Analysing the Thargoid Simulation

[Ian Doncaster]

Positing attacks from HIP 8525 or especially Liu Huang is certainly the sort of thing which if it works suggests you're onto something. I had a bit of a look at whether there might be some reason that set of 4 wasn't attacking and for 20 July to 10 August a rule of "closest controls with targets attack the furthest targets they can reach" would also do pretty well; it wouldn't quite fit with this week, though. Still, it gives similar results to my basic guess, but for almost entirely different reasons (especially; which one is the 5th if there's only 4) so I thought that's worth adding to the table too.

Anyone else got any ideas they want to put up?

Null hypothesis	Narwhal hypothesis	INIV hypothesis	Nearest-Control hypothesis
Cephei Sector AV-Y b2	Cephei Sector AV-Y b2	Cephei Sector AV-Y b2	Cephei Sector AV-Y b2
HIP 4041	HIP 4041	HIP 4041	HIP 4041
HIP 3006	HIP 3006	HIP 3006	HIP 3006 (if there's a fifth)
Lahua	Lahua	Lahua	Lahua
Muchihiks	Muchihiks	Muchihiks	Muchihiks
BD+77 84	BD+77 84	BD+77 84	BD+77 84
HIP 21386	HIP 21386	HIP 21386	HIP 21386
Cephei Sector AF-A c21	Cephei Sector AF-A c21	Cephei Sector AF-A c21	Cephei Sector AF-A c21
Cephei Sector AV-Y b6	Cephei Sector AV-Y b6 (if there's a fifth)	Cephei Sector AV-Y b6	Cephei Sector AV-Y b6
Luggerates	Luggerates	Luggerates	Luggerates

Liu Huang, Akbakara, HIP 8525 and HIP 7338 are Oya's four most-distant Controls, which could form part of some explanation for why they aren't able to attack (if that does continue next week, of course!).

 

[Aleks Zuno]

Positing attacks from HIP 8525 or especially Liu Huang is certainly the sort of thing which if it works suggests you're onto something.

Hopefully! Actually there is something missing from those tables, on which that prediction relies—the fact that Daruwach should be able to attack Cephei Sector ZE-A c8, were the closer HIP 3006 not present. That entry appears between Muchihiks and BD+77 84, separating the list a bit in a way unlike with week 35, such that following the pattern in weeks 35–37 ought to allow those unlikely attacks this time.

Cephei Sector ZE-A c8 and HIP 3006 are very, very naughty target systems for what they appear to be doing! Were those two guarded, M. Oya should start behaving.

 

[Ian Doncaster]

Interesting. So on 27 July when Daruwach and HIP 10778 between them attacked ZE-A c8 and Gliese 9035, and Daruwach skipped HIP 3006 to do so ... which one do you have going for which?

(And of course on 6 July, Sambaho joins in and all three get attacked)

 

[Aleks Zuno]

Interesting. So on 27 July when Daruwach and HIP 10778 between them attacked ZE-A c8 and Gliese 9035, and Daruwach skipped HIP 3006 to do so ... which one do you have going for which?

Goodness; let me see—that would be those week 34 targets which suggest that special rule, applicable to Daruwach and without which M. Oya would simply have lost that attack, leaving Cephei Sector ZE-A c8 clear. I think the end result is that Daruwach attacks Cephei Sector ZE-A c8 and HIP 10778 attacks Gliese 9035.

If I may introduce a few new terms:

• Evading means that a target escapes being attacked, but the attackers can still strike another target and no budget is lost.
• Missing means that an attack attempt is made but fails, and one attack budget is lost. Weirdly, none of those possible attackers are then allowed to attack anything else.

Then, the pattern has it that:

• HIP 3006 evaded Daruwach first.
• Cephei Sector ZE-A c8 evaded HIP 10778 second.
• By some special rule which needs more evidence to be sure, being evaded earlier then allows Daruwach to recover and attack Cephei Sector ZE-A c8.
• HIP 8525 and HIP 7338 missed against BD+77 84.
• By the same special rule, somehow that recovered attack takes the place of the missed one as far as the budget is concerned, otherwise there would have been only four that week.
• Additional: Kanus then attacks Cephei Sector FB-X b1-1, then finally HIP 10778 attacks Gliese 9035 after being evaded but not having missed.

That exact situation does not arise elsewhere, at least not recently in weeks 35–38 and involving systems where the misbehaviour criterion is met, so I would consider it unknown for the moment. Week 38 comes close, but now Daruwach itself is being evaded by Cephei Sector ZE-A c8 (and by HIP 3006, as it happens) with no other attackers around.

You have a good point with examining week 34 a bit more; if the more simple rule is that Daruwach cannot be evaded twice then that means it attacks Cephei Sector ZE-A c8 this week, and still fits my answer above regarding who attacked whom earlier. To continue following week 34, that attack should then take the budget place of Tougeir missing against Lahua and give us five total attacks with the other four as predicted earlier, so that can be a small item of discovery if everything else seems to fit!

For the moment a completely absent case is a misbehaviour with a double-evasion but without a miss. It is possible though, if the misbehaving systems are consecutive but not too high in the list, where the effect on the budget definitely is unknown on our part. That is to say, we know not whether a miss is required to provide the budget for that attack, or if it just happens to recycle said budget if it can but otherwise can spend anew.

 

[Ralph Vargr]

Systems misbehaving- stars throwing their planets around the pool table....

 

[Ian Doncaster]

Cephei Sector ZE-A c8 and HIP 3006 are very, very naughty target systems for what they appear to be doing! Were those two guarded, M. Oya should start behaving.

There's more to it than just those two, though, isn't there? It's not as permanent as guarding them, but on 6 July both were Alerted and defended, so going forward from there, for 13 July and 20 July neither should have been a valid target.

13 July was a mostly-normal week but still outside the pattern followed by the other seven: 4 of 4 possible Alerts, but HIP 8525 and 7338 deferred to their more distant targets (or were evaded by their closer targets, depending on theory) to go for Muncheim and Poqomathi rather than DQ-Y b2 and Bumbo.
20 July was definitely not: the Liu Huang/Akbakara group (then still including BV-Y b2) missed HIP 21386, and Daruwach missed Aowicha. Neither of those two target systems should have been interfering that time, though.

 

[Aleks Zuno]

There's more to it than just those two, though, isn't there?

Correct! There are four naughty systems at M. Oya, and they are linked in ways which seem to be unfortunate in the context of a particular glitch. Given that the misbehaviour occurs in pairs, we have:

• Cephei Sector AV-Y b2 and Cephei Sector ZE-A c8.
• Cephei Sector ZE-A c8 and HIP 3006.
• HIP 3006 and HIP 21386.

Hence, guarding Cephei Sector ZE-A c8 and HIP 3006 would remove all pairings. HIPs 8525 and 7338 are not causing a problem directly, and nor are their targets; they are simply being derailed by the above. That is to say, misbehaving targets are evaders and other innocent nearby targets get missed.

There are many other misbehaving pairs around other Maelstroms, including some which form links together as above, but none with a pair in the top few systems. In that sense, problems with targeting could be associated with having fewer targets, by means of bringing such pairs closer together. By chance, that has been accomplished at M. Oya without also clearing away the pairs themselves!

If the criterion¹ turns out to be predictable, I think M. Cocijo would be the easiest for producing it elsewhere. We would need quite a bit of clearing closer in to move the target Col 285 Sector ZE-P c6-10 further up, realistically in a week where Col 285 Sector WY-F b12-3 cannot attack (leaving Col 285 Sector ZE-P c6-13 or Col 285 Sector ZE-P c6-7 around), then a bit more clearing to bring Col 285 Sector YT-F b12-4 close enough to it.

1. It now has a quite funny name, but that would lead to describing it all and I am still quite worried that its pattern in weeks 35–37 turns out to be nothing more than coincidence.

 

[Ian Doncaster]

1. It now has a quite funny name, but that would lead to describing it all and I am still quite worried that its pattern in weeks 35–37 turns out to be nothing more than coincidence.

This is definitely my worry for my own predictions too - the pattern "closest controls attack the furthest targets they can" would, leaving aside its lack of ability to explain "4 or 5" reliably, have got the correct targets for every week 22 June to 10 August ... before failing spectacularly this week.

Fortunately for my reputation I didn't spot it as a possible pattern until this week had already started.

 

[Aleks Zuno]

Of course, just as I decided to describe it anyway, I noticed a couple of problems.

Firstly its original form does not quite work in full for week 36—in the result here, while looking only at the green colour I missed that the prediction missed that Tougeir did attack Jeng. That said, I can see a very specific difference between week 35 Jeng and week 36 Jeng which matches the original evasion criterion. Applying that difference to week 38 could change the prediction a little though, which I will describe in a moment.

Looking back at them, I can include another link in the misbehaviour chain above: Jeng and Cephei Sector AV-Y b2. For Jeng not to receive an attack in week 35, I think the earlier appearance of Cephei Sector AV-Y b2 could have turned Jeng into an evader. With Cephei Sector AV-Y b2 out of the list for week 36, Jeng is then vulnerable, and apparently Tougeir can attack it for a reason which may be related to the misbehaviour range (after having missed, little else also explains why Tougeir was not allowed to attack Cephei Sector CV-Y b1).

Applying that to week 38, it is likely that HIP 3006 turns HIP 21386 into a future evader, despite its direct conflict being with Cephei Sector ZE-A c8. If it does, indeed we have Cephei Sector AF-A c9 attacking Cephei Sector AF-A c21 instead of Liu Huang attacking HIP 21386, which seems a bit comforting that such a correction brings the prediction closer to expectation. I think HIP 8525 still attacks BD+77 84, though!

The second problem was difficult to spot; at M. Raijin it predicts that Pegasi Sector RE-N a8-4 should have evaded last week, and actually skipped down a few to attack Pegasi Sector RE-N a8-3. At present I see no sensible change which makes the attack on Pegasi Sector RE-N a8-4 succeed, and I think it suggests there are more aspects to be captured. A quick way to fix it is to allow the final attack to succeed—or only to apply the pattern at M. Oya.

 

[Aleks Zuno]

Even though the idea should be considered nascent at best, I still wanted to describe it! I will start with the more central aspects of it which seem more likely to become the basis for something better later, then speculate a bit about what appears to be happening as a result. It is very much open to simplifications!

For this purpose, a target is a system with a non-evicted Control nearby which did not attack last week; that is to say, a guarded system is removed from the list, but all secondary targets for an attacker need to be included. To start with the main criterion:

• At any given Maelstrom, let the coordinates of each target system use the Maelstrom as the spatial origin.
• If there are any targets for which any one ordinate of the three coordinates equals the negative of any for another target, call it a Negative Ordinate Conflict (NOC).

For example, HIPs 3006 and 21386 form a NOC because the HIP 3006 x-ordinate is −7.125 and the HIP 21386 y-ordinate is +7.125. HIP 3006 and Cephei Sector ZE-A c8 share ±16.375, Cephei Sectors ZE-A c8 and AV-Y b2 share ±8.96875, and I noticed earlier that Cephei Sector AV-Y b2 and Jeng share ±6.90625. Even if the remainder of this description turns out not to fit well at all, I think the idea of a NOC is at least a fairly consistent way to foresee targeting trouble.

Then, the major problems seem to occur when a NOC appears within five consecutive target entries. Stated more usefully, it appears to use a five-target window which moves down the target list normally one at a time, but pauses when a NOC pair is present in the window. The targets prior to the first NOC target are attacked if possible, all others in the window are either evaded or missed, then the entire window moves down by five.

Exactly what happens and how it proceeds there is highly speculative at the moment, but the basic idea is that the NOC attackers are evaded and can attack again, while the other attackers miss then have incredible trouble finding other targets. The only attacker which has missed then attacked eventually is Tougeir versus Jeng in week 36; the least bad idea seems to be:

• Attackers which miss can only attack again if they do not appear elsewhere in the window, or perhaps they can attack only their last target (Tougeir attacking Jeng after missing Cephei Sector CV-Y b1 in week 36).
• Attacking (or encountering?) one of a NOC pair earlier makes the other corresponding system evade later (Tougeir not attacking Jeng similarly in week 35).

The truth of that last rule would predict Cephei Sector AF-A c21 attacked this week instead of HIP 21386. If it is the case that an attacker cannot be evaded twice as with Daruwach attacking Cephei Sector ZE-A c8 in week 34, the same ought to happen again this week, perhaps also recovering the lost attack due to the Tougeir miss against Lahua.

Looking at M. Taranis in week 35, specifically at Haushu, it seems also that one system can conflict with itself and that having a zero-ordinate is an automatic NOC that way. This seems needed to explain why the attacks continue normally despite Trianguli Sector ZJ-A b1 conflicting with HIP 25840; the zero at Haushu is encountered first, which freezes the window.

As I mentioned a while ago, little of that makes any reasonable sense; it just seems to meet the incredibly strict requirements for M. Oya while, for the most part, avoiding cases elsewhere. The big exception is M. Raijin in week 37, where the fifth attack against Pegasi Sector RE-N a8-4 seemed to be fine despite the five-target window including Pegasi Sector GW-V b2-3, which shares ±22.21875. Interestingly M. Taranis came close to having a NOC in week 36; had we not guarded Ildano, it would have shared ±10.53125 with Hyades Sector PN-T c3-7 and perhaps skipped several entries.

As an example, consider week 35:

• The window starts at Cephei Sector AV-Y b2, HIP 4041, HIP 3006, Lahua, Muchihiks with 5 budget; Ardhri attacks Cephei Sector AV-Y b2.
• The window is now HIP 4041, HIP 3006, Lahua, Muchihiks, BD+77 84 with 4 budget; Sambaho attacks HIP 4041.
• The window is now HIP 3006, Lahua, Muchihiks, BD+77 84, HIP 21386 with 3 budget; HIP 3006 (−7.125 x) and HIP 21386 (+7.125 y) are a NOC pair.
  • HIP 3006 evades nobody.
  • Tougeir misses Lahua.
  • HIPs 8525 and 7338 miss BD+77 84.
  • HIP 21386 evades Cephei Sector BV-Y b2, Liu Huang and Akbakara.
• The window moves down to Cephei Sector AF-A c21, Cephei Sector CV-Y b1, Cephei Sector DQ-Y b2, Bumbo, Jeng with 2 budget; HIP 13179 attacks Cephei Sector AF-A c21.
• Tougeir skips Cephei Sector CV-Y b1.
• HIPs 8525 and 7338 skip Cephei Sector DQ-Y b2.
• HIPs 8525 and 7338 skip Bumbo.
• Tougeir can attack Jeng, but Jeng evades because its NOC-paired Cephei Sector AV-Y b2 was encountered earlier.
• HIP 7338 skips Cephei Sector FB-X b1-0.
• HIP 7338 skips Muncheim.
• Cephei Sector AV-Y b3 attacks Cephei Sector AV-Y b0, exhausting the budget.

This yields four attacks against Cephei Sector AV-Y b2, HIP 4041, Jeng and Cephei Sector AF-A c21.

Consider week 36:

• The window starts with an immediate NOC pair; HIP 3006 and HIP 21386 again.
  • HIP 3006 evades Daruwach.
  • Tougeir misses Lahua.
  • HIPs 8525 and 7338 miss BD+77 84.
  • HIP 21386 evades Liu Huang and Akbakara.
• The window moves down with 4 budget; Niu Yun attacks Cephei Sector AF-A c22, 3 budget remains.
• Tougeir skips Cephei Sector CV-Y b1.
• HIPs 8525 and 7338 skip Cephei Sector DQ-Y b2.
• HIPs 8525 and 7338 skip Bumbo.
• Tougeir attacks Jeng, its last attack having no further appearances in the window, and 2 budget remains.
• HIP 7338 skips Cephei Sector FB-X b1-0.
• HIP 7338 skips Muncheim.
• Daruwach attacks Aowicha, 1 budget remains.
• HIPs 8525 and 7338 skip HIP 2422.
• Cephei Sector AF-A c9 attacks HIP 11111.

Consider week 37:

• Immediate NOC pair; HIP 3006 and Cephei Sector ZE-A c8.
  • Cephei Sector WO-A b4 and Cephei Sector DQ-Y b1 attacked.
  • HIP 3006 and Cephei Sector ZE-A c8 evade.
  • HIPs 8525 and 7338 miss BD+77 84.
• Cephei Sector FB-X b1-1 and Gliese 9035 attacked.

Then, this week 38:

• Ardhri attacks Cephei Sector AV-Y b2.
• NOC pair; HIP 3006 and Cephei Sector ZE-A c8.
  • Sambaho attacks HIP 4041.
  • HIP 3006 evades Daruwach.
  • Tougeir misses Lahua.
  • Suspecting that it cannot be evaded twice, Daruwach possibly attacks Cephei Sector ZE-A c8 and recycles the miss, either way leaving 2 budget.
• HIP 8525 attacks BD+77 84.
• HIP 21386 probably evades Liu Huang and Akbakara, HIP 3006 having appeared earlier.
• Cephei Sector AF-A c9 attacks Cephei Sector AF-A c21.

I imagine one will want to see soon first whether there turns out to be any merit with that week 38 description, but ultimately the hope is that some of the above gives pause then some simplification emerges which clears up some of the anomalies in a better way than supposing that an extra rule exists and having no other cases to support it!

 

[Ian Doncaster]

Recapture of HIP 7338, which I don't think changes any of the predictions as none of the four models had it attacking this week anyway (for one reason or another)

 

[Aleks Zuno]

Indeed; that last worked example still seems fine!

To revisit this problem a little:

The big exception is M. Raijin in week 37, where the fifth attack against Pegasi Sector RE-N a8-4 seemed to be fine despite the five-target window including Pegasi Sector GW-V b2-3, which shares ±22.21875.

Rather than allowing the final attack to succeed arbitrarily, a better process adjustment would be:

• Place the five-system window as before.
• Stage attacks within it, up to the present budget limit, where a non-possible staged attack does not cost any staged budget.
• Only check the NOC criterion among those entries which have been encountered.
• Proceed as before.

That would make a tiny bit more reasonable sense insofar as it starts to approach the discovery of a reason why such a window might exist in the first place (other than to cause problems, that is). It would cure the week 37 exception at M. Raijin because that window has only 1 budget but needs 4 to reach the NOC pair, while preserving all M. Oya cases, including week 35 where the 3 remaining budget is just enough to span the full window due to Sambaho (real) and Tougeir (staged) having attacked already. The week 38 example is a lighter version of that, where the window has 4 budget available but gets encountered fully using only 3.

Attempting to simplify it by removing the fixed window and simply staging all budgeted attacks does not work, for Niu Yun would then miss against Cephei Sector AF-A c22 in week 36. Repairing that by suggesting that the staging stops at a NOC pair also fails, for HIP 8525 would then attack BD+77 84 in week 37. It really seems to use a fixed five-target window!

 

[Ian Doncaster]

While there have been a lot of recaptures this week, the only ones which make a difference to the predicted attacks are in Leigong, which loses everything except

Target	Distance	Controls	Confidence	Attackers	Inhabited?	Advance
Arietis Sector ZP-P b5-1	31.758	1	1	Arietis Sector MX-U c2-19;	No	-0

which it was always going to get

As far as outcomes go, the seven normal Maelstroms attacked the predictable five targets. So, on to Oya:

Null hypothesis	Narwhal hypothesis	INIV hypothesis	Nearest-Control hypothesis	Actually attacked?
Cephei Sector AV-Y b2	Cephei Sector AV-Y b2	Cephei Sector AV-Y b2	Cephei Sector AV-Y b2	No
HIP 4041	HIP 4041	HIP 4041	HIP 4041	No
HIP 3006	HIP 3006	HIP 3006	HIP 3006 (if there's a fifth)	Yes
Lahua	Lahua	Lahua	Lahua	No
Muchihiks	Muchihiks	Muchihiks	Muchihiks	No
BD+77 84	BD+77 84	BD+77 84	BD+77 84	No
HIP 21386	HIP 21386	HIP 21386	HIP 21386	No
Cephei Sector AF-A c21	Cephei Sector AF-A c21	Cephei Sector AF-A c21	Cephei Sector AF-A c21	Yes
Cephei Sector AV-Y b6	Cephei Sector AV-Y b6 (if there's a fifth)	Cephei Sector AV-Y b6	Cephei Sector AV-Y b6	Yes
Luggerates	Luggerates	Luggerates	Luggerates	Yes
Cephei Sector AV-Y b0	Cephei Sector AV-Y b0	Cephei Sector AV-Y b0	Cephei Sector AV-Y b0	Yes
20%	60%	0%	80%	They're on to us...

Well, that's new information, at least.

AV-Y b0 appears to be a deferred attack from AV-Y b3 which didn't attack AV-Y b6 or AV-Y b2 (and neither did Ardhri for b2 - they must have attacked b6 instead). As an uninhabited->uninhabited move it's not unprecedented, so it looks like - for Oya - I need to add these options to the tables as well. Doing that might fix the "nearest Control attacks furthest target" hypothesis if there's no prioritisation of inhabited systems - Ardhri attacks b6 then b3 attacks b0 would fit with that, I think, but I need to actually run the tables.

(and having done that fix, run it for the attacks on the 17th to see if it explains the gap there)

 

[Aleks Zuno]

Regarding the additional uninhabited targets:

• Having Cephei Sector AV-Y b0 in the list would indeed take the attack away from Cephei Sector AV-Y b2; however—
• Given that Sambaho must have attacked Luggerates, having Cephei Sector ZE-A c8 in the list would have taken the Daruwach attack away from HIP 3006.

Attacking furthest targets definitely is quite good, and I think if there is one thing I can see which could exclude Cephei Sector ZE-A c8 and thus let the furthest Daruwach target be HIP 3006, it is that Cephei Sector ZE-A c8 is where the first NOC occurs (perhaps two at once; sharing ±16.375 with HIP 3006 and also ±8.96875 with Cephei Sector AV-Y b2 earlier).

I will check that idea a bit more later; the hope is to discover that striking the furthest targets sometimes causes four total attacks naturally, or secondarily to seek a consistent indication for that!

 

[Ian Doncaster]

I will check that idea a bit more later; the hope is to discover that striking the furthest targets sometimes causes four total attacks naturally, or secondarily to seek a consistent indication for that!

Yes - the "nearest control attacks furthest target" idea, while seeming to get reasonably good indications of targeting, has no way at all to explain why sometimes (more often that not) there are only four attacks.

 

[Ian Doncaster]

Narwhal Nose report generated 24 August

In general this might be a record low week for inhabited attacks - only 9 predicted at non-Oya maelstroms, some of which could themselves be blocked.

-------------------------------------------

Systems are listed in order of priority, and in general the first five would be expected to be hit by Alerts, with a variable number of reserve targets shown depending on confidence levels. Predictions made early in the week may be reissued later if Control recaptures materially affect them.

Table key
Name: Name of system considered at threat of Alert.
Distance: The distance in LY to the Maelstrom. A primary component of Thargoid prioritisation.
Controls: The number of Controls which can theoretically attack this system this week.
Confidence: The confidence that any Control can attack this system this week. Numbers under 1 should be treated as increasingly unlikely, while numbers above 2 are almost certain within the constraints of the model.
Attackers: A list of potential Controls to launch the attack from, partial if there are 4 or more. Controls disconnected from their parent maelstrom are marked in italics; controls attacking a system closer to the maelstrom than themselves are underlined; controls which may be skipping an inner uninhabited system to strike an inhabited system further out are struck-through; controls containing a barnacle matrix are bold.
Inhabited: Whether the system contains standing human population.
Advance: The approximate direction of the Alert relative to the Maelstrom. +1=towards Sol, -1=away from Sol, 0=perpendicular to Sol

-------------------------------------------

Calculating Taranis
Read 42 control systems and 953 targets.
Total connected: 39
Disconnected Controls

System	Distance
Hyades Sector BV-O b6-4	11.71637476
Trianguli Sector EQ-Y b0	16.57010439
Hyades Sector EG-N b7-1	18.06744634

Target	Distance	Controls	Confidence	Attackers	Inhabited?	Advance
Hyades Sector FB-N b7-3	15.476	7	7	Hupang; Hyades Sector FB-N b7-4; HIP 23816; etc	No	0.4
Trianguli Sector EQ-Y b3	17.317	1	1	Trianguli Sector EQ-Y b0;	No	0.6
Trianguli Sector FL-Y b5	19.749	2	2	Trianguli Sector BA-A d85 (9.921 LY); Trianguli Sector FL-Y b4;	No	0.7
Trianguli Sector KR-W b1-4	21.129	1	1	Hyades Sector HW-M b7-4;	No	0.7
Hyades Sector GW-W d1-88	22.107	3	3	Hyades Sector HW-M b7-0; Trianguli Sector FL-Y b4; Hyades Sector HW-M b7-1 (10.021 LY);	No	0.3
Hyades Sector EQ-O b6-2	22.572	4	4	Hyades Sector EQ-O b6-1; Hyades Sector DQ-O b6-3; 5 Mu Leporis; etc	No	-0.9
Hyades Sector QN-T c3-11	22.991	4	4	Hyades Sector DQ-O b6-3; Hyades Sector QN-T c3-12; HIP 24834; etc	No	-1
Hyades Sector PN-T c3-5	23.218	2	2	Ixbalan; Hyades Sector GW-W d1-99;	No	-0.9
Hyades Sector EG-N b7-3	24.638	1	1	Hyades Sector EG-N b7-1;	No	0.1
Hyades Sector GW-W d1-82	25.603	1	1	Hyades Sector BV-O b6-3;	No	-0.7
Hyades Sector GB-N b7-3	25.796	2	2	Hyades Sector CV-O b6-2; Hyades Sector CV-O b6-4;	No	-0.6
HIP 26534	28.300	1	1	Hyades Sector CV-O b6-4;	No	-0.8
Hyades Sector CV-O b6-6	29.624	3	3	Hyades Sector QN-T c3-12; HIP 24834; Hyades Sector CV-O b6-5;	No	-1

No further targets possible
Taranis getting to the point where its Matrix systems will maintain a perimeter against further shrinking

-------------------------------------------

Calculating Leigong
Read 61 control systems and 937 targets.
Total connected: 55
Disconnected Controls

System	Distance
Hyades Sector GR-W d1-93	16.95933786
Hyades Sector GH-M b7-1	19.61536885
Hyades Sector GH-M b7-2	19.92191789
Hyades Sector GH-M b7-0	20.74929404
Hyades Sector FH-M b7-2	21.17212292
HIP 8467	23.45403583

Target	Distance	Controls	Confidence	Attackers	Inhabited?	Advance
Arietis Sector KM-W d1-93	20.925	4	4	HIP 9709; Arietis Sector BQ-P b5-2; Arietis Sector KM-W d1-94; etc	No	0.9
Pathamon	22.064	3	3	HIP 7277; Arietis Sector JM-W d1-56; Arietis Sector MX-U c2-17;	Yes	0.5
Jementi	25.007	2	2	Arietis Sector NX-U c2-19; Patollu;	Yes	0.9
Ceti Sector XJ-A c18	25.359	1	1	Patollu;	No	0.9
Hyades Sector LY-I b9-0	28.457	1	1	Arietis Sector MX-U c2-17;	No	0.4
Haitchane	29.580	1	1	Hyades Sector NT-I b9-1;	Yes	0.2
Hyades Sector NT-I b9-2	31.119	5	5	HIP 6570; HIP 6572; Hyades Sector KX-U d2-94; etc	No	-0.2
HIP 5629	40.949	1	1	Hyades Sector NT-I b9-4;	No	-0

No further targets possible
Leigong bringing most of its perimeter close-in though there's some scope for recapturing to adjust its targeting a bit

-------------------------------------------

Calculating Indra
Read 141 control systems and 853 targets.
Total connected: 141

Target	Distance	Controls	Confidence	Attackers	Inhabited?	Advance
Arietis Sector LM-V b2-3	20.115	4	4	HIP 19098; Obamumbo; Arietis Sector LM-V b2-5; etc	No	0.1
HIP 19157	21.650	1	1	Arietis Sector IG-X b1-4;	Yes	-0.6
HIP 20890	21.889	10	10	64 Tauri (10.013 LY); HIP 20577; HR 1354; etc	Yes	0.9
Arietis Sector NH-V b2-0	22.472	1	1	Obamumbo;	No	-0.2
HIP 20527	22.540	16	16	Hyadum I; HIP 20916; HIP 20577; etc	Yes	0.9
HIP 20899	23.086	16	16	HIP 20916; HIP 20577; HR 1354; etc	Yes	1
HIP 20491	23.341	5	5	HIP 19934 (10.022 LY); HIP 20712; Arietis Sector OS-R a5-0; etc	Yes	0.4
HIP 20741	24.478	16	16	HIP 20577; HR 1354 (10.023 LY); Hyadum II; etc	Yes	1
HIP 22203	25.394	4	4	HIP 22496; 92 Sigma-2 Tauri; HIP 22422; etc	No	0.8
69 Upsilon Tauri	26.471	3	3	HIP 20712; HIP 21008; Arietis Sector JR-V b2-3;	Yes	0.8
90 c1 Tauri	26.476	2	2	79 b Tauri; 92 Sigma-2 Tauri;	No	0.8
Arietis Sector DQ-Y c18	26.589	4	4	Arietis Sector HG-X b1-3; Arietis Sector HG-X b1-4; Arietis Sector DQ-Y c9; etc	Yes	-0.4
HIP 21380	26.610	4	4	HIP 21008; Arietis Sector TY-P a6-0; Arietis Sector TY-P a6-1; etc	Yes	0.8
HIP 20850	26.986	9	9	77 Theta-1 Tauri; 70 Tauri; HIP 20815; etc	Yes	0.9

63 additional targets not listed, 77 total
A fairly normal week at Indra, though earlier recaptures are starting to place uninhabited systems back onto its targeting list

-------------------------------------------

Calculating Oya
Read 28 control systems and 966 targets.
Total connected: 25
Disconnected Controls

System	Distance
Tougeir	16.22455941
Liu Huang	17.75687587
Akbakara	18.17006789

Target	Distance	Controls	Confidence	Attackers	Inhabited?	Advance
Lahua	19.363	1	1	Tougeir;	Yes	0.9
Muchihiks	19.561	1	0.2	Tougeir;	Yes	0.9
Cephei Sector ZE-A c8	19.896	1	1	HIP 10778;	No	-0.4
BD+77 84	20.166	1	1	HIP 8525;	Yes	0.9
HIP 21386	20.509	2	2	Liu Huang; Akbakara;	No	0.7
Cephei Sector AF-A c22	21.188	3	3	Niu Yun (9.908 LY); Warnones; HIP 13179;	No	0.5
Cephei Sector CV-Y b1	22.039	1	0.2	Tougeir;	No	0.9
Cephei Sector DQ-Y b2	22.149	1	0.2	HIP 8525;	No	0.5
Bumbo	22.416	1	0.2	HIP 8525;	Yes	0.9
Jeng	23.038	1	0.2	Tougeir;	Yes	0.5
HIP 2422	24.009	1	0.2	HIP 8525;	Yes	0.5
HIP 11111	24.118	1	1	HIP 13179;	Yes	0.6
Poqomathi	24.136	1	0.2	HIP 8525;	Yes	0.6
Cephei Sector DQ-Y b6	24.953	1	0.2	HIP 8525;	No	0.9
Gliese 3050	25.579	1	0.2	HIP 8525;	No	0.6

No further targets possible
Applying the rule that the closest control attacks the furthest target it can gives the following sequence:
- HIP 10778 attacks ZE-A c8
- Niu Yun attacks AF-A c22
- HIP 13179 attacks HIP 11111 (these first three are the only targets)
- Tougier now has quite a few choices and Jeng is the most distant, also inhabited - this is what they did three weeks ago
- if there is a fifth attack, it would be Liu Huang against HIP 21386, though also no surprise if there isn't a fifth attack.
- if recaptures knock out any of those, HIP 8525 against either Poqomathi or Gliese 3050 would be the reserve, and no further attacks are possible beyond that (edit: except that Akbakara could have a go at HIP 21386 if Liu Huang and several others went)

-------------------------------------------

Calculating Cocijo
Read 107 control systems and 888 targets.
Total connected: 103
Disconnected Controls

System	Distance
Col 285 Sector OS-T d3-106	23.03364557
Col 285 Sector YT-F b12-2	24.23140798
Col 285 Sector VN-H b11-2	25.80003104
Col 285 Sector OM-J b10-0	30.84914783

Target	Distance	Controls	Confidence	Attackers	Inhabited?	Advance
Col 285 Sector YT-F b12-6	20.034	1	1	Mapon;	No	1
HIP 38718	20.765	2	2	Mahlina; Col 285 Sector ZE-P c6-7;	No	0.9
Col 285 Sector SS-H b11-0	23.090	2	2	Col 285 Sector VY-Q c5-17; Col 285 Sector VY-Q c5-18;	No	0.5
Col 285 Sector SS-H b11-3	23.190	1	1	Col 285 Sector VY-Q c5-18;	No	0.5
Canaharvas	23.548	1	1	Col 285 Sector ZE-P c6-7;	Yes	0.9
Col 285 Sector KM-V d2-69	23.672	2	2	Col 285 Sector SS-H b11-6; Col 285 Sector KM-V d2-109;	Yes	0.1
Col 285 Sector YT-F b12-7	23.878	1	1	Col 285 Sector YT-F b12-2;	No	0.7
Col 285 Sector YT-F b12-4	24.162	2	2	HIP 36901; Col 285 Sector YT-F b12-0;	No	0.9
Col 285 Sector SS-H b11-2	24.329	1	1	Col 285 Sector UN-H b11-2;	No	0.5
Paitra	24.880	6	6	Col 285 Sector UN-H b11-4; Col 285 Sector UN-H b11-5; Col 285 Sector UN-H b11-3; etc	Yes	0.7
Col 285 Sector BA-P c6-4	25.044	1	1	Col 285 Sector OS-T d3-106;	No	0.5
Col 285 Sector XY-F b12-1	25.325	1	1	Col 285 Sector OS-T d3-69;	No	0.1
Col 285 Sector UD-G b12-2	25.947	2	2	Kurumanit; Col 285 Sector WY-F b12-1;	Yes	0.4
Col 285 Sector OS-T d3-115	26.097	1	1	Col 285 Sector DA-E b13-4;	No	-0.1
Col 285 Sector ZT-F b12-5	26.825	1	1	Col 285 Sector ZT-F b12-4;	No	0
Col 285 Sector VD-G b12-1	27.061	1	1	Col 285 Sector XY-F b12-6;	No	-0.5
Col 285 Sector WT-Q c5-22	27.359	3	3	Col 285 Sector UN-H b11-3; Col 285 Sector XT-Q c5-1; Col 285 Sector XT-Q c5-22;	No	0.7

22 additional targets not listed, 39 total
A normal week at Cocijo

-------------------------------------------

Calculating Thor
Read 73 control systems and 922 targets.
Total connected: 66
Disconnected Controls

System	Distance
Rajuarpai	11.4695589
Chanyaya	16.10879151
HIP 19894	16.20887223
Col 285 Sector IG-O c6-14	17.71585625
HIP 20024	19.46077284
Col 285 Sector UH-C b13-1	21.81126877
HIP 18857	24.61621433

Target	Distance	Controls	Confidence	Attackers	Inhabited?	Advance
HIP 19600	20.316	2	2	Col 285 Sector LV-F b11-1; Col 285 Sector LV-F b11-0;	No	-0.1
Col 285 Sector UH-C b13-2	21.239	1	1	Col 285 Sector QB-E b12-0;	No	-0.5
Col 285 Sector UH-C b13-0	22.796	2	2	Col 285 Sector UH-C b13-1; HIP 18857;	No	-0.2
Col 285 Sector IA-G b11-3	24.220	1	1	Col 285 Sector IG-O c6-18;	No	0.6
Col 285 Sector JG-O c6-5	24.842	2	2	Col 285 Sector IG-O c6-14; HIP 20024;	No	0.3
Unktety	25.357	1	1	HIP 19198;	Yes	1
Col 285 Sector OC-V d2-81	29.906	1	1	Col 285 Sector LV-F b11-0;	No	-0.1

No further targets possible
Thor also spends another week targeting close-in systems

-------------------------------------------

Calculating Raijin
Read 123 control systems and 871 targets.
Total connected: 122
Disconnected Controls

System	Distance
Hosan	27.87701436

Target	Distance	Controls	Confidence	Attackers	Inhabited?	Advance
Nu Guang	19.330	1	1	Snoqui;	Yes	0.9
Pegasi Sector DG-X c1-6	21.188	2	2	HIP 115162; Nibelaako;	No	0
Pegasi Sector NY-O a7-1	22.028	2	2	Pegasi Sector JH-U b3-9; Pegasi Sector MY-O a7-5;	No	0.5
Pegasi Sector UK-L a9-0	23.218	10	10	Pegasi Sector VK-L a9-0; Pegasi Sector TK-L a9-2; Chakma; etc	No	0.3
Ixbaksha	23.247	9	9	Pegasi Sector MN-S b4-7; Chakma; Pegasi Sector UK-L a9-3; etc	Yes	0.5
Pegasi Sector EB-W b2-2	23.770	3	3	Pegasi Sector IH-U b3-4; Nibelaako; Pegasi Sector MY-O a7-0;	No	0.1
Isla	23.900	1	1	Pegasi Sector LC-U b3-2 (9.952 LY);	Yes	-0.1
HIP 116542	24.019	2	2	Chnemine; Pegasi Sector HW-V b2-1;	No	-0.1
HIP 111081	24.703	13	13	Pegasi Sector MN-S b4-8; Pegasi Sector MN-S b4-5; Pegasi Sector TK-L a9-2; etc	No	0.2
Pegasi Sector NY-O a7-2	24.719	1	1	Pegasi Sector MY-O a7-5;	No	0.5
Aurus	25.723	5	5	Pegasi Sector PI-S b4-0 (10.004 LY); Pegasi Sector KH-V c2-20; Iduni; etc	Yes	0.1
Pegasi Sector OI-S b4-1	25.760	2	2	Pegasi Sector OI-S b4-0; Hosan;	No	-0.4
Pegasi Sector GW-V b2-3	25.918	2	2	Pegasi Sector KC-U b3-6 (9.914 LY); Pegasi Sector BQ-Y d71;	No	-0.6
Pegasi Sector RE-N a8-3	26.050	1	1	Pegasi Sector JH-U b3-8;	No	0.7
Pegasi Sector FB-X c1-22	26.084	1	1	Pegasi Sector KC-U b3-2;	No	-0.5

56 additional targets not listed, 71 total
As within Indra, recent recaptures at Raijin are starting to add more uninhabited systems to its inner targets

-------------------------------------------

Calculating Hadad
Read 64 control systems and 929 targets.
Total connected: 63
Disconnected Controls

System	Distance
Col 285 Sector VS-Z b14-3	24.11759116

Target	Distance	Controls	Confidence	Attackers	Inhabited?	Advance
HIP 29596	18.225	1	1	Montioch;	Yes	0.8
Col 285 Sector SH-B b14-2	18.310	2	2	Col 285 Sector QM-B b14-1; Col 285 Sector QM-B b14-2;	No	-0.7
Col 285 Sector KW-M c7-31	20.046	6	6	Col 285 Sector KW-M c7-7; Col 285 Sector KW-M c7-4; HIP 31648; etc	No	-0
Col 285 Sector RM-B b14-8	20.607	1	1	Col 285 Sector RM-B b14-3;	No	-0.9
Col 285 Sector PM-B b14-3	20.783	7	7	HIP 31648; Col 285 Sector KW-M c7-5; Col 285 Sector KW-M c7-2; etc	No	0.3
Col 285 Sector KW-M c7-29	21.254	4	4	Col 285 Sector KW-M c7-4; Col 285 Sector QM-B b14-7; Col 285 Sector KW-M c7-30; etc	No	-0.4
Col 285 Sector WN-Z b14-5	21.482	2	2	Col 285 Sector KW-M c7-14; Col 285 Sector KW-M c7-15;	No	-0.6
Col 285 Sector RM-B b14-4	21.892	2	2	HIP 31223; Col 285 Sector VS-Z b14-3;	No	-0.9
Col 285 Sector RM-B b14-6	22.792	3	3	Col 285 Sector QM-B b14-7; Col 285 Sector KW-M c7-30; Col 285 Sector KW-M c7-28;	No	-0.4
Col 285 Sector SX-Z b14-2	22.853	2	2	Col 285 Sector US-Z b14-1; Col 285 Sector SX-Z b14-0;	No	0.3
Col 285 Sector VS-Z b14-1	22.869	1	1	Col 285 Sector KW-M c7-15;	No	-0.7
Col 285 Sector SX-Z b14-1	23.453	1	1	Col 285 Sector SX-Z b14-0;	No	0.4
Col 285 Sector VS-Z b14-6	23.515	1	1	Col 285 Sector VS-Z b14-3;	No	-0.6
Col 285 Sector NR-B b14-1	23.674	1	1	Col 285 Sector PM-B b14-1;	No	0.9
Col 285 Sector VS-Z b14-5	24.456	1	1	HIP 32764;	No	-0.4
HIP 29217	25.118	1	1	Col 285 Sector QM-B b14-2;	No	-0.6

6 additional targets not listed, 22 total
If the Hadad attacks continue to go for Montioch there's a chance that Hadad will be on 5/5 uninhabited targets for the first time in a while.

Spoiler: Exhausted Controls

-- Recaptures above (2 if confirmed, probability otherwise; exhausts below at exactly 1)
-- Taranis
HIP 25654:1
Hyades Sector GB-N b7-2:1
Trianguli Sector CA-A c22:1
Trianguli Sector CA-A c15:1
Hyades Sector BV-O b6-4:1
-- Leigong
Arietis Sector MX-U c2-19:1
-- Indra
71 Tauri:1
Arietis Sector JR-V b2-4:1
HIP 20679:1
HIP 20480:1
86 Rho Tauri:1
-- Oya
Ardhri:1
Daruwach:1
Cephei Sector AF-A c9:1
Cephei Sector AV-Y b3:1
Sambaho:1
-- Cocijo
Col 285 Sector OS-T d3-96:1
Col 285 Sector WY-F b12-3:1
Col 285 Sector SS-H b11-1:1
Col 285 Sector VY-Q c5-15:1
Col 285 Sector ZE-P c6-13:1
-- Thor
Chanyaya:1
Col 285 Sector JA-G b11-2:1
Col 285 Sector NG-E b12-5:1
Col 285 Sector EA-Q c5-5:1
Col 285 Sector NG-E b12-2:1
-- Raijin
Pegasi Sector IH-U b3-0:1
Pegasi Sector QE-N a8-1:1
Pegasi Sector NN-S b4-10:1
Sholintet:1
HIP 113535:1
-- Hadad
Col 285 Sector US-Z b14-2:1
Col 285 Sector RN-T d3-78:1
Col 285 Sector RH-B b14-2:1
Col 285 Sector US-Z b14-7:1
Col 285 Sector SH-B b14-4:1

 

[Aleks Zuno]

I looked back through weeks 34–38 again; it can be explained almost entirely by allowing normal attacks up to the first system of the first NOC pair, removing those normal attackers and targets from the list, then switching to use nearest-attacks-farthest!

The remaining minutiae are:

• Detection of the first NOC pair can now occur more simply just by examining the targets encountered while placing those five normal attacks. That limiting factor is still needed (see M. Indra and M. Raijin!).
• During nearest-attacks-farthest—
  • An attack upon either of that first NOC pair is skipped, and this attacker becomes the last (further targets are still considered).
  • An attack upon any other NOC system succeeds, and this attacker becomes the last.
• Based only on the unique circumstances of Daruwach in week 34, it seems that a single attacker which can see both of the offending NOC pair may skip the first and attack the second; regardless of the reason, it is clear that this attack must occur just before proceeding with nearest-attacks-farthest. Attempting to apply nearest-attacks-farthest to Daruwach here fails in two ways:
  • Cephei Sector AV-Y b3 is nearer than Daruwach, but did not attack Cephei Sector AV-Y b0.
  • Aowicha and Gliese 9035 are farther than Cephei Sector ZE-A c8, but Daruwach attacked the latter.

That reduces the total number of rules a lot, and I can use them to write fully-worked examples for weeks 34–38 if interested!

That said, looking at week 39 it may not be worth trying to continue that pattern yet, because it makes the quite unlikely prediction that M. Oya just attacks its normal five targets (top six without Muchihiks) due to having no NOC pairs this week:

• If that does occur, something must have changed after week 33, because weeks 32–33 displayed minor misbehaviour despite also having no NOC targets.
• If just applying nearest-attacks-farthest gives at least a good prediction, we are looking for something else which causes the behavioural change.
  • If not at least related to NOCs in some way, it will be one of the most amazing coincidences that they correlate in weeks 34–38!

 

[Ian Doncaster]

Starting off in closest first and then switching to nearest-attacks-farthest might well be where this is going.

I'm looking at something like:

• start in classic mode
• if you reach an attack from a control system which could also attack a further target, defer the attack and switch to NAF targeting
• if NAF tries to deploy the 5th attack before it's got far enough down the control list to reach the deferred attack, Thargoid logistics get confused about what they're supposed to be doing, and the 5th attack doesn't happen
• if in NAF the next target can be seen by the deferred control, use it for the attack even if it's not the nearest control at this point (added to explain week 34, doesn't seem to break anything else)

It explains why there were only 4 alerts in weeks 35 to 37, then this week Ardhri instantly switches into NAF and is close-in so it gets the full 5, and it also seems to work fine for weeks 32 and 33.

In week 34, you then get the following:
- WO-A b4 classically targeted, WO-A b3 has no alternative but to attack
- DQ-Y b1 classically targeted, AF-A c9 has no alternative but to attack
- HIP 3006 classically targeted, Daruwach can attack Aowicha, defer Daruwach and switch to NAF
- HIP 10778 is the closest control, its furthest target is Gliese 9035. Daruwach can see that and is scheduled to attack, so use Daruwach
- HIP 10778 is still the closest control, its furthest target is now ZE-A c8 which it attacks
- Kanus is now the closest control and attacks FB-X b1-1
- AV-Y b3 attacking AV-Y b0 is next on the list, but there's no budget left.

If that's the case, this week: Tougier can see Jeng, so it instantly switches the targeting to NAF and defers; being 4th on the control list the attack will then go through and so the 5th attack from Liu Huang should happen.

 

[Aleks Zuno]

I presume the deferral system needs very much an Maelstrom-Oya-only stipulation though; for example, if I try it upon M. Hadad in week 38 where all sectors are Col 285:

• US-Z b14-2 attacks KW-M c7-12 normally (only target).
• RN-T d3-78 defers attacking RM-B b14-2 (sees KW-M c7-29 and RM-B b14-0):
  • RH-B b14-2 at 9.38 Ly attacks Muruidooges.
  • US-Z b14-7 at 10.06 Ly attacks VS-Z b14-0.
  • Montioch at 13.34 Ly attacks TS-Z b14-3.
  • KW-M c7-4 at 13.88 Ly sees KW-M c7-29:
    • RN-T d3-78 strikes instead (Daruwach stand-in rule)...
    • But fails (fifth attack rule).

For a Maelstrom which attacked actually targets 1–5 that yielded targets 1, 3, 4 and 11 with a failed attempt at target 8!

Incidentally, including into the list the systems which are in respite has a positive effect for the NOC and NAF approach! Week 33 then works, and weeks 34–38 are unaffected, although week 39 still maintains an unlikely insistance that everything will be normal this time. I am trying to pester V into making a quite big list which only considers Control versus non-Control to see if that coerces yet more cases into succeeding via NAF, then similarly it will need rechecking that no other Maelstroms are affected wrongly.

 

[Ian Doncaster]

I presume the deferral system needs very much an Maelstrom-Oya-only stipulation though

Yes. With all the others behaving exactly as the classic model predicts, I'm assuming that there's something specific about Oya which is enabling this alternative targeting mode.

If it continues to explain well next week, I might go back and try it on the very early weeks to see if it also explains the more sparse branching strategy in use then: instinctively it feels like it should give that sort of behaviour while a maelstrom is relatively short on perimeter controls but not actively being contained. Of course, those also have the further budgetary complexity of uninhabited systems costing 1/4 as much and so a lot more alerts to place.