The Dynamic Gamemode

[Skull132]

AKA: Democracy. The Nuclear Option.

AKA: Skull has 1 week break from school and pretty much all responsibilities and so he's decided to start coding again.

Assorted PR: https://github.com/Aurorastation/Aurora.3/pull/11008

Okay so, let's take stock. Our current voting system is "Winner takes all", which means that there is very little point in voting anything other than extended and secret, unless you manage to get on a voting spree. This is not great, and many, many suggestion threads have been made, attempting to propose a fix. Though staff have resolutely not touched any of these, as trying to over complicate game mode selection is a bad idea and we'd just have a thread asking for another voting implementation next time around (see: how the transfer vote limit got jerked around a lot).

Instead, let's remove game modes. Aurora's backend allows us to compose game modes on the fly. So let's do that instead.

Thus, the players will now vote for 2 things, back to back.

Intensity. This will determine just how "Intense" of a round the players want. Do they want a lot of action, or a lot of action. The scale is from 0 to 3, with the lowest being "Extended" -- guaranteed no antagonists. We did originally discuss how to get by without this, but figured that the playerbase would be in quite the uproar if we didn't include a clear cut option. I also couldn't figure out a good way to make the antagonist selection not maximize the intensity output, so it works out. The other options, 1 - 3, dictate what antagonist types (not game modes) are up for grabs. Each antagonist type also has their own intensity level: 1 - 3, and so for every round, the antagonist types selected will have a summary intensity equal to or less than (in case of no other choices) to the round's intensity.

For example, a round with intensity 2 can have: 2x antag types with 1 intensity, or 1x antag type with 2 intensity.

This system is a bit simple for now, and I'm not fully happy with it, but it works.

Antag voting. Once intensity has been decided, players will be allowed to vote for an antag. This is not winner takes all. Instead, the antags that get selected will be randomized. The list is weighted, so that the antag that was voted for more will have a higher chance of being used, but that antag's presence is not guaranteed. Though, I think how the function will play out numerics wise, there will be some relative threshold past which it's impossible to not get picked? We'll see. This should demolish most "voting strategies", while also making your vote more useful. Every vote given to your favourite antag will increase the likelihood of their presence in the round: you no longer have to pick between two evils.

Did I say that the "game mode" will be dynamically composed? You could in theory have shit like ninjas and traitors and wizards. And it'll be very interesting. If not horribly broken. We'll see!

[Scheveningen]

This seems neat.

[WickedCybs]

Storytellers and such are a concept I've always loved, one of the main things I've enjoyed on a server like Eris. Will be interested in seeing how this works out.

[Kintsugi]

I suppose fundamentally I don’t care about whether or not this system is implemented. It isn’t better or worse than what we current have - what I do care about is extended and wizard. For one, doesn’t this mean secret extended is going out the window? Additionally, I hope this doesn’t impact the frequency of extended in any way.

 

anyway, remove wizard while you’re at it. 

[Carver]

I'm somewhat surprised the system goes for "amount of votes that specific antag receiving vs the total number of antag votes issued" compared to just looking at the roles players have toggled on, a dynamic system made me assume it'd just intermix based on both intensity and what the antag players have enabled in their prefs (aka if you had an intensity that enabled traitors+lings+vampires and you only had people toggling for traitor, you'd only get traitors and never lings or vampires). Such a version would have the curious effect of also showing what antag types people like to play as the most.

[Skull132]

Predicting who can join as what antag is complicated if not impossible. Going off of the toggles would also introduce a lot of noise, and the selection process would likely steer towards completely random within a closed set. So. Not good.

Also. Seeing what antags people have toggled on is already possible by just inspecting the database. So the curiosity is irrelevant - - the information is already there, available to us.

[Carver]

A shame, though I'd have enjoyed the random chaotic jank (and moreso the lack of knowledge when the antag type isn't dictated by known vote). Unorthodox combinations of multiple roles in a round is generally more fun to me than knowing the roundtype by the first 1 or 2 antag reveals - so I can only hope this system works out with a lot more mix+matching than the very limited present combo-modes.

Out of a personal curiosity for the toggles, has that particular set of data ever been released?

[Skull132]

4 minutes ago, Carver said:

Out of a personal curiosity for the toggles, has that particular set of data ever been released?

No.

[Roostercat]

While this does not remove extended, it DOES remove secret extended, meaning people are always going to expect antags (and even a suspected list of which at that) and char development and relaxed roleplay is going to be damn near impossible as everyone is waiting for the bad men to appear. Not to mention normal extended needs to supercede all the other votes which is very unlikely to happen often with the current crowd. And thus I really don't like this.

[Skull132]

6 minutes ago, Roostercat said:

While this does not remove extended, it DOES remove secret extended, meaning people are always going to expect antags (and even a suspected list of which at that) and char development and relaxed roleplay is going to be damn near impossible as everyone is waiting for the bad men to appear. Not to mention normal extended needs to supercede all the other votes which is very unlikely to happen often with the current crowd. And thus I really don't like this.

Could be viewed as a positive. The idea that only extended permits meaningful character development is a questionable trend that was acquired at some playerbase turnover. Low intensity rounds should still permit rp while keeping some sort of interesting shit happening.

The super majority requirement will be removed for now. But I'll trail back to it a month or two down the line, once we have data on how people vote without it.

And in the future, the selection of pure antags will be complimented by other event roles. So it isn't all about making 2d sprite go horizontal. Though this plan is a long one.

[Roostercat]

Sounds good to me, overall worth a try I suppose.

[Kintsugi]

I have re-evaluated what I think in regards to this, and I am going to say I tentatively believe it to be better than the status-quo. Secret extended being removed is... Alright, I guess. To be frank, secret extended sometimes felt like the worst of both worlds - the paranoia of secret and the lack of action of extended. True extended is great - but part of why it is so great is the assurance that your round will not suddenly be derailed. Generally speaking, I do strongly feel that extended is the only gamemode where you can reliably, and consistently, develop your characters and their relationships with other people.

 

Moving on to why I think this is a good change - because it gives the player more choice in regards to gamemodes. Previously, unpopular gamemodes could be quite common - why? The secret rotation arbitrarily chose gamemodes without regard for how the players actually voted. With this system, antagonists that are actually popular will more consistently be picked - and antagonists that generally are not popular will not. With that in mind, and with the assurance that there is no plan to remove extended or marginalize it, I think we're better off using this.

 

In the long term, I do think we should focus on a stronger storytelling system, where the emphasis is placed on event roles and third party ships, and less on pure antagonists - this could be a good forbearer of that concept.

[Lmwevil]

Storyteller system ideas can be yoinked from open source Eristation code and probably fits well

 

given I even suggested dynamic in the other thread, i think this is a brilliant idea

[Skull132]

Current state of affairs:

• Super-majority requirement from extended removed from extended. We'll see how the voting habits play our over a month or so.
• Antag selection now works.
• Finally selected antags will not be announced.

The TM will remain active for a day or two to spot issues if all goes well.

[NerdyVampire]

I was almost about to make a similar (but not as sophisticated) suggestion to this but searched first and found this active post. I really like the idea of something more dynamic being implemented and can't wait to see it in action. Full support and love from me ❤️

Edited January 22, 2021 by NerdyVampire
Clarification

[Skull132]

I'm also not too happy with how the rounds are currently composed.

The system of 0 - 3 means that there is never, ever a chance of 3 + 1 (merc + traitor, for example) happening. However, the algorithm might be a bit too merrily optimizing towards maximum insanity, so upping the intensities by 1 doesn't sound like a good idea either: they'll get over utilized.

In the grand scheme of things, if we end up adding more "non-hostile" event type antags, it may buff out. But a temporary fix would be to add a 25% of a 1 intensity antag being added to any dynamic picking resolution which only contains 1 gamemode if intensity >= 2. This would lead to compliments of large gamemodes (like merc or cult) alongside a "supporting cast" of traitors, for example. Which should be pretty neat.

Also. Revenant and other gamemodes that require processing will not work for now. LMAO.

[Kintsugi]

50 minutes ago, Skull132 said:

-snip-

I think you need to be careful with having too many antagonists show up. Changelings and mercs and traitors? That's half a calamity round. Too many cooks spoil the soup.

[Skull132]

5 minutes ago, DanseMacabre said:

Changelings and mercs and traitors?

ye except even with my proposed system, this can never happen.

[Carver]

I'm very curious to see how conts/fellows get reintroduced under this system, whether you'll tie them together or separate them wholesale.

[niennab]

I could be wrong but I believe when its voted extended, the 'check round info' button simply says that the round is dynamic. If that is the case, could it be changed? I'd hate to not know it was extended only because I missed the initial vote and round start.

[Skull132]

3 hours ago, niennab said:

I could be wrong but I believe when its voted extended, the 'check round info' button simply says that the round is dynamic. If that is the case, could it be changed? I'd hate to not know it was extended only because I missed the initial vote and round start.

I'll be making the winning intensity level show on round status in all cases.

[WickedCybs]

Just feedback I also posted on the discord. Like how things are so far, dynamic particularly works well for antags like Loner as they can be thrown into something like wizard without upsetting the balance of the round badly.

[Butterrobber202]

I am really enjoying this new voting sytem alot. It's functional and instead of set mixes, we get a pretty random handbag, so all in all, big fan! 

[StationCrab]

The random mix has lead to some really interesting scenarios.

[Scheveningen]

I've enjoyed this, though I have a question. Is there greater weight in favor of something like if mercenary gets voted by 7 people vs. when borer gets 1 vote, or is the vote actually equal? I've seen dominant mercenary votes for dynamic yet did not see mercenaries spawn through when the round started, is there an explanation for that - or is it just really bad luck?