Investigating the new states

[Ian Doncaster]

I should post my current theory on how these states actually work here... I forgot about this thread. If anyone has an ultra-low population (double figures ideal) with no traffic which they'd like to use to look into the actual causes (and counters?) of these events, then I'm happy to collaborate on designing the tests, but I have nowhere remotely suitable near Colonia to actually carry them out.

New events theory - thank you to Brap and Jane for each providing some classified information on activity to fine-tune it.

Event type states are:

• based on buckets like the old pre-3.3 states
• bucket size probably related to system population: events more likely in lower population systems
• buckets leak - evidence: factions with no events at all, despite some activity in the system
• buckets reset / can't be filled while an event is active or recovering
• first bucket to fill triggers the event, presumably some priority order if multiple buckets fill on the same day

Other than Outbreak, events have no pending period (Outbreak has 3 days pending) but it does take at least one tick to fill the buckets, so the minimum time between events is 15 days (14 days recovery from the last one, 1 day to fill the buckets, then it goes) - this appears to be a change from 3.3 when Pirate Attack had a hidden pending period giving a minimum time between events of 25 days. That being the case, figuring out what actions cause which event should be practical by conventional analysis methods.

A quick data table on minimum times between event pairs (first event on the left, second event at the top) based on Colonia region data. Note the longer transition to Outbreak with its visible pending period.

Fv T> Out Pir Dro Bli Ter Inf Nat Pub ANY
Expan 56  --  --  --  --  44  36  --  36
Outbr 19  18  26  19  16  16  26  15  15
Pirat 26  18  15  19  18  17  --  19  15
Droug 21  29  25  25  28  16  --  20  16
Bligh 21  --  16  23  20  17  23  15  15
Terro 28  17  30  18  20  28  28  17  17
Infra 19  15  15  15  20  16  22  15  15
Natur 32  21  25  22  --  24  --  24  21
Publi 24  18  19  77  17  16  30  15  15
ANY   19  15  15  15  16  16  22  15  15

(Natural Disaster is almost certainly the same as the others, but its so rare out here that the data is too sparse to be sure)

Hypothesis: the way that Infrastructure Failure and Natural Disaster cause events on other factions is by direct filling of those buckets, which could therefore be countered by direct emptying actions, if people knew what they were.

 

[Dommarraa]

Interesting low pop (794) system. Has only had 1 or 2 days with ANY player activity since middle of August, but has still had Drought, Bust, Civil Unrest, Lockdown and Outbreak.

INARA

The community site for Elite: Dangerous, Starfield, Kingdom Come: Deliverance with the game databases, information, gaming news and other handy content...

inara.cz

 

[FrogsFriend]

Interesting low pop (794) system. Has only had 1 or 2 days with ANY player activity since middle of August, but has still had Drought, Bust, Civil Unrest, Lockdown and Outbreak.

INARA

The community site for Elite: Dangerous, Starfield, Kingdom Come: Deliverance with the game databases, information, gaming news and other handy content...

inara.cz

Do you think that was all driven by the (presumably unmitigated) drought back in Sept?

 

[Dommarraa]

Do you think that was all driven by the (presumably unmitigated) drought back in Sept?

Probably. Its just that people keep telling me that nothing happens without CMDR action. I think it demonstrates that things also happen with CMDR inaction.

 

[Ian Doncaster]

Probably. Its just that people keep telling me that nothing happens without CMDR action. I think it demonstrates that things also happen with CMDR inaction.

Well, or at least that the effects of CMDR action can play out over multiple ticks, leaving cause and effect quite distant.

 

[goemon]

Well, or at least that the effects of CMDR action can play out over multiple ticks, leaving cause and effect quite distant.

that's not "over multiple ticks" only. that's 2-3 months.
4.8: outbreak (1). There is a tiny influence move a week later (next date somebody updated the system).
12.9.: Draught. There is a month between the two, or three weeks after the last influence movement record.
19.9.: Bust, Civil Unrest - i assume the Draught hit security and economy slider.
18.10. Outbreak (2), Bust, lockdown with tiny influence movement - i assume, the outbreak hit economy and security slider again, triggering lockdown.
(we suspected an attack originally, so we did nothing about all that).

there is something really interesting about the 2nd outbreak. if an outbreak bucket would have been emptied after the first one, there shouldn't have been a second one so soon. activity in system would have moved influence, which it didn't (beside two times in really small amounts)

so ... my sketch would be:

• Outbreak(1) raised the probability of Draught above 0 (as well as other probability states?).
• Dice rolls on chance of Draught, Draught finally happens after 3-4 weeks.
• Draught fills the outbreak bucket (or bust or civil unrest fill the outbreak bucket?)
• Dice rolls on outbreak, finally a month later outbreak happens.

the "bucket" on draught, outbreak (?), infrastructure failure would be in this case a percentage of the state to happen.

I'm not sure whether this could be actually tested ... because with such a model you'd trigger draught et. sim. at various levels. It just would get more probable over time to trigger it.

now - it will be interesting, whether that system gets another draught ever - if nobody is active there. or whether the probability is 0 after the state has happened. that could explain why systems with all factions in no-state are a very good indicator of no activity.

 

[Ian Doncaster]

that's not "over multiple ticks" only. that's 2-3 months.
4.8: outbreak (1). There is a tiny influence move a week later (next date somebody updated the system).
12.9.: Draught. There is a month between the two, or three weeks after the last influence movement record.
19.9.: Bust, Civil Unrest - i assume the Draught hit security and economy slider.
18.10. Outbreak (2), Bust, lockdown with tiny influence movement - i assume, the outbreak hit economy and security slider again, triggering lockdown.
(we suspected an attack originally, so we did nothing about all that).

Under the theory that negative actions only/primarily change where positive influence redistributes from, perhaps the only in-system actions were negative? Not an area I know much about.

Drought causing Bust + Civil Unrest seems pretty normal.

Outbreak does not appear to cause security or economy damage, though ... so unless those states were still left over from the aftermath of the Drought?

there is something really interesting about the 2nd outbreak. if an outbreak bucket would have been emptied after the first one, there shouldn't have been a second one so soon. activity in system would have moved influence, which it didn't (beside two times in really small amounts)

I'm fairly sure - because of the number of system-factions which get the same event repeatedly, and also because busy systems don't guarantee events (of any sort) recurring with minimum possible period - that any event occurring empties all event buckets at the end of its recovering period anyway.

 

[goemon]

Under the theory that negative actions only/primarily change where positive influence redistributes from, perhaps the only in-system actions were negative? Not an area I know much about.

if you apply -inf actions, even if you only apply -inf actions, you get influence movement. which isn't the case there.
if not ... if not certain -inf actions have no effect during that state. like combat actions during election. that's something i don't know much about. but for exampel smuggling has no effect during war/civil war.

I'm fairly sure - because of the number of system-factions which get the same event repeatedly, and also because busy systems don't guarantee events (of any sort) recurring with minimum possible period - that any event occurring empties all event buckets at the end of its recovering period anyway.

with you, which means either draught or bust or civil unrest will have raised the possibility of outbreak. even if my understanding is, a states bucket gets emptied once filled and state triggered (so at pending, not at end of it). but that's a mainly academic difference.

Outbreak does not appear to cause security or economy damage, though ... so unless those states were still left over from the aftermath of the Drought?

something have to have happened to move the security slider from civil unrest to lockdown.

 

[Ian Doncaster]

with you, which means either draught or bust or civil unrest will have raised the possibility of outbreak. even if my understanding is, a states bucket gets emptied once filled and state triggered (so at pending, not at end of it). but that's a mainly academic difference.

When the bucket is emptied isn't exactly relevant, except that if it's emptied before end of recovery I don't think any actions to fill it (for any event) get counted during recovery (of any event). If they did, I think we'd see smaller gaps and potentially more variety in event types, in the really busy systems.

something have to have happened to move the security slider from civil unrest to lockdown.

Definitely. Are there actions which harm security but not influence, perhaps?

 

[goemon]

Definitely. Are there actions which harm security but not influence, perhaps?

well - usually not. but during outbreak combat actions have "no effect" (first FDEV BGS update) - i could imagine them not having an influence effect by now, but still moving security slider to trigger civil unrest -> lockdown.

 

[goemon]

maybe we have to come up with a list of:

• "slider states" (state gets active by lowering or upping a slider; i think this is what we are used to call buckets, but the current implementation is actually not one of a bucket style)
• "probability states" (state has a probability to happen, and actions raise or reduce the probability; see combat bond effect on pirate attack and outbreak; again i think a "bucket" of probability is a misleading analogy)

... and than map effect of states on states.

 

[Factabulous]

Surely the first thing to do is find what fills the buckets?

I've tried a few commodities that I thought might kick states off (Toxic Waste / Chemicals etc), but never seen anything happen. I assumed that maybe I wasn't providing enough (though this was low pop systems), or that it needed X days of supply to fill a bucket. From what Ian writes it seems that it shouldn't be time-based, so maybe I just chose the wrong commodities.

I did notice that my Power Generator supply system for the last CG (Hesa) got a Terrorist Attack soon after - though I think that was a coincidence rather than causation (unless ppl think buying goods can fill buckets?)

 

[Ian Doncaster]

From what Ian writes it seems that it shouldn't be time-based, so maybe I just chose the wrong commodities.

It doesn't appear to need more than a day (provided it's not blocked by other states), and in a sufficiently small system it seems to be possible for a single player to cause the effects.

I did notice that my Power Generator supply system for the last CG (Hesa) got a Terrorist Attack soon after - though I think that was a coincidence rather than causation (unless ppl think buying goods can fill buckets?)

I'm not going to rule out that it doesn't, though it seems unlikely.

 

[goemon]

Surely the first thing to do is find what fills the buckets?

I've tried a few commodities that I thought might kick states off (Toxic Waste / Chemicals etc), but never seen anything happen. I assumed that maybe I wasn't providing enough (though this was low pop systems), or that it needed X days of supply to fill a bucket. From what Ian writes it seems that it shouldn't be time-based, so maybe I just chose the wrong commodities.

I did notice that my Power Generator supply system for the last CG (Hesa) got a Terrorist Attack soon after - though I think that was a coincidence rather than causation (unless ppl think buying goods can fill buckets?)

i really think it is not about "buckets" in a traditional sense.

traditional bucket style: trade x tons to fill boom bucket, once filled boom gets pending.
i seriously don't think that's the case for the new states, and the sliders tell us a different approach on the old states as well.

now:
a) somewhat "bucket style": trade x tons, economy slider moves positive until boom is hit, boom gets pending.
b) "probability bucket": trade x tons, probability for pirate attack to happen goes up. might go active or not.

question on b): can that probability be 100%, and is there a fixed amount of x to get it there?

 

[Factabulous]

i really think it is not about "buckets" in a traditional sense.

traditional bucket style: trade x tons to fill boom bucket, once filled boom gets pending.
i seriously don't think that's the case for the new states, and the sliders tell us a different approach on the old states as well.

now:
a) somewhat "bucket style": trade x tons, economy slider moves positive until boom is hit, boom gets pending.
b) "probability bucket": trade x tons, probability for pirate attack to happen goes up. might go active or not.

question on b): can that probability be 100%, and is there a fixed amount of x to get it there?

Surely to investigate either case you need the thing that fills the bucket / moves the slider. No point worrying about which mechanism it is without any idea of how to cause it. At the moment it seems to be a case of 'we did things, and things happened' - zero or more of these may be causative.

Hmm, I was wondering how to earn my ARX this week - maybe time for more experiments

 

[goemon]

Surely to investigate either case you need the thing that fills the bucket / moves the slider. No point worrying about which mechanism it is without any idea of how to cause it. At the moment it seems to be a case of 'we did things, and things happened' - zero or more of these may be causative.

Hmm, I was wondering how to earn my ARX this week - maybe time for more experiments

in most cases we know what moves the sliders though. for boom it's trade and exploration data, and drawing from experiences with dommarraas hydrogen boms, it's tonnage or tonnage per commodity not profit, that moves the economy slider.

we do know, some of what raises the probability of pirate attack, and one action that raises the probability of outbreak (redeeming bonds). we don't really know what raises the probability of infrastructure failure, blight, drought, terrorist attack, and other that effects outbreak. but we do know that states can raise the probability of other states to trigger.

 

[Factabulous]

in most cases we know what moves the sliders though. for boom it's trade and exploration data, and drawing from experiences with dommarraas hydrogen boms, it's tonnage or tonnage per commodity not profit, that moves the economy slider.

we do know, some of what raises the probability of pirate attack, and one action that raises the probability of outbreak (redeeming bonds). we don't really know what raises the probability of infrastructure failure, blight, drought, terrorist attack, and other that effects outbreak. but we do know that states can raise the probability of other states to trigger.

Sliders are completely known - not sure why you mention them - we did experiments 2 years ago moving them around. I'm only interested in the bucket states (i.e. the ones not listed on either slider).

 

[Ian Doncaster]

For Colonia system:

most recent inter-state gaps are:
Jaques: 23, 20, 21, 25, 19, 24 (5 Pirate Attack, 2 Public Holiday)
Council: 29, 25, 24, 23, 22 (2 Pirate Attack, 2 Public Holiday, 2 Inf Failure)
People: 23, 23, 23, 25, 17, 26 (1 Pirate Attack, 4 Drought, 1 Blight, 1 Inf Failure)
Holloway: 30, 27, 24, 27, 20 (1 Drought, 3 Pirate Attack, 1 Public Holiday, 1 Inf Failure)

It's a difficult pattern to get much sense out of, but there's definitely trends there.

Meanwhile in quiet CEI systems with single-figure traffic, there's usually a ~60-70 day inter-state gap on the controlling faction, and incredibly long gaps on the others.

 

[goemon]

Sliders are completely known - not sure why you mention them - we did experiments 2 years ago moving them around. I'm only interested in the bucket states (i.e. the ones not listed on either slider).

yes, i got that. but i don't think they are bucket based in a traditional sense - as explained above.
they trigger imho by chance, and you can raise the chance for them to trigger.
imho this also explains the different times they occur in Ian D.'s exampel.

which would mean - in two identical systems with same starting conditions and same actions applied, they could happen on different ticks, or even the same tick.

you are not filling a traditional bucket, you are upping the chance for them to happen.

 

[Factabulous]

yes, i got that. but i don't think they are bucket based in a traditional sense - as explained above.
they trigger imho by chance, and you can raise the chance for them to trigger.
imho this also explains the different times they occur in Ian D.'s exampel.

which would mean - in two identical systems with same starting conditions and same actions applied, they could happen on different ticks, or even the same tick.

you are not filling a traditional bucket, you are upping the chance for them to happen.

uh huh yes - but either way - without the trigger being known it doesn't matter. You're saying 'maybe this is not a 100% chance?' when we don't even know what 'this' is. It's impossible to disprove your assertion that 'stuff might be just random' - so better to test hypotheses that can disprove it - then at least we have a smaller problem space.

Ah well, 'this' discussion isn't going anywhere - time to experiment