Jump to content

Round Voting Rework


Nikov

Recommended Posts

Say we have a round type vote.


2 vote Changeling.

2 vote Vampire.

2 vote Traitor.

3 vote Extended.


It is clear that six of the nine players in this vote want some sort of antagonist on the station, but the voting majority, itself a minority, wins the vote. This is because our current voting system is winner-take-all for the largest individual selection. I consider this less than idea.


Consider an alternative. In this, the voting scripts add the total number of votes (X), randomly selects a number 1 through X, and selects the game mode based on the number.


So given the above votes we have...


1-2 are Changeling.

3-4 are Vampire.

5-6 are Traitor.

7-9 are Extended.


Total is 9. The script chooses a number 1-9, compares it to the table, and then begins a round of Secret of that game type. In this way we can all vote for our favorite game mode, instead of our most likely to win preference, and the dice will determine which mode. Votes simply stack the odds for player preferences.

Link to comment

That's actually pretty good, but...


One thing which always bugged me was not being able to know when it was extended and when it wasn't. And by association, other gamemodes such as cult, nuke, rp-rev, etc. Some characters just fit certain gamemodes better than others and I like that we're capable of having a heads-up of what we're heading into.


For example, playing janitor through extended is a maximum pain, but playing janitor through rp-rev can be very funny.


I'd just hate to lose that :(

Link to comment
That's actually pretty good, but...


One thing which always bugged me was not being able to know when it was extended and when it wasn't. And by association, other gamemodes such as cult, nuke, rp-rev, etc. Some characters just fit certain gamemodes better than others and I like that we're capable of having a heads-up of what we're heading into.


For example, playing janitor through extended is a maximum pain, but playing janitor through rp-rev can be very funny.


I'd just hate to lose that :(

 

We could just have it show the round type that wins according to this proposed system. I can understand not wanting to play certain characters for certain round types and I have always really liked a system like this in similar games. Gives that one person voting voting wizard or ninja a chance.

Link to comment
We could just have it show the round type that wins according to this proposed system. I can understand not wanting to play certain characters for certain round types and I have always really liked a system like this in similar games. Gives that one person voting voting wizard or ninja a chance.

That seems like a really good solution!

Link to comment

This is a pretty good idea, however I'd like to suggest a couple of refinements:


First, I'd suggest that any time the majority of the total votes goes for a single round type then that round type should win outright and be announced as such. But, because it's more fun (usually) not to know, when the vote is mixed it should come up as secret and display the possible round types and the odds of it being each type so that players can still make informed choices while retaining an element of surprise.


Second, when the round-type is this new style of secret and the round fails to start, it should automatically pick a different one from the options. That way we don't have to wait 6 minutes or whatever just 'cause some dipshit voted for mutiny during dead hour when there's no chance of it actually happening.

Link to comment
I like Lady's proposed modifications with one caveat. If the vote goes to extended, it should say it's extended.


I like extended, but secret extended is very different then voted extended and also it is the worst, by the hour and a half mark, 75% of the crew have gone SSD.

 

A valid point. I see nothing to object to here.

Link to comment

I don't like the idea of a 51% 'tyranny of the majority'. Perhaps a 66% supermajority could force a game mode, but a simple 51% still means 49% don't want it. More often than not, a 51% vote will win its choice anyway. I must therefore disagree with the proposed change.

Link to comment
I don't like the idea of a 51% 'tyranny of the majority'. Perhaps a 66% supermajority could force a game mode, but a simple 51% still means 49% don't want it. More often than not, a 51% vote will win its choice anyway. I must therefore disagree with the proposed change.

 

I'd be fine with requiring a 66% to force a given game mode, but I don't think it'd be as much of a problem as you seem to. The reason we end up with almost all of the votes split between two (or maybe three) game modes is because our current system encourage players to switch their vote from the mode they'd like to one which maybe their second or third choice but at least has a chance of winning.

Link to comment

So 66% forces a game mode, otherwise its the vote-weighted RNG. I would like the odds announced after the game mode so people can select their characters appropriately, but am still against declaring the game mode with the exception of the 66% push. I think it is important for balance purposes that teams aren't stacked against the antagonist, e.g. a nuclear round filling all Security slots with only one oblivious engineer. The other problem is that it does constitute metagaming to select a character based on the round type. This is just a statement of fact, not a condemnation.

Link to comment

I am very much against using the RNG (specifically, BYOND's RNG) as it's not exactly unbiased. Sure, it's a good idea we've got here, but it's not actually random - to the point where I can get 15 "heads" in a row on a coin flip. I'd much prefer the

if we can actually script it in. In short, you can freely pick your favorite round type without worrying that the extended peasants will win just because you want to vote wizard, because you voted for nuke as your second pick and secret as your third.
Link to comment

I'm not sure what to say to that other than "fix the RNG", and I doubt thats in our power. If we accept the RNG is too bad to use for this voting system, then its too bad to use elsewhere. Yet we do, so the RNG is good enough to use here as well.

Link to comment

Pseudo-random number generation is kinda a big deal in computer science and cryptography, and generating pseudo-random numbers with a reasonable amount of entropy takes time. If you need a bunch of random numbers fast (like for a game), they're going to be shit. That's just the nature of the technology.

Link to comment

Let me specify yet again that we have alternative voting methods without leaving even more up to the RNG. It's literally called the alternate vote. Let's say we've got a vote as follows:


We have eight people playing on the server, voting in the order in which they would prefer to play a given round type.


Brage votes for secret, wizard, and extended.

Delta votes for secret, then extended.

Hive votes for secret, nuke, and rp-rev.

Meowy votes for extended, then nuke.

Voltage votes for extended and doesn't want to play any other game mode.

Doomberg votes for nuke, wizard, and extended.

Frances votes for wizard, changeling, and rp-rev.

Alberyk votes for changeling, nuke, and extended.


In the end, the tally for first-pick votes is as follows from highest to lowest:


Secret - 3

Extended - 2

Nuke - 1

Wizard - 1

Changeling - 1


Under our current system, secret would win the vote and we'd be playing that round, even though a majority of the players (5/8) clearly have no interest in playing a secret round.


But, because each person lists alternative votes under this system, we know where they would go if they'd decided to change their vote because their given game mode didn't win, which prevents their vote from simply not counting because they like a game mode that other people do not. So the lowest game modes voted in are nuke, changeling and wizard, so we look back at what the alternative votes for each person are - Doomberg lists nuke and wizard, but both are eliminated, so his vote goes to extended. Frances, on the other hand, is all over the place and voted for not one but three round types that no one wants to play, today. Sorry, Frances. Alberyk lists changeling and nuke, but has extended to fall back on, so his vote goes to extended as well. The end results are much different.


Secret - 3

Extended - 4


By the end of the process, we come up with a scenario that more of the player base will be satisfied with (4/8 of the voters ended up having their votes count towards it, and even some of the secret players listed it as secondary.) We're now playing a game mode that a majority of players actually want to play, rather than allowing a minority of 37.5% dictate the round type for us.

Link to comment

The RNG for byond is not that bad. Any randomizer will throw out sequences because the existence of sequences are an inevitable outcome of randomization.


I was sure I was right, but byond IS pretty weird so I ran a test.


This is my test code:

 

/client/verb/rng_test()
set background = 0
set category = "Testing"
set name = "Test randomizer"

var/list/results = list(0,0,0,0,0,0,0,0,0,0)
for (var/x = 0; x < 1000000; ++x)
	var/index = rand(1,10)
	results[index]+=1
world << "Randomization results:"
for (var/x = 1; x <= 10; ++x)
	world << "[x] = [results[x]]"

 

For those of you who don't speak Byond, I rolled a random number between 1 and 10 a whole lot and recorded which number was which. This is the results after rolling a million times:


Randomization results:

1 = 99290

2 = 100134

3 = 99938

4 = 99614

5 = 100647

6 = 100110

7 = 100267

8 = 100296

9 = 99375

10 = 100329


So yeah. The byond rng is fine.


I'm in favor of a pure weighted die-roll like the OP suggests, without the caveates of 'if most people vote for a thing it goes in automatically'. The point of the die roll is that there's no point to strategic voting. I really like that idea.

Link to comment
  • 2 weeks later...
×
×
  • Create New...