WILD LIFE - April 2023

---

Goodness me, what a mad March. TEP's delegate abdicated. A caterpillar evicted wolves from New Dinosaurtopia. And a guinea pig sits over England. Could the world get more surreal? Of course it could! Thanks to Jiangbei, we can now see how a nation would look if it was run by Donald Trump. Wait...

On a more mundane note, new frontiers are on the horizon. And I finally took a more detailed look at General Assembly repeals.

But if you're just thinking, "I don't care about any of this. Why did I get a notification?", then this section is for you.

You've probably already heard about The Incident, but just in case you haven't, here is your opportunity to point and mock. This is the story of how a big "W" became a big "L".

Democratically elected, Altys started their term as WA Delegate of The East Pacific on 3rd March. Eleven and a half days later, they resigned. By accident. The over 550 endorsements they had accumulated since the end of January were wiped out.

The story is that Altys pressed "ctrl+W" on their tablet to close a page. Unfortunately they had an R/D script installed, where the hotkey to resign from the WA is "W". (If it's any consolation Altys, you're not the first person to lose their job due to automation). With over 500 endorsements (now the most in TEP), Vw53aland Herbie hereby became delegate.

Although there has been a campaign to get people to endorse Altys once more, writing as at 31 March, Altys still needs another 50+ endorsements to regain that Delegate title. That will be an uphill struggle, and the gap won't have been closed by the time this has been published. They'll be back as delegate before the end of their term though. Won't they?

That caterpillar. Being a nation themed around a kids' picture book, I can't not shout out this cute dispatch. I would say "Eric Carle, eat your heart out", but I worry that Luca may have eaten it already. (For those who don't know, Eric Carle is the author of The Very Hungry Caterpillar).

The tale begins with Osiris' Sekhmet Legion* invading England at the minor update on 12th March. Quickly backed up by other raiding organisations, soon over 100 players were occupying England. This number rose to 200 over the next few days. Unable to topple this pile, defenders turned to alternate tactics. At first they went with "delegate tipping". And then they went to the Security Council to get a Liberation.

If the raiders could add a secret password, they'd prevent any more defenders getting in. They could take their time about ejecting everyone else from the region. But a Security Council Liberation removes any password, and so keeps pressure on the raiders. It doesn't actually remove the raiders though. Usually at some point after an SC Liberation is passed, raiders get bored and go home. Raiders consider it a win because they leave at a time of their own choosing. Defenders consider it a win because the raiders leave.

So the metaphorical tiger went to the Security Council with their draft Liberation. Which was written as an acrostic of the word "guinea-pig". Really. Check for yourself - the little sunshiney sky icon that hovers over England links to the successfully passed Liberate England resolution. I sh*t you not. There is a guinea-pig sitting on top of England. And literally thousands of people voted for this.

As noted by Valtarre:
"One could defecate onto a scroll and present it to this council, and as long as it had the word "Liberation" on it and targeted an active occupation by adherents to raider unity, it would slide as easily through this body as" dysentery.
(Actually, Valtarre's choice of phrase was "as an arrow through a besieger", but please allow me some artistic license).

The South Pacific's Office of WA Legislation described the resolution as being:
"Creatively written with an acrostic"
I don't think I've used the phrase "cringe" before, but this might be a good moment. If this is what passes for creativity in TSP, I... I dunno. I despair.

A legitimate concern around this Liberation was the timing. In history, the "Liberate The Mystical Council" proposal got stuck behind some other proposals in the queue for the Security Council. By the time the resolution got to vote, the region had already been destroyed and refounded by raiders. In the case of England however, only one proposal (written by the same author as "Liberate England") was going to be at vote before the Liberation, leaving plenty time to save England. However the author decided to pull that first proposal from the queue, allowing the Liberation to go through four days earlier. It wasn't necessary but that was the author's choice. I'm not sure why they made that decision. But I suspect they wanted that resolution through quickly before everyone started questioning why England should be dominated by a guinea pig for the foreseeable future.

* I'm a junior member.

Artificial Intelligence has becoming more and more prevalent in real life over the past few years. With news stories about racist chatbots, art generators trained on stolen intellectual property, and AIs writing essays, it wasn't too surprising to hear that Italy has banned ChatGPT. But at the same moment that real-world nations started putting on the brakes, some NS nations went full throttle on AI.

In a moment of Artificial Genius, Jiangbei created a script that allows an AI to answer NS Issues for a nation. Each nation managed by the AI needs to be associated with a "prompt". The prompt is presented to the AI along with the issue details, and the AI chooses an issue response to match that chosen style. I guess that the challenge for the player is to figure out how to write a prompt that works as expected. The example prompt given is "Who would Donald Trump agree with". I'm left wondering how the virtual nation will measure up vs the USofA!

Jiangbei's thread in the Technical forum has more details - go take a look. And if you're interested in how the nations fare, you can check them out in Nationstates AI.

Frontiers. There's screeds to be written on this topic, but sadly I won't be writing them. I'd love to, but I have a day job, and (somehow) a life outside NS. So here's what I think is important right now. There's a risk I've misunderstood some things (wouldn't be the first time!), but this is my understanding:

First, moderator Sedgistan is hoping that the Frontiers concept will go live on NationStates in April. For those of you wondering what the Frontiers concept is, the first post in this thread will give you an idea. Some of the specifics have changed in the nearly two years since that post, but it's broadly representative of what is to come.

Second, once you know what a Frontier is, you may ask yourself: "Should I found a Frontier?"
No. That's a stupid idea. Unless you and a dozen or more friends have WA'd up with the explicit intent of doing this, don't do it. Founding a Frontier from scratch is the equivalent of creating a region as non-executive founder - you have no power to do anything in your own region. You can't fix the typo in the WFE. You can't add a flag. It's just a bad idea.

Similarly, a region founder may ask themself: "Should I transition my region to a Frontier?". Again the answer is no. You shouldn't. If you haven't already made the decision to transition your region to a frontier, then you've got a hell of a lot of homework to do before making that decision. Go read the thread.

Third. If you're a Founder, your role is going to change slightly. You'll now have two roles - Founder, and Governor. The Founder role will no longer have powers. It will just be a label, showing who clicked the "Found Region" button. All Executive powers will now sit in your Governor role. So you can still do all the same stuff. But you have an extra label. And there's also some new stuff you can do:

Choose to transition to a Frontier (DON'T. See point two above)
Appoint a Successor

Four. There's a new Power, called Successor. It's a way to pass on the Governor role to another nation. From that first post:
Like regions currently are, but the Governor can appoint a Successor, who takes over Executive status should the Governor CTE.
Governors with a live Successor appointed can abdicate their position without needing to wait until they CTE.

That's really it for how they're different. The intention is for these to provide an additional layer of security for regions uninterested in R/D, and to reduce the need for messy refoundings, which erase regional history. There is a slight double-edged sword element to it, as yes, someone could infiltrate a region long-term in order to be appointed Successor.

A Successor is appointed as a Regional Officer position, and only one can be named [AWT - this has changed. There can be a heirarchy of Successors]. As it's an RO position, the chosen Successor can therefore resign the position. A Successor takes over immediately on the Governor ceasing to exist. If the Successor has CTEd before the Governor does, then the Successor does not take over the position.

Abdication is a subsequent addition to this plan, and is necessary because otherwise a Governor wishing to hand over the reins would have to CTE first - we don't want to incentivise that behaviour. At present this is only intended to be possible when a live (non-CTEd) Successor is appointed; if the Governor wants to turn the region into a Governorless (non-executive Governor) one they should consider Frontier status.

Should you appoint a Successor? Maybe.

If you don't appoint a Successor, and you CTE, your region could get raided, just like it can now. But there's a new raider tactic available now. In the past, raiders could either "tag" your region and move on, or stay and eject everyone and then refound the region. Now they also have the option to try and transition your region to a Frontier and then back to a non-Frontier to get Governorship of your region.

If you do appoint a Successor, then if you CTE, you definitely lose control of your region. This is fine if you're bored of NS and never want to play again, but could be frustrating if you change your mind. If you've been away for over four weeks, then your chosen Successor may not want to give you the region back. After all, your absence just demonstrated a lack of commitment to the region.

There's a lot more to it - What if you did transition? What if you're in a founderless region? I'd like to answer that, but I'm not too clear on all the nuances, and am short of time to go finding out.

Which regions will become Frontiers?

Tinhampton has said that Sophia will become a Frontier. Tin is the only person I know of that is going all in. It's a brave move, and I sincerely hope that this is a success.

IIRC, The Grey Wardens, The League, and 10000 Islands all intend to have colony Frontier regions. That way the main region stays safe with a founder, while opening up an avenue of new recruits. I think Europeia will also be going down this route. [edit, 5 April: Comfed has kindly pointed me to this dispatch where Europeia officially declared its intention to become a Frontier region, so also going all in.]

Three out of those five regions are defender regions, and their Frontier colonies will be receiving a steady supply of new players. Potentially this could shift the raider / defender dynamic in favour of defenders - unless more Frontiers are created, 30% of new nations could be founded in defender colonies.

It's good for new players to land in regions that have experience. Still, the influx of new players to these regions might be a bit of a shock to the system. So far, all new recruits to these regions have made an active choice to go to that region rather than just randomly spawning. So the typical attitudes of new players could be different to before. The Feeder regions have had two decades of learning how to deal with undesirable new players. Hopefully the new Frontier regions have been taking lessons!

What will happen to the Feeders?
The feeders will only get 50% of the flow of new nations that they get currently. Long term, you'd expect their populations to half as existing nations move out / CTE, and there is a smaller stream of replacements. But how fast that happens depends on the rate of existing nations moving / CTEing. For example, TNP seems to have a strong retention rate, and so could well have a slower decline than the other feeders.

As the population drops, you'd expect the number of endorsements on the delegate to drop too. Which should make the feeders marginally more prone to coups.

And what about the rest of us?
Many players will probably never notice the difference. Which seems kind of weird, given the huge (but uncertain) impact this will have on Gameplay. It's a new frontier.

After a year of playing Repeal or No Repeal, I finally got around to looking at the statistics behind the game. This didn't lead to the outcome I was hoping for. But here's the story anyway.

I pulled down details of all the General Assembly resolutions via the NS API, then started pushing the data around in a spreadsheet. 193 of the 646 resolutions were repeals, meaning just 453 were original resolutions. So really, 193 of 453 resolutions got repealed. 193 of 453 is 42.6%. That's a high ratio of repeals (though maybe not as high as I expected!).

I ranked each repealed resolution by how long it lasted before being repealed, and drew the first chart below. That's quite a range - some of those repeals happened over fourteen years later! We can see that roughly 30% of resolutions are repealed within three years, and about 23% within a year. When we've been playing the game though, we've been looking at recent legislation - generally less than two months old. The second chart zooms in on that period, looking only at repeals that happened within the first 60 days. That 60 day window still contains about 13% of all resolutions.

One of the quirks I like about the second chart is that it reflects the four day voting cycle of the World Assembly. We can see lots of repeals happened on days 8, 12, 20, 24, 28 and 36 (all multiples of four). Of course, if there were days when nothing was at vote, the four-day cycle got broken, explaining why not all repeals follow that multiple-of-four pattern.

The other thing I wondered was - how does vote share affect repeals? If a resolution passed with only 51% in favour, intuition says it's more likely to be repealed than if 99% were in favour. I drew some more charts to find out for sure.

The first chart below counts how many resolutions fall into each vote-share "bucket". The colours then show how many of those were repealed in 60 days, repealed in over 60 days, or not repealed at all. You can see that very few resolutions get over 90% of the vote, and that none of those have been repealed.

The second chart stretches out those bars to look at each bucket in isolation. It shows what portion of each bucket was repealed in 60 days, repealed in over 60 days, or not repealed at all. It's pretty clear that if a resolution gets a low vote-share, it's more likely to get repealed than if it got a high vote-share. Nice to know that intuition was right!

So how can this information be used in the context of the game? We now know that "No Repeal" is the more likely outcome. So correct guesses of "Repeal" should score more points than correct guesses of "No Repeal". And vote-share should be considered too. Correct guesses of "Repeal" on high vote-share resolutions should score more points than correct guesses of "Repeal" on low vote-share resolutions. But how many points? Here's where the gambling industry could help us out.

From that last chart, we can read the empirical probability of a repeal happening within 60 days. Empirical probabilities aren't the real probabilities, just the evidence from history. Look at that 90-100% of vote share for example. The empirical probability of a repeal is zero. But we know that the real probability must be higher than zero - just because we haven't seen a 90-100%-vote-share-resolution repealed yet doesn't mean that it can never happen. (Note that empirical probability is also called relative frequency, but since we're in a nation simulator, I think "empirical probability" has a nicer ring).

The gambling industry tells us that the payout on a bet should be calculated as:
(Amount bet / percentage probability) - Amount bet.
That should also be a reliable method to figure out scores for the game.
Assuming a $100 bet, and plugging our empirical probabilities into that formula, we'd get payouts as follows:

Vote Share	Empirical Probability of Repeal	"No Repeal" Bet Payout	"Repeal" Bet Payout
50 - 60%	25.4%	$34.04	$293.75
60 - 70%	18.9%	$23.33	$428.57
70 - 80%	10.3%	$11.43	$875.00
80 - 90%	5.4%	$5.71	$1750.00
90 - 100%	0.0%	$0.00	infinite

That last row is problematic. I can't really give out infinite points for a correct guess. That empirical probability of zero is the culprit. If I had a better estimate of the real probability, I could fix that. But my Statistics 101 course didn't give me a better method.

If I got that issue fixed, I still don't think this works well as a scoring system. Although the scores are fair, the payoffs for "Repeal" guesses are just too big. It could take years for "No Repeal" voters to catch up with the score of a "Repeal" voter who got lucky. So I'm going to abandon the plan of awarding points for correct guesses.

That wasn't the outcome I hoped for, but it was a good rabbit-hole to go down. I got some nice charts, and learned some new things. Like "don't take the GA so seriously". Not that I did.

You will have received a notification if you have been mentioned somewhere in this dispatch, and have a tick in the "Dispatch mention" box of your subscriptions page.

Almost all nations mentioned will have:
- done something I felt was worth writing about in an article
- voted in a recent Where The Wild Things Are poll

If you would prefer not to be notified in future, your choices are:
- Stop doing cool stuff worth writing about
- Don't interact with our RMB or polls
- Un-tick in the "Dispatch mention" box of your subscriptions page

But hopefully you enjoyed this dispatch, and don't mind the ping.
If you would like to subscribe, please post the word 'Subscribe' on Where the Wild Things Are RMB, or send me a TG, and I'll take a note to ping you next time.