64 Player Map Design


#41

Any progress on a Toral Map?


#42

I really think you should give a new design a go for a while Jay.
Not only is the current one getting just a tad old, but it is incredibly hard on players in the middle. I speak from experience.


Rainbow Trade Aliance
#43

Yep, we have a plan for a new map. I just need some time to load it up and test it. Perhaps later this week?


#44

That’s great, Jay!


#46

Sorry, I got totally swamped last week and this week is not looking super good either.


#47

Any plans for a new map? It’s been almost a year with the same map and it would be interesting to change it up every now and then so that 64p games get more interesting


#48

I’m taking the new game Blight to Pax Australia in three weeks and will be really rushed until then, but once that is out of the way I’m going to spend a little time on NP and I agree, a new map is needed!


#49

Honestly, people have pretty much figured out this map; every game is pretty much the same. There are basically a few snowballing alliances, and towards the end players from each of the alliances form two giant coalitions that fight it out.

Hopefully you’ll have time to work on this soon.


#50

@xjhdexter what map should we run next? I will set it up next week!


#51

I really do not know . . .

1 A
Jay said that Three Neighbors map 64L was too wide

1 B
That is why he choose Three Neighbors map 64D, which is the quad_flower being played now.

2
My next suggestion could be Three Neighbors map 64E, which maybe looks like a “pretzel”.

This is a variation of map 64D

3
I was slowly working on this spiral

Which I think is an unfair map.

I have made slow to little progress with spirals.

4
The three neighbors maps are based on the concept of fair distances ( 1D ) between neighbors.

I had another concept, which is based on fair defensive areas. ( 2D )
So if two neighbors are a little closer, then a third neighbor should be farther away. So there would be some push and pull between neighbor distances that would be balanced by fair defensive areas.
If there is some shrinkage in one area towards one neighbor, then there should be some expansion in other areas towards another neighbor. The hope is that with fair areas, there is a fair opportunity to scatter random stars.

Just thinking out loud, but I have made little progress with the math.
I have thought about structures found in nature. Spiral galaxies, trees and their branches, rivers meandering down mountains, joining branches in a watershed, etc. . . There often seems to be two or more forces at work to create the curves and branches.

I like maps to be fair. This is difficult to do.

The standard 8 player maps are random hex maps. Randomness is a different kind of unexpected fairness.


#52

What are the maximum dimensions?


#53

It is possible that I mis-spoke here ?

I think Jay’s opinion was that on map 64L, most of the players were in the outer loops, and had less opportunity to interact with other players from other parts of the galaxy. Map 64D allowed more player interactions in the center, where there are four branches joining together instead of just three.

Yes, that is what he said. EDIT


#54

In my opinion, 64L is better because it is more different from 64D, which has been in use for a long time.

I think that three branches in the center is actually good; when I am part of an alliance in the center, we get cramped due to two neighbors.

Of course, this is just my opinion; we won’t know until the maps are actually tried out.


#55

hmmm . . . still needs more work ? or maybe . . . .IDK ?

some empires have two neighbors, some have nearly five . . . :unamused:

I could attempt to apply the fair defense area effect, and see what happens . . . ?


#56

One problem with the equal neighbour maps is they are often not equal at the start because the random layout can completely isolate people or only give a player one neighbour.


#57

Do you think we should come up with a new random map generator that works better for 64 players. (The current algorithm just makes a big blob)

We could divide the galaxy into a grid and populate each section with 4 player starts. This grid is 4x4 then further divided 4x4 but you might want to make the smaller grid 3x3 so that there are only 5 empty cells for every 4 player starts.

Do you think it would make a map with interesting gaps?


#58

I’m not exactly sure what you mean with the above grid. I’ll come back when I’ve slept and hope for clarity.

A cluster based map generator with knowledge about jump distance would be nice.

First generate the home stars. Then ensure there is a X hyperspace range walk from every home star to every neighbouring home star, where X is some value and neighbour is decided based on the above equal neighbour discussion. Then ensure every home star has the same number of stars that it is the shortest path from (that it can get to faster than other empires, taking into account HR limits). After this assign resources so each cluster has the same resource spread.

The above method will leave you with clusters which are fair. If everyone plays perfectly during the exploration phase they get the same star count, resources and neighbours. The differences are in a varied layout and players.

Another idea, though not relevant to 64 player maps, is to create a random spread then duplicate in a circular layout. You would need to rotate the layout each time you tile it so it has rotational symmetry.


#59

So is the map going to change any time soon?


#60

I don’t want to promise anything, but now that Pax is out of the way I have set aside a full day every week to NP.

Last Thursday I set up a Node server and started working on a new javaScript server that will generate new 64 player maps.

Things are slow because I had to evaluate a bunch of different hosting options. I chose Heroku because they have a pricing structure I can understand (unlike Google, Amazon, and Microsoft) and their front end design is so nice and polished I expect it reflects their back end and documentation.

Its also slow because I’m learning Node and and there are a few new concepts I need to get my head around.

It might take a while to find a function that makes the scattering “interesting” and also fair. I want the map to have areas where there are no stars. There will be players with more neighbors that others. I want the new maps to be a lot more random to make every game more different.

Also worth noting, now that this part of map generation will be done on a remote server, there will be an extension to the existing mag generator that knows how to read custom maps with every single star specified.

You can follow development here.

https://github.com/jaykyburz/tritonsnode


#61

@JayKyburz would it be possible to use an RNG to cycle through all three 64 player maps ? Custom Galaxy Three Neighbors map 64L xjhdexters-hex_loops , map 64D xjhdexters-quad_flower , map 64E xjhdexters-pretzel ?