Analysing the Thargoid Simulation

[sathlei]

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 displays. Checking for signals, hyperdictions and/or being able to plot with a fleet carrier are not reliable methods due to randomness/other bugs.

{ SystemAlliance : Thargoid } is one of the few clear attributes for this conflict reported by the game.

Controlled systems are not guaranteed to contain any thargoid war signals. At any point in time there may be some, or there may be zero. Additionally we also know there are hidden/bugged controlled systems that don't show thargoid war signals under any circumstances (around Leigong).

Again, I've made a mistake saying the list was constructed based on TG War signals. Here system allegiance was the only criteria for classifying system as Thargoid controlled.

I'm using TG Was for detecting invaded but still human-controlled systems.

And yes, as mentioned before, the currently FDEV implemented signals/attributes are not adequate for the complete outside-the-game conflict monitoring.

 

[Ian Doncaster]

Noting https://forums.frontier.co.uk/threads/thargoid-war-rebalancing.612479/ it might be a few more weeks before we have a good difficulty model for systems, since the data so far is going to need throwing out.

Some interesting bits on Recovery systems - they seem to be a static 4 weeks (if stations need repairing) regardless of the number of stations, or 1 week (and autocomplete) if all stations are mostly okay because the invasion was defeated in Week 1. It also appears that the Recovery progress bar doesn't reset - which makes some sense, the Thargoids can't reinforce it on their move - so we may need to look at these in a few weeks time to see what happens to incomplete ones. Equally, the difficulty rebalance may let Op IDA finish HIP 23716 early.

This week, Leigong's alerts are visible, but there's none visibile at Oya or Cocijo. The 11 control system discrepancy remains. Most Alerts remain targeted to uninhabited systems by about a 10:1 ratio, and the expansion shell of Alerts continues to be omnidirectional at a glance - lots of uninhabited systems behind the Stargoid are being selected. The total number of Alerts is slightly higher than might have been expected from previous weeks. It looks like the Thargoid "Alert budget" is a global one rather than per-Maelstrom, given the wide variation in Alert count between Maelstroms and between weeks at the same Maelstrom. Next week's Alert count will be very important for determining where the Thargoid's budget comes from:
- if it's from Maelstroms, they'll probably get about 80 again next week
- if it's from Control, they'll probably get about 100 next week.

 

[Ian Doncaster]

For those who like graphics, here's a summary flowchart of the Thargoid State machine. Still a few uncertainties but hopefully some of those will be sorted out this week.

 

[Virkful]

@Ian Doncaster Have you checked how many alert systems were sent for week 4? (My motherboard died so can't login to check )
Wonder if it's related to controlled/invasion system, or a static number to each maelstroms.
(Assuming the "last week" and "this week" on first post at page 1 is referring to week 2 and week 3)

 

[Ian Doncaster]

@Ian Doncaster Have you checked how many alert systems were sent for week 4? (My motherboard died so can't login to check )
Wonder if it's related to controlled/invasion system, or a static number to each maelstroms.
(Assuming the "last week" and "this week" on first post at page 1 is referring to week 2 and week 3)

Here's the progression.

Alert​	Invasion​	Controlled​	Maelstrom​
0​	9​	20​	1​
8​	25​	50​	3​
26​	41​	106​	5​
38​	59​	208​	8​
81​	50​	245​	8​

There's a very good fit for "Control systems in week X determines Alert systems in week X+1" when looking at the war as a whole, but there's basically equally as good a fit for "Number of Maelstroms in week X determines Alert systems in week X+1".

Next week, when the number of Control systems is likely to jump by another 50 or more, but the number of Maelstroms is predicted to stay at 8, might provide a way to distinguish the two theories.

 

[yttrbio]

Watch out for that 26 alerts in week 3 number. If there were 5 invisible alerts around Leigong, that's 31, in case that adjusts your analysis. No gains by Leigong this week, so there were no invisible alerts around it last week.

 

[Ian Doncaster]

Watch out for that 26 alerts in week 3 number. If there were 5 invisible alerts around Leigong, that's 31, in case that adjusts your analysis. No gains by Leigong this week, so there were no invisible alerts around it last week.

Redoing with 31 that week instead, it makes very little difference: both Control-Alert and Maelstrom-Alert correlations get marginally weaker.

 

[yttrbio]

It is probably worth nothing that, of the 158 alerts that have happened, all but 2 have been within 10 ly of a control system the previous tick. The 2 that were further are HIP 20890 (10.013) and HIP 113785 (10.001), both of which could be just rounded down to 10 in whatever algorithm is being used (or maybe the range is slightly higher). I've been hypothesizing that the range of alerts is 10 ly, and it seems to be holding up so far. That gives an maximum on expansion rate of about 5 ly per week if they go in a straight line (which they don't, so it's somewhat lower)

(You can see my dataset here if you want to investigate yourself)

 

[sathlei]

When you de-clutter the current war map, you may observe something interesting.

The eight commanding invasion Thargoid Maelstroms create a form of a bag, partially surrounding the Solar system:

(it's better visible on the 3D model)

Maelstroms are situated at quite similar distance to the Sol - between 136 and 232 ly.

 

[sathlei]

When you de-clutter the current war map, you may observe something interesting.

The eight commanding invasion Thargoid Maelstroms create a form of a bag, partially surrounding the Solar system:

View attachment 339021
(it's better visible on the 3D model)

Maelstroms are situated at quite similar distance to the Sol - between 136 and 232 ly.

This plot is probably a bit clearer:

 

[Ian Doncaster]

Maelstroms are situated at quite similar distance to the Sol - between 136 and 232 ly.

That's quite a range when the bubble is vaguely spherical and they're all on the edge of it, though. Any 8 randomly-selected fringe AWs would probably have that property.

Raijin is interesting in this respect, I think - it started out at a substantial Z>0 coordinate, but rather than aiming for the nearest part of the bubble, "crossed" most of it on the way in to end up joining the other seven.

I feel their positions are significant, but probably not significant with respect to Sol.

 

[HappyMoonMonkey]

I'd put my money on " the simulation ran best and with the most even distribution of frontline throughout" with the maelstroms in these positions.

 

[sathlei]

I feel their positions are significant, but probably not significant with respect to Sol.

Another plot, with Thargoid controlled systems and HIP 22460 added:

An outlier at 411.375, 9.09375, 261.46875 is HIP 62154 which experienced Thargoid incursion a year ago.

Note: Thargoid controlled systems were determined based on the SystemAlliance = Thargoid status in EDDN messages from December.

 

[Ralph Vargr]

I'd put my money on " the simulation ran best and with the most even distribution of frontline throughout" with the maelstroms in these positions.

Where did all those escape pods go...

 

[Felix DiCestria]

This plot is probably a bit clearer:

View attachment 339022

The dish shape is striking though at this point I'm inclined to view it as an artefact of forming around the edge of the Bubble.
There doesn't seem to be any determined drive towards Sol.

 

[sathlei]

The dish shape is striking though at this point I'm inclined to view it as an artefact of forming around the edge of the Bubble.
There doesn't seem to be any determined drive towards Sol.

I haven't seen any Thargoid invasion movement analysis, so I cannot say what is the invasion target.

The early static picture suggests Thargoids may be creating a form of barrier between the Human space and their own older (core?) territories.

But that's a hypothesis

 

[Ian Doncaster]

I haven't seen any Thargoid invasion movement analysis

So far? Roughly spherical - including taking uninhabited systems even further outside the bubble than the Maelstrom's initial control sphere extended.

There's no obvious attempt to expand in any particular direction; equally, depending on various unknowns a bit of initial defensive consolidation before more focused expansion might be a sensible move for them.

 

[yttrbio]

I haven't seen any Thargoid invasion movement analysis, so I cannot say what is the invasion target.

I've been trying to do this numerically, but there is very little consistent in the results. What I've done is generate some sets of eligible systems for alerts and attempted to see what percentile distance the actual alerts fell from a 'zero' point compared to randomly generated sets of systems.

4 different hypothesis tests:
Each week, the average alert moves further from the maelstrom than a randomly-selected set of systems within 10.05 ly of a control system
Each week, the average alert moves further from the average of systems within 10.05 ly of a control system than a random selection of such systems
The average alert around each maelstrom is further from the maelstrom than a randomly-selected system within the range of the furthest alert
The average alert around each maelstrom is further from the average of systems within the range of the furthest alert than a random selection of such systems

The last two are inherently biased in favor of seeing result, because I don't have any other way of figuring out the range.

My results, though they may vary somewhat based on your assumptions. This approach also wouldn't detect if expansion was in opposite directions.

As is, I don't see great evidence, and whether there is or not depends highly on how you compare to a non-directional state.

 

[Kerissimo]

Galnet has hinted that exobiology may play a part in defeating the Thargoids. How?

As for xenobiology helping with fighting the Thargoid. Has anyone checked if there's an atmospheric world in Thargoid controlled space and if this world (preferably with an Ammonia atmosphere) has new biologicals?

I think checking every atmospheric worl in thargoid controlled systems, and especially in maelstrom ones, could be usefull, maybe they are doing something, terraforming?
Maybe we could glearn something interesting, about them.

Problem: probably that would require 2 commanders, 1 pilot, and one taking samples, unmanned ship probably will be destroyed very fast.

 

[Maolagin]

It is probably worth nothing that, of the 158 alerts that have happened, all but 2 have been within 10 ly of a control system the previous tick. The 2 that were further are HIP 20890 (10.013) and HIP 113785 (10.001), both of which could be just rounded down to 10 in whatever algorithm is being used (or maybe the range is slightly higher). I've been hypothesizing that the range of alerts is 10 ly, and it seems to be holding up so far.

Perhaps explained by cubic vs spherical distance metric? The fact that BGS uses cubes rather than Euclidean distance would make me suspect that the war sim does the same, with the same annoying effects on statistics computed on distance.