or if they change their sets in such a way that one of them doesn't kill off everyone other than dgrazian anymore, or if they change their parity scheme to force me to clue 2 or 3 instead of 1 to get this nice set.

Predicted outcome of the game: TIE between green and blue.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 4:08 pm 
How come nobody figured out MOLE? Like FOOTLAMBERT is a unit, and MOLE is the only unit that's a codename. They have MOLE at the bottom of their queue because it's an animal with feet, but the unit connection is at least an order of magnitude stronger. Also, my cluing for 6 was intended to get them to dump their cmcclena queue, since 6 was the highest possible amount I could legally clue. (The last 3 words are incredibly weakly clued, though ankle BELTs are a thing.) But realistically I was hoping the clue would get them to hit SHOE, RAY, and MOLE, with the bystander ANGEL and whatever they decided to hit for adamyedidia and dgrazian.

In that case cmcclena and ixa would be outed, destroying their clues for those, and then I'd roll the dice with GEOMETRY, 3 or SHAPES, 3 in an attempt to get DICE (cubes), BELT (circle), and FLUTE (shape of a glass). It probably would only have a 20% chance of succeeding in winning in 2 days though so it's not that bad. But I'm still pretty sad about only 1 super obvious word being hit when FOOTLAMBERT clued 3 words very strongly, and my team continuing to use the top of their queue despite a clue for 6 (especially on cmcclena who I was *specifically* trying to prevent them from hitting, otherwise I would have done AVOGADRO'S, 2)


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 4:13 pm 
Oh I also remember why I didn't clue BANGKOK on day 1, which was because pravinas told me that she trusted Lucy and a few other people so I should strongly clue them if they weren't green. And BANGKOK -> CAPITAL was very dangerous, so I took the extra bystanders to avoid breaking coordination of my team. Though given my team shot only bystanders day 1, maybe BANGKOK would have been better, with me giving a parity bit later to say lucy was green?


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 4:16 pm 
Oh now my team changed the parity-check thing so it's no longer possible to have a provable win in 3 (since MOLE and RAY are no longer in the same set, they can't double-out from the queue alone). Though it probably won't be that hard as long as they interpret a clue for 2 correctly. Maybe since my team keeps misfiring on cmcclena, I should make her be the one they have to POE to death at the end. Anyway they'll probably change the queue again at 9:45 like they always do so I shouldn't spend too much time thinking about it :P Maybe they'll make it a possible win in 2 even (only works if they manage to put all 3 truenames in the same set)


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 4:16 pm 
OMG KSEDLAR IS THE BEST SHE REMEMBERED THAT MOLE IS A UNIT!!!

*very happy*


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 4:25 pm 
But instead of shooting at it, they're moving it to the top of the queue, which means the parity check no longer allows me to clue for 2 True Names :(


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 6:05 pm 
So it seems FOOTLAMBERT was this game's BREAKROOM, with the team guessing all the clues successfully (even FLUTE which suprised me)...but putting them all at the bottom of the queue.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 6:51 pm 
Also, ksedlar wants to espionage cmcclena to figure out how much HP she has left. She also knows red fired on cmcclena early-game. Will she realize that red might have damaged one of Casey's true names? In particular if they "know" cmcclena is at 2HP they'll fire on her 3x tomorrow which is useless since I'm trying to out casey *because* of all the strong bystanders that landed on her before my team figured out she was red. And my large clues have *specifically* been trying to get them to dump the cmcclena queue. It currently atleast has RAY on it, which is nice, if I decide to go for the 1 parity.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 6:55 pm 
There's also the possibility that if I clue small, and they think cmcclena's at 2HP, maybe they'll hit RAY anyway? I think my best option might be cluing for 3, which gets MOLE (outing ixa, who they're probably not going to guess more than they need to). The problem with a 3 clue is that they might click 3 times for the clue, instead of going down the cmcclena queue. But if I manage to clue adamyedidia's DICE, and one of dgrazian's words, they might hit RAY anyway if they think cmcclena's at 2?


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 7:28 pm 
Ok my clue is ICOSAHEDRON, 3, unless I think of a better one. This gives a very small chance of winning tomorrow, which only happens if my team somehow decides BELT and FLUTE (both shapes) are stronger than MOUNT for dgrazian. Which is basically not happening. But hopefully it sets up a win in 3:

- ICOSAHEDRON -> D20 -> DICE is extremely strong.
- ICOSAHEDRON -> BELT, FLUTE for being shapes. FLUTE is probably stronger, which is nice because BELT is easier to clue. These probably won't get guessed tomorrow but will hopefully appear on the queue for Day 7.

Meanwhile, the clue has the right parity such that they'll guess MOLE, unless they somehow decide ixa has a stronger clue. PLASTIC is the bystander I'm most worried about, but imo is certainly weaker than MOLE. However ksedlar knows jasonye and I made icosahedra for a Thing-A-Week out of plastic straws and tape so this is still pretty worrisome. I certainly don't want to do CUBE though or I'll hit everything that's a rectangle, like SCREEN, MAIL, etc, while not really solving the PLASTIC problem. Probably the best clue to avoid that would be RANDOMNESS which just hits DICE, but the fact that it has to be RANDOMNESS, 3 because of parity issues will force them to hit other, random things that I wasn't trying to clue.

So my predictions is they'll hit NOTE, DICE, MOUNT, MOLE with their main 4 shots, and then have 2 left over. Best case is they go down the queue to find RAY, and then also hit FLUTE. In this scenario, I'd only need to clue BELT the final day. Worst case is they decide ORANGEs being spherical are like icosahedra, click that, miss FLUTE and BELT, click PLASTIC instead of MOLE, and somehow miss the D20 connection (given ksedlar plays a lot of D&D I really really really think they won't miss this). But I think this clue should at the very least give me 2 True Names, and probably 3? And if not they should hopefully have a really good Day 7 queue? Also blue did really well today so hopefully they realize they need to win on Day 7, and don't split up the queue too much.

Another thing I'm worried about is cmcclena's health being overestimated means in the case where there are still 2 or 3 red players alive on day 7, my team might give me an illegal parity bit. For instance, if cmcclena and dgrazian are alive with 1 HP each, and my team gives me a parity bit where '3' is the correct answer, I'm fucked. But mod 3 should still be fine hopefully.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 8:43 pm 
Nvm. Blue guessed 5 correctly today, and will almost surely win tomorrow. We messed up at the worst possible time, gg :(


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sat Dec 30, 2017 9:06 pm 
Not sure if there was anything I could have done differently there at the end, with 6 words to clue and 2 days to clue them, with a queue full entirely of fakenames. That's just a really bad position. Maybe AVOGADRO'S, 2 would have gotten adamyedidia and ixa outed, and then I'd somehow have to kill cmcclena and dgrazian tomorrow with a clue for 4, which doesn't sound easy at all. I think my biggest mistake this game was that clue for 0. It was so tempting because the queue looked like

FAKENAME
FAKENAME
FAKENAME
FAKENAME
TRUENAME
TRUENAME
TRUENAME
TRUENAME

but since it didn't give any info other than removing 2 fakenames it ultimately left me in the situation of clearing out the queue, making my life much harder in the future. Giving the UPLIFT, 7 or similar clue right there would have added a bunch of misses over the next two days, but would have given me more leeway to clue smaller/use parity better as opposed to having to clue the entire set of words because the queue had no truenames on it.

In fairness I was *going* to give SOARING, 3 as the clue that day to get EAGLE, ALPS, and SOUL, but my team imposed a bunch of parity restrictions and changed their queue with 5 minutes left, and when I saw the new queue had that format I didn't like SOARING as much because it'd result in them shooting a bunch of fakenames at the top of the queue, so I knee-jerked changed to accomodate the queue, but probably should have stuck with SOARING. SOARING was also a better clue than UPLIFT in that it had way fewer UP* and *LIFT bystanders, and if my team's queue had true names on it I wouldn't be super sad about them going down the queue. UPLIFT, 7 was necessary because the queue was all wrong, so I had to clue to completely replace it. And then it was still all wrong after that, hence the ill-fated FOOTLAMBERT experiment. I actually thought there were 3 solid names in that (RAY, SHOE, MOLE) but the fact that the queue ended up with me having a bit to choose between three situations which each had exactly 1 True Name was really awful, and probably that bit of RNG completely destroyed our winning chances. (Technically, we could still tie, but the probability of this is ~2% right now.)

The other big mistake was, maybe BANG would have been better than BANGLADESH day1, since BANGLADESH only got 2 more truenames (PORT, EMBASSY) at the cost of about 6 bystanders (the countries mostly, plus putting food-related stuff on the queue). I was trying to do everything but my team shot nothing useful day 1. In fairness, clues are less useful without team info, so giving a huge ambiguous clue day 1 and having the bystanders cleared day 1 isn't the worst thing in the world, so I still think I like BANGLADESH. If lucy didn't have CAPITAL then BANGKOK would have been way better. Maybe BANGOR (a port city in Maine, also BANG is much more obvious)? That probably would have been best, and I did see that day1, I just wanted EMBASSY to exist. I also didn't really think people would weight countries as highly as they did, except for INDIA which was a great bystander for screwing with blue. So idk there, I think the clue for 0 was probably a bigger mistake though.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sun Dec 31, 2017 10:24 pm 
I think BANGLADESH was a pretty damaging mistake. Lots of guesses not only on bystanders, but on the wrong team.

The clue for 0 was unfortunate. That was a night I was busy all day and couldn't force decisions earlier.

It also sucked how often we were presenting you with bad options (3 sets with 1 true name in each, 2 sets with 0 true names in each).


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sun Dec 31, 2017 10:41 pm 
When you clued FOOTLAMBERT, we were sticking you with godawful queues and choices, but we only had 6 words left, and a 2-day goal. I think there probably should have been some way to get this done.

adamyedidia:
dice
michikyo/ixa:
mole
cmcclena:
ray
shoe
dgrazian:
belt
flute


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sun Dec 31, 2017 11:02 pm 
I think that BANGLADESH was definitely better than BANGKOK or BANGOR. I'm not sure if green would have figured out the BANGLADESH = BANG connection if I hadn't proposed it, and I only proposed it because I thought BANGLADESH and BANG were actually semantically related (in that people at ET tap each other on the shoulder and say "Bang-ladesh" as a form of trolling.) I didn't think it was legal to clue substrings of a word unless they were semantically related to the word.

I definitely approve of trolling blue by making Jackie look red, and (I advocated for not clicking INDIA at all), but I think that trolling should be done with fake names, not real ones. I think it's quite bad that green ended up shooting at TWO BLUE true names (CHINA and INDIA) and 0 red ones on day 1. If red had been more competitive (instead of blue being the main competitor) that could have been disastrous. Note that with EGYPT being a true name, green could plausibly have shot that too if it weren't for the RED SEA clue. Add in FRANCE, CANADA, TURKEY, GERMANY, HIMALAYAS, and you end up with 8 words that are IMO better fits than nearly any of the intended target words (though maybe we should have estimated EMBASSY to be higher, given how hard it is to clue.)

I think what happened is that Adam overestimated the strength of the BANGLADESH->BANG connection in comparison to looking at BANGLADESH itself. Afaik I was the only one who even noticed it, and I think it's also the case that I might not have even recognized the connection if it weren't for me proposing "BANG" as a clue last game. Probably "BANG" on its own would have been better. It would have solidly gotten HEART, SHOT, and MUG, at least. PANTS, POUND, SLUG, and RACKET are the most worrying bystanders, but at least none of them are true names of another team. And PANTS/POUND could maaaybe be avoided, on account of them being on greens.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sun Dec 31, 2017 11:12 pm 
Josh and I definitely saw BANG in BANGLADESH before you did Katie. I think the main issue was that we didnt really start discussing the clue/deciding what to shoot as a group till like an hour before rollover.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sun Dec 31, 2017 11:32 pm 
Oh, okay, sure. No one else had told me they liked BANGLADESH -> BANG before I brought it up, so I thought I had independently thought of it.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sun Dec 31, 2017 11:44 pm 
Amusing clue possibility that could backfire easily: CHARLES for 4. RAY CHARLES was the inventor of SOUL music. Music is connected to BELT (especially for singing) and FLUTE.


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sun Dec 31, 2017 11:49 pm 
I think "EQUIP" might have worked well.

My favorite clue is "ELAN", cluing Soul and Heart from french word and also cluing the character from OoTS


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Sun Dec 31, 2017 11:56 pm 
I think that TREES was a really good clue. I think that we shouldn't have asked for a parity check to constrain the size for day 3, especially while not saying what "large" was and what "small" was. EXPONENT was not enough, though it effectively anticlued LOG and ROOT quite well. Probably should have paid more attention to PIE, in spite of PIE not having been on the queue at that time.

++ ELAN. I think ELAN would get HEART, SOUL, STICK, DICE, FLUTE, and maybe even BELT, right? Though google isn't actually helpful for getting HEART or SOUL out of it, and LIFE would be a bad bystander if it were done too early. Also COMIC and NAIL for sure (Nale is Elan's evil brother, if I recall).


Top
  
 
 Post subject: Re: python's thoughts
PostPosted: Mon Jan 01, 2018 1:41 am 
Wow; there's a lot here I hadn't realized; green was both being much more clever than I thought throughout the game and in a trickier position than I thought near the end.

Well done, Python and greens.


Top
  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 32 posts ]  Go to page 1, 2  Next

All times are UTC - 5 hours


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group