Consider uu distance for HW placement at universe generation

[swaq]

HW distribution definitely seems uneven. The closest pair is about 150 uu apart (though still 8 jumps) while another HW has at least 410 uu separation from the nearest enemy HW. And the highest separation between neighbors is nearly 600 uu.

[Oberlus]

HW distribution definitely seems uneven. The closest pair is about 150 uu apart (though still 8 jumps) while another HW has at least 410 uu separation from the nearest enemy HW. And the highest separation between neighbors is nearly 600 uu.

The 150 uu distance is only secondarily relevant (it allows to see your neighbour sooner), the one that really matters is the number of jumps and the distance that those starlanes would take.

Allocation of HWs is done after systems and starlanes have been deployed, in this function (called from here). Do you know python and some network analysis to make it more even?

[swaq]

HW distribution definitely seems uneven. The closest pair is about 150 uu apart (though still 8 jumps) while another HW has at least 410 uu separation from the nearest enemy HW. And the highest separation between neighbors is nearly 600 uu.

The 150 uu distance is only secondarily relevant (it allows to see your neighbour sooner), the one that really matters is the number of jumps and the distance that those starlanes would take.

Allocation of HWs is done after systems and starlanes have been deployed, in this function (called from here). Do you know python and some network analysis to make it more even?

The distance of the starlanes will usually be shorter if the HWs are closer. Which would mean fewer turns before clashing with the neighbor. Probably not a huge issue but could be annoying in certain circumstances. A side effect of the current algorithm is that HWs will naturally be closer where systems are more dense (shorter starlanes). This could also be beneficial on some layouts as it would keep one player from monopolizing a ton of systems in one particularly dense area. So there isn't really one correct way of doing things.

I haven't done any serious coding in python but I am familiar with it. It looks like the selection is done here and it prioritizes maximum number of planets within 3 jumps of the HW system. I could imagine adding some criteria for minimum uu separation as well either there or in the calling function. I don't have the time to really look into it right now though, unfortunately. There are probably higher priority issues for making games "fair."

[Oberlus]

I've split the topic for easier reference in the future since I think it is relevant (although not pressing right now).