So for what it's worth I did this calculation - the best fit sphere is about 185 ly radius, centered about 4 ly from Gebel. It's actually an okay fit - the worst outliers are Taranis (21 ly inside the surface) and Thor (16 ly outside). The closest lore system to the center, I believe, is Beta-3 Tucani, 11 ly from the center point. Amusingly that also means Cocijo (the star system) is near the center too. So as I suspected, basically a bubble-sized sphere centered roughly on the center of the bubble.
Analysing the Thargoid Simulation
- Thread starter Ian Doncaster
- Start date
Analyzing stuff based on just the maelstrom coordinates seems too simple to provide any real insight while there's probably a way to get something useful from visualizing the maelstrom + all it's occupied systems with some generated scalars -- direction from the maelstrom, deviation from the average direction of the systems of the maelstrom or something along those lines.
I don't think it's a puzzle as such, but just getting the right visualization to maybe spot a noticeable pattern.
I don't think it's a puzzle as such, but just getting the right visualization to maybe spot a noticeable pattern.
I did this early on, there wasn't much to look at. I suppose you could do the tests again with more info, now.
forums.frontier.co.uk
Analysing the Thargoid Simulation
Then why does it state "TG War signal detection based list of Thargoid controlled systems" ? As noted above - that was a mistake. I'm using a couple of different signals to identify the system war states. To check the allegiance you have to visit the system and see what the ship's HUD...
Yeah I actually agree with that. I did that sphere fit because, once I made the suggestion, I was curious to know the answer, but that's not typically the kind of puzzle solution that ED goes for. Especially in light of recent comments in the livestreams, I tend to suspect there's more to be learned by carefully observing NPC scripts in and around the war zone than from any kind of coordinate numerology.
The farthest current alert is Col 285 Sector JG-O c6-18, 38.71 ly from Thor. But there are control systems further out. The farthest current control system is Ceti Sector DB-X b1-0, 49.49 ly from Leigong
Finally, some data that's not completely ambiguous!
Here's a plot of alerts in inhabited systems versus alerts in uninhabited systems, per week, per maelstrom, ignoring the first few weeks for each when they were doing that weird oscillating thing. (I've added a bit of fractional random fuzzing to the counts so you can get an idea of how many points are clustered together)
As you can see, six of our Maelstroms can be explained by a very simple model: they get 20 points a week to spend on Alerts, and inhabited Alerts cost 4 points each. So they can do 5 inhabited, 4 inhabited+4 uninhabited, 3 In + 8 Un, and so on. None of them have picked "we're feeling peaceful this week, 20 uninhabited" yet, but it seems to be an option at least theoretically available to them.
Thor and Leigong (highlighted with '+' rather than 'x' markers) do not follow this model - they generally (but not always) get fewer total Alerts than would be expected. Possibly their long range alerts are more expensive too, or maybe something else is going on? This needs more investigation.
EDIT: it may just be that Thor and Leigong have not yet exited / only recently exited their initial deployment phase - Leigong's last three weeks have been on the normal pattern, and maybe Thor just hasn't settled at all yet. All "initial deployments" have resulted in fewer Alerts than expected.
Why Leigong and Thor are taking so long to establish themselves I don't have any idea about - it's certainly not because we're excessively resisting them!
Here's a plot of alerts in inhabited systems versus alerts in uninhabited systems, per week, per maelstrom, ignoring the first few weeks for each when they were doing that weird oscillating thing. (I've added a bit of fractional random fuzzing to the counts so you can get an idea of how many points are clustered together)
As you can see, six of our Maelstroms can be explained by a very simple model: they get 20 points a week to spend on Alerts, and inhabited Alerts cost 4 points each. So they can do 5 inhabited, 4 inhabited+4 uninhabited, 3 In + 8 Un, and so on. None of them have picked "we're feeling peaceful this week, 20 uninhabited" yet, but it seems to be an option at least theoretically available to them.
Thor and Leigong (highlighted with '+' rather than 'x' markers) do not follow this model - they generally (but not always) get fewer total Alerts than would be expected. Possibly their long range alerts are more expensive too, or maybe something else is going on? This needs more investigation.
EDIT: it may just be that Thor and Leigong have not yet exited / only recently exited their initial deployment phase - Leigong's last three weeks have been on the normal pattern, and maybe Thor just hasn't settled at all yet. All "initial deployments" have resulted in fewer Alerts than expected.
Why Leigong and Thor are taking so long to establish themselves I don't have any idea about - it's certainly not because we're excessively resisting them!
Last edited:
What's more, it went for Ceti Sector DB-X b1-2 (47,22 ly) and Ceti Sector DB-X b1-3 (49.59 ly) this week. Not sure if it's pushing in that direction specifically or if there's a reason.
For more complicated mechanics we might have to consider adjacency bonuses like things near multiple populated systems would be more costly to invade while invading stuff near tharg controlled systems gets cheaper. Even then it wouldn't be a clear computation because the easiest thing to do then would be to just expand away from the bubble and what counts as "near" would have to be a guess.
Taking this brilliant work of Ian and thinking in a powerplay analogy, what has been surprising to me is that they just cap out at 20 "thargoid command capital", or neither maelstrom has expanded enough yet to gain an additional point - biggest current maelstrom being Cocijo with 154 systems. Though it might be they still want to fine tune the war simulation better before "releasing" this variable. To make it really accurate the thargoid pushback in bars would also need to cost points.
Thor and Leigong are both starved for nearby systems. Thor in particular has nothing nearby. There were 23 systems within 10.02 ly of a control system this last tick, only one of which was populated. Liegong has only recently started getting into comparable numbers to the others. There is likely something about the in-range systems that potentially reduces the number of alerts to put out, either because some systems are just bad, or it has a more direct correlation. Still, I'm not seeing any clean results we'd be able to use to actually predict things, and it may eventually become moot once all maelstroms become big enough to spit out their 20 points of alerts.
What's the current best data source for a list of thargoid controlled systems? Historical data might be neat too to do visualizations.
https://dcoh.watch/systems should be solid. You can also use my dataset. It's pretty manual and could possibly have errors, but it's probably the easiest way to see historical data without writing something to query the DCoH API, and I believe Ian was using it for the work above (as he pointed out one of the errors). https://docs.google.com/spreadsheets/d/1wAJA8ORpH8wxIa4ryD9shPVJgkaRhgZR/
If you want visualisations over time, DCoH also has this: https://dcoh.watch/edmap/historical/2023-03-02
It's very hard to explain the zero-alert weeks which most but not all maelstroms had in the early stages through either lack of suitable targets or extra costs, though.
That could also explain the larger spread of the lower two maelstroms, if there's a total system count that each is aiming for.
It's very hard to explain the zero-alert weeks which most but not all maelstroms had in the early stages through either lack of suitable targets or extra costs, though.
Yup. I've been looking at all the variables I can think of (number of systems in range, number of NEW systems in range, number of systems in range scaled by point value, number of control systems), there is nothing particularly clean. I wonder if there's a global alert count as well, and the systems with the least nearby systems get a lesser cut of the point total.
Thor being at 22 uninhabited, 1 inhabited should still have been able to use either 20:0 or 16:1 expansion patterns.
I wonder if there's a rule that as well as being within 10 LY of a control system, all Alerts must be generated by a different Control? Or even "any Control can only generate an Alert once every two weeks" which would explain some of the early oscillations and why Leigong (which started out with just 11 Controls) and Thor (which is in very sparse space) took so much longer to settle?
It would fit Leigong's early behaviour:
- starts with 11 controls
- puts down 11 alerts, uses each control once
- next week, places no alerts as all its controls are recovering
- week 3, places 17 alerts from 25 controls (on a standard 16:1 pattern)
- next week, only places 2 alerts (possibly some of its unused inner controls from last week had no targets left by now, stopping it going to 8)
and explain why Leigong and Thor in sparser space take longer to settle as they need to expand further to get a perimeter where they have 20+ edge controls which can be used in alternate weeks without crowding.
Does that explain Thor's slowness, perhaps?
If it's only able to expand from edge control systems, would that mean it's certainly able to expand from systems where there's border AXCZs?
So figuring out the conditions for those to spawn in uninhabited systems might help too.
So figuring out the conditions for those to spawn in uninhabited systems might help too.
I had naturally assumed this to be the case - whenever I've looked at the map, I've never seen two (or more) alerts next to only one control system. Never bothered to study the numbers to confirm it though.
Under the proposed system, what would happen if a control system that has launched an alert/invasion is taken back?
I wouldn't expect anything to, certainly not in general - though it might be interesting to test on a Control which is the only link back to the Maelstrom for its Alert. The one right on the end of Leigong might be a reasonable candidate and far enough out to be practical to do with a large enough group.