One Hour One Life Forums

Well, that'll tell you whether each server's population is growing and shrinking as intended. But what effect that has on the common impression that "My family keeps dying because the babies dry up" will be harder to gauge, since that's essentially anecdotal at this point.

The necessary data is in the birth and death logs, but I don't think anyone has yet done an analysis that could answer those sorts of questions objectively and definitively. thundersen is doing amazing work with that data (see his visualization of the birth locations) so it'll probably get teased out before long.

Regarding competition:

You can create inter-server competition by not seeding the server selection. With seeding, assuming a constant global population, a server's birth rate will be equal to its death rate; players who die on ServerX will be reborn on ServerX. But without seeding, all servers will have the same birth rate; each will get an even share of the global death rate. But each server will still have its own death rate. Servers that do a better job of keeping their population alive will see their population rise, and the others will fall. The same will be true of families within an individual server, with the added skewness of the yum and heat mechanisms.

Given that your birth parentage (who your mother is and where she is located) within a server is supposed to be random and independent, i.e. it's not expected that you'll keep getting reborn to the same family if there are multiple families on the server, I don't think there's any reason to make births "sticky" across servers.

The reconnect functionality is very good; perhaps there's a way it could be preserved during a life without creating sticky sessions between lives. Maybe the reflector could query the "sticky" server and find out if the player is alive? If so reconnect, if not re-randomize for a new connection.

Regarding thin populations:

How a town's population feels has a lot less to do with its population and much more to do with its birth rate. Basically, towns are either calm, booming, or dying. Absolute size means very little: a four-person Eve settlement can feel "full", a twelve-person end-stage town can feel "empty". But everyone can tell when the babies are coming fast and furious, and everyone gets depressed and miserable when the baby stream has dried up.

With the current reflector, every time a server is activated, that server has a boom and its upstream neighbor has a bust. And every time a server is de-activated, that server has a bust (and is exterminated) and its upstream neighbor has a boom. This is not a good dynamic.

If you spread the connections evenly across all the active servers, and allow them all to evenly share the global population increases and declines, then each one will feel roughly stable rather than having booms or busts. This is what Server1 and Server2 feel like almost all the time now, since both of them are usually held constant at their soft cap of 50% and therefore have neither booms nor busts.

And regarding competition, again:

There will always be at least two or three lineages on a server thanks to lineage bans (if the lifespans are long) or birth cooldowns (if the lifespans are short). So there will be at least two or three camps / settlements / towns competing for the incoming baby streams even within a single server, no matter how low the population gets.

Here's a design that might work:

On the reflector, keep state for "number of active servers" and "last reported population of server[x]". When a new connection arrives, add up the last reported population for all the active servers and divide by the number of active servers; if it's above some high threshold increase the number of active servers, and if it's below some low threshold decrease the number of active servers.

Then send the player to a random active server (seeded by email / twin code). Query that server for its current population and update the last-reported-population state for that server.

This will (roughly) evenly distribute the players across all servers while ensuring that there's (roughly) no more than X and no less than Y players per server. And it doesn't require querying every server on every connection - just one.

I would make the lower threshold very low, so that even when population is falling most of the servers still remain active. Deactivating a server sterilizes all the lineages on that server and is very depressing for anyone who happens to be on it at the time.

If the global population and active server count is close to an inflection point (either approaching or just past), send twins to Server1 instead of randomly, since there's a chance that the players might have connected at moments when there were a different number of active servers, which would cause the random seeding to send them to different servers.

[ General gripe ] I can count on one hand the number of people who understand and appreciate what Jason is trying to do with the Apocalypse.

Anyway:

Excellent analysis. Hopefully Jason will figure out something workable and worthwhile eventually.

This absolutely should be true. I like the aesthetics of the current Apocalypse, but those aesthetics demand that ending the Apocalypse requires the same kind of sacrifice that starting it did.

preach it brother

I'll stay in bad camps just to help the other players have a good experience; besides, the challenge can be fun. Occasionally I'll give up on camps which are hopeless to the point of pure frustration.

THIS IS AMAZING

I'm guessing that even at the zoomed-in scale each "dot" represents a multitude of births all close to a single camp? The width of each dot looks like it's maybe 100, which is a huge area at the individual player level. So each dot is an Eve camp, and the fatter dots are where the camp lasted long enough to spawn some outposts? Although I'd expect most camps to have a few births at some remove from the camp, as some of the women give birth while they are out gathering supplies.

What does it look like zoomed WAAAY in?

In the meantime, I encourage everyone to "CURSE ELLA MOIST" as soon as you have a token free. Even if Jason is okay with bots, we the players should send them to Donkey Town immediately and permanently.

Beside the general annoyance of repeating stock phrases, and the economic drain of eating food while doing nothing productive, the bot also was opportunistically meddlesome. She picks up items (such as carts) while you are using them, and will apparently attempt to kill people when a knife is available.

Please curse Ella Moist.

Can you zoom in on the central spirals to show them with clearer detail?

Greep forgot to provide instructions on how to use his One City Server, so here they are: https://onehouronelife.com/forums/viewtopic.php?id=4868

In addition to:

• Tea's suggestion to use a low-population official server

• Greep's suggestion to use his One City Server

• lionon's suggestion to use his practice server

I'll give you my suggestion to run your own private local server. Instructions: https://onehouronelife.com/forums/viewt … 243#p41243

So many options! Have fun!

The man in the straw hat gave the correct answer.