Our Survival: a lesson from Population Science

Post Reply
User avatar
Dokimi
Member
Posts: 227
Joined: Mon Sep 19, 2016 03:52
GitHub: DokimiCU

Our Survival: a lesson from Population Science

by Dokimi » Post

I've been reading the latest incarnation of the "Is Minetest too small" debate.

I have training in population modeling... so a troubling thought occurred to me. I think it deserves some more rigorous discussion than the "Too Small" debate usually gets.


The problem:

The risks of being too small.


My aims for this discussion:
  • - Discover if we are in the danger zone

    - Decide what we can do to exit/stay away from the danger zone.


To explain, a crash course in population science!



Every population has two features (among others):
- a minimum viable population size (mvp). Below this you enter a death spiral and crash to extinction.
- random fluctuations in size. e.g. + or - 100 individuals every year.


For small populations this results in a huge danger:

The risk of unpredictable spontaneous extinction.


You feel safe. Then you die.



A graph to illustrate:
Image
This is a population which is too close to it's minimum size (50). Although it seems to be doing fine, it is at risk. A run of bad luck can make it go below mvp. Death spiral. Extinction.
Spoiler
This is an extremely simple model in excel, but shows the basic principles. If you want to try it yourself use this formula:

=IF(B2>50,B2 + RANDBETWEEN(-10,10),0)

That is, the initial population (B2) gets randomly changed up or down by 10 with each time step. If it goes below 50 the population crashes to zero.

Playing with this gives you a very good feel for the effects of randomness... sometimes the population grows, sometimes it instantly crashes, but in all cases the risk remains - unless it escapes the danger zone.

On a related point:
Collaborative projects like ours work by self-organization. As I learnt making my Self-organizing Systems mods, these mechanisms work better with larger numbers (we are a bit like the Eusocial Bot in Ecobots). It's a minimum size thing: too small and the self-organization become dysfunctional... breaks down... potentially goes extinct.


The difficult part:
- mvp is unknown (we might be able to make a good guess of it though)
- Randomness is also unknown. In fact we can never know what the biggest run of bad (or good) luck could be. People forget that surprises are surprising!


The moral of the story:
Step the hell away from that cliff edge!

The smaller we are the bigger the risk.


Questions

1) Have we got the data? How many active participants do we have? What are the trends, the fluctuations etc?

2) Is our self-organization working at our current scale? Would it work better with more people?

3) What might out minimum viable size be?

4) How close to the minimum are we? Are we in the danger zone?

5) What are some potential disasters we face? Would we survive them?

6) What are we going to do to avoid the dangers of being too small?



Let's figure this out!
Attachments
randomextinctiongraph.PNG
randomextinctiongraph.PNG (19.46 KiB) Viewed 207 times

User avatar
Wuzzy
Member
Posts: 4786
Joined: Mon Sep 24, 2012 15:01
GitHub: Wuzzy2
IRC: Wuzzy
In-game: Wuzzy
Contact:

Re: Our Survival: a lesson from Population Science

by Wuzzy » Post

The MVP of Minetest equals 1. As long there is someone caring to volunteer for Minetest or there is someone willing to play it (even if just in singleplayer), Minetest's not dead.
Minetest started as a 1-man project, it originally was even named after its creator, this is strong evidence that Minetest does just fine even if there is only one person on planet Earth who cares about Minetest.

If literally nobody cares about Minetest anymore, it's dead. But even then Minetest could be rediscovered and revived later by someone else.

Your analogy with real living populations sucks. :P
Why do you think you can just take the theory from Population Science and apply it to the Minetest community? That's not exactly the same.

twoelk
Member
Posts: 1482
Joined: Fri Apr 19, 2013 16:19
GitHub: twoelk
IRC: twoelk
In-game: twoelk
Location: northern Germany

Re: Our Survival: a lesson from Population Science

by twoelk » Post

one difference to real life organisms is that software projects may indeed be dead for a while and yet can be revived if interest sparcs up again. the far bigger dangers for software are rather an advance in technology that might leave not updated or maintained projects so far behind that they become irrelevant or won't work on the current work-environment anymore (winkwink at LuaJIT).

an interresting real life example of a chaotice population may be the sheep on Soay a remote island of the west coast of Scotland. this example shows that a real equilibrium may never be reached in a population.

User avatar
voxelproof
Member
Posts: 1087
Joined: Sat Aug 05, 2017 08:13
Location: Europe

Re: Our Survival: a lesson from Population Science

by voxelproof » Post

My opinion on the subject is that it's not about Minetest in particular. Probably Minetest is the first prominent instance of a new genre echoing from the stunning success of Minecraft --- I'd propose to call it MC-likes by analogy to roguelikes. I am pretty sure this is to stay here for longer and surely is not endangered by extinction :)
To miss the joy is to miss all. Robert Louis Stevenson

User avatar
Dokimi
Member
Posts: 227
Joined: Mon Sep 19, 2016 03:52
GitHub: DokimiCU

Re: Our Survival: a lesson from Population Science

by Dokimi » Post

The difference between reincarnation and survival:

My apologies for not making myself clear.

It is self-evident that Minetest-the-game can survive even if one person has a copy. Any idiot knows this. I was not seeking data on this point.

What I was looking for was information on Minetest-the-community. Us, you and me, the ability of this population of human beings, and it's associated structure of servers and additional Minetest related content, to survive.

As we are a population of human beings, we are subject to the same laws of nature as any population on planet earth. Births, deaths, immigration, emigration. Even minimum size effects.

Below a certain scale we would lose the ability to maintain servers, old mods, etc. Updates would be less frequent, potentially of lower quality.... the community would lose it's vitality. More people would leave. We would enter a death spiral.

Thankfully software is free from these laws of population dynamics. Perhaps if this community collapses one day, Minetest could indeed be reborn.



Data please!
I personally do not "feel" like Minetest is at risk of imminent collapse. However I lack the data. I saw in the "Is Minetest too Small" debate the question "Too small for what?"

So it struck me, there is at least the possibility that we are too small to be sustainable in the long-run. But without the data I have no way of knowing if this is the case.

Does anyone have the facts about the matter?

User avatar
voxelproof
Member
Posts: 1087
Joined: Sat Aug 05, 2017 08:13
Location: Europe

Re: Our Survival: a lesson from Population Science

by voxelproof » Post

Dokimi wrote: What I was looking for was information on Minetest-the-community. Us, you and me, the ability of this population of human beings, and it's associated structure of servers and additional Minetest related content, to survive.
Well, there're a few overlapping communities within the 'Minetest population' ;) If you mean the part which covers the server users, it is more volatile group and prone to swings in numbers due to the changing fashions. If you take into account active Forum users providing new content and developing new mods and games, I think it's more stable and many if not most of them will continue their interest regardless of gross number of users of Minetest because it's their hobby or even passion. To that part the analogy taken from the population science doesn't apply imho.
To miss the joy is to miss all. Robert Louis Stevenson

User avatar
Dokimi
Member
Posts: 227
Joined: Mon Sep 19, 2016 03:52
GitHub: DokimiCU

Re: Our Survival: a lesson from Population Science

by Dokimi » Post

Its not an analogyyyyyyyyyyy!!!!! :-)

Obviously we are a weird population (or cluster of populations) compared with say, sheep on an island e.g. mostly controlled by immigration/emmigration rather than births/deaths. But still.... you can talk about a population of Jelly Beans and apply science to that (extremely high death rates, very sad!). It makes sense to me that we could we could apply some science to the Minetest community!

Talking about "extinction" I was simplifying.

Obviously 1 copy of the game, or a small committed core could go on. What I mean is practically extinct (remember practically infinite???!!). The point at which Minetest becomes of near zero relevance to anyone beyond those few survivors.

Seeing as no data has turned up, I realized we might get a snap shot of the health of the community by looking at it's age distribution (i.e. the age of accounts).

This data is pretty rough, but interesting (I added up the account ages for 51 members on recent topics on the forums)

Image

It's hard to interpret this data in isolation but... a few possible lessons:

1) We might be undergoing a rapid growth spurt (lots of relatively new members)
2) We might be struggling to retain members (a lack of people sticking around after 3yrs)
3) That committed core (willing to stick around forever) would be on the small size.

Too sum up!

I realize I may have blinded people with science. My point here was to encourage a risk assessment.

This is a pretty normal thing for any organization to do, makes sense we do this kind of thing too!
Attachments
MTagedistribution.PNG
MTagedistribution.PNG (15.73 KiB) Viewed 207 times

User avatar
Wuzzy
Member
Posts: 4786
Joined: Mon Sep 24, 2012 15:01
GitHub: Wuzzy2
IRC: Wuzzy
In-game: Wuzzy
Contact:

Re: Our Survival: a lesson from Population Science

by Wuzzy » Post

If # of Minetest players or devs drops to 0, who cares? Nobody! Because there is no one left to care. There's always the chance of people moving on. There are worse problems. There are other interesting games (and engines) as well.

I don't get this doomsday rhetoric …
Jesus Christ, it's just a game engine.

User avatar
Dokimi
Member
Posts: 227
Joined: Mon Sep 19, 2016 03:52
GitHub: DokimiCU

Re: Our Survival: a lesson from Population Science

by Dokimi » Post

To clear up misconceptions:
1) This is not an analogy
2) This is not prophecy
3) This was not originally even an expression of actual concern.

It was a question!

I have seen two conflicting opinions around here:
1) We need to grow
2) Everything is fine just the way it is.

I am trying to figure out who is right.

The biggest question the "Everything is fine" camp needs to answer is the point raised here. Are we big enough to be sustainable.

However, if as has been suggested, no one actually cares if this projects crashes and burns then... that itself raises some big questions:

What is the point of investing our time and effort here?

Is this place built to fail?

twoelk
Member
Posts: 1482
Joined: Fri Apr 19, 2013 16:19
GitHub: twoelk
IRC: twoelk
In-game: twoelk
Location: northern Germany

Re: Our Survival: a lesson from Population Science

by twoelk » Post

Dokimi wrote:--
I have seen two conflicting opinions around here:
1) We need to grow
2) Everything is fine just the way it is.

I am trying to figure out who is right.
--
Both!

we exist
so something must be right
we have existed for a while
so something must be going right for some time
we are somewhat growing
all the better

if you really are ambitious you might want to study the history of minetest and study the effect of events outside of the community like :
  • Minecraft updates
  • Landmine harasses servers
  • engine-game-split-->-content-moves-into-mods
  • Minecraft gets bought
  • Irrlicht development slows down
  • LuaJit development almost dies
  • Server map databases get too large to be handled well with sqlite
  • RBA dies
  • other individual devs leave or just vanish
  • players on mobile devices appear in masses
  • Illegal clones outnumber official clients on many servers
  • OldCoder gets annoyed and annoys even more
  • GitHub gets bought
  • Laws in Europe change
  • Stuff I forgot
Last edited by twoelk on Fri Sep 14, 2018 19:20, edited 1 time in total.

User avatar
voxelproof
Member
Posts: 1087
Joined: Sat Aug 05, 2017 08:13
Location: Europe

Re: Our Survival: a lesson from Population Science

by voxelproof » Post

Dokimi wrote: Talking about "extinction" I was simplifying.


Hopefully ;-)
Obviously 1 copy of the game, or a small committed core could go on. What I mean is practically extinct (remember practically infinite???!!).


Yes I do! :) However talking about "extinction" you put the weight on the massiveness of the playerbase and I prefer to think about Minetest in terms of the sheer quality - both of the very engine and of the best games which go along with it, and that of the most active i.e. creative part of its users.
The point at which Minetest becomes of near zero relevance to anyone beyond those few survivors.
Like the Noah's Ark ;) For me the number of users really doesn't matter. A few of them with whom I can share the similar interests and the wondrous joy of experiencing beautiful and engaging imaginary realms is perfectly enough.
To miss the joy is to miss all. Robert Louis Stevenson

Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests