Provable randomness

[PokerFace]

I allow randomness in all my games. I really don't see how you being random limits other players

even if you random vote with dice, someone can still vote you for choosing to be random over choosing to scum hunt legitimately. Hell i'd want to vote you for that

[PokerFace]

Really, the problem with verifiable randomness is that

town can come up with a plan for everyone to follow, and scum can't secretly make their choice non-randomly.

So town trades away their ability to use their greater numbers on reads, but it disrupts scum influence and crucially eliminates means scum can't COMBINE their influence. Because scumhunting is not that much better than randomness, the gamestate can reach points when this enforced mass plan is better than scumhunting. Boo!

One dude random voting isn't really the problem. The problem is that verifiable randomness allows plans that aren't really healthy for the game. Banning all verifiable randomness is just the smart solution because it's bright line enforceable whereas "no mass plans involving verifable randomness" really isn't.

And yes, as Ether notes, it needs to be "no verifiable randomness" and not "no dice tags" because there are a TON of ways to generate a sufficiently random number.

This is really the issue I think. If you are scum and you object to the plan, it makes you look scummy, and you can't secretly avoid voting your partners.

If that's the issue then I think the real problem is that the game setup is not balanced well. If your setup gives town an advantage by being random then your setup sucks and you need to re-balance it before running it again. Don't ban randomness. Just don't play badly designed games

[PokerFace]

AITP is not a bad mafia setup. Its not mafia at all. The game has informed majority vs uninformed minority, so its not mafia. Some people play it and its fun, i play it alot. Because its not mafia, it does not need to be played like mafia. Scumhunting is actually anti-town in AITP.

Provable Randomness is slightly less than 50-50 in AITP. You have 1 king and 1 assassin. If you roll the dice to get the player lynch schedule and its got the assassin way before the king, then town should win. If you roll it and king comes before the assassin or assassin and king are the last few, then town will loose. It likely gives town a 45% chance at winning. But since scumhunting gives the town only a 33% chance of winning, you want to be random in AITP

[PokerFace]

For a bad mafia setup that gives town a 70%+ advantage when using randomness you want Texas Justice

Dice has been used to decide that game before and it helped town alot!

And because that setups sucks, Mr. Flay and Myself have designed many alternatives

[PokerFace]

Proving randomness is a

Good

strategy in AITP. Town has 45% chance the king will be in a good place when they randomize the player list to decide who gets lynched when.

Its not broken

strategy as Assassin should still have at least a 55% chance he will win.

How is PF estimating these numbers you ask? Well in AITP assassin wins by killing the king or by being the only player alive with the king at the end of the game. If king is before assassin in player listings than Assassin will win. If Assassin is one of last 2 names in the list than Assassin wins. If assassin guesses right, he can still suicide bomb the king when he dies

50% assassin is after king when list is randomized. 50% assassin is before king when its randomized. HOWEVER You need to subtract from that 50% since assassin wins if he is in last 2 and he also wins if he guesses right when he dies. I'd say the odds of either of those two things happening is at least 5%. It may be more than 5% but either way since you are subtracting from 50% Town's chance of winning will be less than 50%. And since Town's chance will still be less than 50%, its not broken.

If it were broken, town's chances would be greater than 50% like how they are 72% in Texas Justice

[PokerFace]

Texas Justice is a hugely townsided setup if you just randomly shoot, and I do believe that Town did randomly pair people up to shoot there.

Yes, that is correct. 72% is hugely town sided and they did random pair who would shoot who

As far as your other comment I'll have to do some math and get back to you. I was retired/on hiatus OCT 2011. So I'm not sure how correct your findings there are. Give me sometime to look into it and I will get back to you

[PokerFace]

ok I have a question. You have a chart with (Minimum Day Assassin Will Lynch vs Town Win %)

That chart has town win chance start at 60% and increase to 64%. Why would town get a better chance at winning each day? Logically the assassin should get a better chance each day as there are less and less wrong answers for him to choose from. So is your chart backwards? Because if its not, then I think your math is wrong

Also is the fact assassin and king being last 2 alive makes Assassin win getting factored into your code? Because it should get factored in

[PokerFace]

So your plan is not true randomness but rather loosely guided randomness since you have lyncher's picking lynchees. Correct?

So your code is only trying to find the odds the assassin gets lynched? Or does your code have anything that factors assassin's chance of making the right suicide bomb once he has been lynched?

[PokerFace]

Ahh ok then i think i understand what you did. True randomness is roughly 45%-33% town sided. Your guided randomness is obviously better than that. But I'm not about to say your idea is greater than 50% until I do the math myself. I should have time to do the math tomorrow evening or Saturday.

[PokerFace]

If that's the issue then I think the real problem is that the game setup is not balanced well. If your setup gives town an advantage by being random then your setup sucks and you need to re-balance it before running it again. Don't ban randomness.

PokerFace are you claiming that any setup with a town sided EV in unbalanced? If so once again I'm going to have to say you're objectively wrong. Go design a 50% EV setup, run it twenty times and see what happens.

No that's not what I'm claiming

I'm claiming any setup where randomness > scumhunting is an unbalanced or broken setup

[PokerFace]

The issue isn't that randomness is a breaking strategy, it's that proveable randomness let's players generate information that isn't at all compromised by their alignment. It's the same issue that we have with the red text of truth and other trust tells. It's not on their level as being able to break the game but the issue is that it's a more winning strategy then actually playing the game. What exactly do you think we gain by allowing it?

Yay I never understood that red text of truth stuff. People can choose to lie and start using the text to lie, right? The only reason they don't is because they value how good they are as town more than how good they are at being scum. Any buddy that does that is an idiot. You should value your desire to win equally. You should want to win as town and win as scum equally. So if you aren't lying with red text next time you are scum, then you aren't playing to your win condition to the best of your ability. Which means you suck as a mafia player.

Back on the provable randomness topic I don't think it always let's you generate info regardless of your alignment. If the math ends up in your favor and you follow the plan, then you are the alignment that plan benefits. If the plan actually benefits scum and you are scum and you trick the town into thinking the plan benefits town, then well done!

[PokerFace]

Okay, so in an n player game of assassin (1 king, 1 assassin, n-2 guards), if
1. A player is chosen at random to decide randomly on the day 1 lynch
2. The lynchee each day decides the next day's lynch at random
3. Town never lynches the king nor the day 1 lyncher

assassin win rate if I did the math right is H(n-1)/(n-1), where H(n) is the nth harmonic number. This can be approximated very well by (ln(n-1) + .577)/(n-1) for values of n that are for some reason too large for you to want to calculate H(n-1) (only useful for theoretical purposes really...). For 9 player AiTP, town has just over a 66% win rate.

what you just described was CITP

Contagion in the Palace

And it is very town sided when the mod locks the thread and correctly reveals roles when players are lynched. If they late reveal or no reveal and leave the thread unlocked you get this

[PokerFace]

So that is what you're claiming?

Town EV's > 50% it is (at least based off data culled from opens / newbies) to dice it then it is to play, MS has for at least the period of several years where I was active preformed worse then random in those areas. Quite literally randomness > scumhunting, it may not be glamourous but that's how it was. Proveable randomness wasn't removed on a lark, it was removed because if it is a legal play and players are playing by rule #1 (play to win) then you can have some truly truly dumb outcomes.

Things were that bad or are still that bad? If they are still that bad then we may want to improve the newbie setup and or run some better opens

How did things get that bad that we were better at being random then we were at scumhunting?

[PokerFace]

The issue isn't that randomness is a breaking strategy, it's that proveable randomness let's players generate information that isn't at all compromised by their alignment. It's the same issue that we have with the red text of truth and other trust tells. It's not on their level as being able to break the game but the issue is that it's a more winning strategy then actually playing the game. What exactly do you think we gain by allowing it?

Yay I never understood that red text of truth stuff. People can choose to lie and start using the text to lie, right? The only reason they don't is because they value how good they are as town more than how good they are at being scum. Any buddy that does that is an idiot. You should value your desire to win equally. You should want to win as town and win as scum equally. So if you aren't lying with red text next time you are scum, then you aren't playing to your win condition to the best of your ability. Which means you suck as a mafia player.

Back on the provable randomness topic I don't think it always let's you generate info regardless of your alignment. If the math ends up in your favor and you follow the plan, then you are the alignment that plan benefits. If the plan actually benefits scum and you are scum and you trick the town into thinking the plan benefits town, then well done!

Post one of each day a dice tag ranging from 1 to X where X is the number of living players will be rolled, each player will vote for the chosen player, this will continue until the game ends.

This plan benefits town, any player who deviates from it is either scum or breaking rule #1.

What a fun game of mafia.

Does provable randomness always mean that plan? I didn't read this entire thread. I was under the impression provable randomness was using the dice

once at game start

. Not every day using it

If provable randomness is always that then we likely need to ban that

If provable randomness is not necessary that or it does not give town greater than 50% win, then it does not need to be banned everywhere

[PokerFace]

to refine what I said earlier: I claimed any setup where randomness > 50 > scumhunting was an unbalanced or broken setup

Math puzzle's are fun, the first time. After that first time, they aren't fun any more. And if you are scum, it doesn't sound like it was ever fun. I think that's why we got rid of all those follow the cop setups in the first place

Provable randomness can reduce mafia to a simple mathematical question of which side can absorb lynches proportionally better.

It is incredibly dumb, I don't understand why people are defending it.

you can chalk it up to naivety. I didn't think provable randomness was as strong as you say it is. Has it actually been done in a game and yielded the exact results you are outlining?

[PokerFace]

It was banned no?

Why exactly do we want to bring it back?

I didn't think this thread was about "bring it back!" I thought it was about "why was it banned?"

[PokerFace]

I'm not trying to provide reasons for it to be unbanned. I'm trying to figure out why it was banned since I didn't think it was strong enough to be banned. I can provide a link to some math I did a long time ago where I ruled random lynching was pro-scum in RITP. As the kingdom is town in AITP and the kingdom is scum in RITP (they are somewhat polar opposite) it makes sense random lynching would still favor the kingdom and thus now favor them with their now opposite alignment. But if your theory of which side can absorb lynches better was true then you would expect random lynching to still favor the town in RITP. And it doesn't

http://forum.mafiascum.net/viewtopic.php?f=94&t=59621

[PokerFace]

1-(6/7 * 5/6 * 4/5 * 3/4) = 58%?

are you sure the 1- is necessary? shouldn't it be just 42%?

[PokerFace]

The software I used to use to compile c# does not seem to want to work. So I did what I could math wise.

I figured if I was gonna find an error it would need to be a big one to send the 60% to below 50% and I did not find said big error. So your math is good

I did find it interesting that 6 player AITP was 50% ev with random lynching. I once believed 5 player was the most balanced version of AITP. Seems its actually 6

I designed something called WITP a while back. Can your code be applied to it to determine when random lynching gives each kingdom 33% odds and assassin 34%? There are 2 kingdoms in WITP so striving for 33%, 33%, 34% would be most balance I could get.

[PokerFace]

well fuck. try this

War in the Palace

[PokerFace]

In playing the game it seems most balanced at either 7 or 9 total players. You can't have 8 players since you got 1 assassin, 2 kings, and 2-3 guards for each king (4-6 total). King A and King B are not on the same team and neither are their particular guards

Which do you think would be better balanced 7 or 9? I lean towards 7 as less players means less info for the assassin to find. In running the game I believe assassin has won 4 out of 6 times. One loss was with 7 and one was with 9. Three of the four assasin wins were with 9 and one of them was with 7

[PokerFace]

because of this thread, I have decided to run
Contagion in the Palace
during the next marathon day with 4 guards, 1 king, and 1 assassin. (6 players total)

Running it should provide some interesting observations. Feel free to join it during the july 2-5 marathon weekend http://forum.mafiascum.net/viewtopic.ph ... 6#p6956066

[PokerFace]

This should effectively limit discussion to zero outside day 1. And the assassin has even less means of protesting their lynch as it only take one player to eliminate the assassin with the disease. This version of AITP truly is a lottery.

What are you intending to get out of this?

According to previous research on AITP with provable randomness, 6 players results in 50-50 odds for both sides

The first running of CITP was ruined by the mod revealing players' roles only

after

they had selected who they would infect next. The player's role should be revealed once the player is infected. This is

before

they select their infection target.

Yes, the game has a lottery element, but it may not be "directly" hindered or enforced. During the game the

thread will not be locked at anytime

. While a player is selecting their target, anyone is allowed to post. Will you help town guards, as they select who to infect? Will you try to confuse the Assassin, as he selects who to infect?

I'm basically running a game where provable randomness results in 50-50 odds. While allowing players to attempt wifom should they perceive it can better their odds. I plan to study if wifom is used in the heat of the moment. And how it is used in a heat that is generated faster than in normal AITP.

I am running this game during marathon weekend as its lottery element makes it unsuited for the main forums and I feel the heat/drama will increase in a marathon format without offending the mind sets of a player. Players get invested in month long games. Peoples feeling get invested in month long games. Peoples feeling get hurt in month long games. Something that is a lottery should not be an investment that could hurt feelings

I plan to update that wiki page shortly as its initial draft lacks some of the information given here

[PokerFace]

I edited http://forum.mafiascum.net/viewtopic.ph ... 5#p6959915

do you mean that sarcastically or are you actually curious, Magua?

[PokerFace]

Well put

I have updated the wiki to better represent the purpose of CITP without dismembering its integrity/balance as much as it had previously been dismembered

[PokerFace]

Ran CITP 3 times over marathon weekend.
http://forum.mafiascum.net/viewtopic.php?f=57&t=62458
http://forum.mafiascum.net/viewtopic.php?f=57&t=62466
http://forum.mafiascum.net/viewtopic.php?f=57&t=62533

Analysis:

Is 6 players balanced?

I would say Yes as the end game was a 1/2, 1/3 guess for the assassin in all 3 games. Results I found, agree with Magua's math

Was wifom used?

In game 1 there was a moderate amount. In game 2 there was too much. In game 3 there was practically none. Game 3 was the most devoid of wifom and discussion of any "In the Palace" game I have ever seen!

Why did kingdom win only game 1?

Game 2 had lots of bad wifom despite the speedy day 1 lynch. Game 3 had almost no wifom, but it took them longest to reach a lynch. 2/5 town players were put at L-1 in game 3 before a lynch occured. One town player also self voted in game 3. So voting made it 1/2 for assassin before day 1 was even over. Game 1 had only 1 player get to L-1 which was the same player lynched day 1. Discussion and WIFOM in that game didn't point toward who was or was not king. It pointed toward who was or was not assassin. This type of WIFOM is not great, but its better than the who is king WIFOM. Also this WIFOM was not used by all players and did not encompass all players. Some players were not used in it or did not use the WIFOM at all. All players using it can effect the game outcome both ways for good or bad pending how each use it. But using all players in it, is what I think guarantees an Assassin Victory. If each player is talked about for being scummy or townie or guard or king or assassin, then discovering who is towniest and king can be very easy for an assassin.

Can WIFOM better a players odds of wining in "In The Palace" games?

in theory yes, but getting every player in a game to use it correctly and the same way is harder to organize than the optimal pro strategy in AITP. It will always be easier for 1 player to formulate a plan for 1 player (assassin) than for 1-X players to formulate a plan for X players (Kingdom). Thus I think the "everybody stay silent and just lynch someone immediately strategy" is always going to be the better strategy in "In the palace" games. Lynching someone immediately is the most important part of that strategy as it aids the cutting down of discussion and WIFOM part. Its the first piece of the puzzle. And without pieces you will never complete the puzzle. "In The Palace" games are more like puzzles than mafia games especially since its informed majority vs uninformed minority.

Is provable randomness the way to go in "In The Palace"?

Technically its not possible because town won't be random while they have something to protect like a king, BUT the closest thing to it (which is the optimal town strategy) is the best way to go. Its not the only way to go, but its the better way to go vs crazy WIFOM, especially in CITP and AITP

"Only you can protect Forest Fires" is practically an AITP game, but as its got a HERO role that can't vote, multiple assassins that can't vote, 1 king that can't vote, and night kills, in theory the same optimal town strategy should be ideal. But the other roles and night kills made it hard to organize, and it can be interrupted by the scum. So an "In The Palace" setup that has more than just 1 king and 1 assassin is a better setup against provable randomness. This same theory also applies to WITP where I imagine its impossible to use these strategies as there is no town. There are multiple factions, but no true town. WITP is likely the furthest from mafia than all the others.

What does this have to do with provable randomness and actual Mafia?

As long as town has something to protect, I don't think you will be able to organize it. Town wants to preserve its power roles and catch scum the hunting way as it cuts down on loosing power roles and in theory could get the scum without casualty to the town. The town has roles it needs to protect. Through in game discussion votes are given direction. The same can be said about power roles. The more direction you have for power roles, the greater they succeed

What if the town didn't have something to protect, no power roles? If the game was nightless and or mountainous I think a provable random strategy could be easily organized by the town. And I think it would be very powerful for town there because it would be an true battle of who's side can absorb lynches the best. When you mix in power roles and nights, things become a battle of

which roles survive the longest

and not

which faction can survive the longest

. The swingier the game is the less likely you can use provable randomness. And the less likely it will be successful and unaffected by power roles. The more power roles (things you need to protect) effect the randomness, the less successful and controllable the strategy becomes.

[PokerFace]

yep. Game 1 had low discussion and good vote patterns. Game two had horrible discussion and good vote patterns. Game 3 had low discussion and horrible vote patterns. You need the right combo of both in royal mafia because having one just won't cut it. And I would argue the good voting patterns are more important

Discussion hurts the most informed faction (because it reduces their information advantage). In regular Mafia, that's the scum

Yes, that is why we scum hunt and should continue to scum hunt in mafia. It should always improve your odds of winning as town better than provable randomness in normal mafia