[SPOILERS] Woden and ljubljana like boats

[Woden]

Here is what I have set up for Russia...
   

I guess I wouldn't call it really a trap but I gave him options. My hope is that he wants to get my frigates out of the game because he has to know I am getting close to Refining. I have had aluminum for a few turn now. Anyways, I have a single frigate on the harbor at Loki and moved a fleet ironclad to protect my shipyard. The options he has is going for the frigate at Loki or going to the trade routes, where I have 4 fleet ironclads waiting. Regardless of what he does (if he moves forwards and does something), I have more ships waiting in the wings that can help in an attack. The worst thing (for us) he could do here is just sail over to England. Luckily, I have a bunch of frigates that will be battleships before he can get over there. 

If there are a large contingent of ships by my marker, I might have to take those out first before sailing north. I would hate to engage his northern fleet and get slammed from behind. The good think is each fleet ironclad is worth 4 privateers. The bad thing here is that he might have 5 per ironclad. 

In other news, as long as I can get 1 of 2 quarry/pasture pillages next turn, I should have enough faith to buy the GA. I finish Natural History next turn and will start whatever civic is in front of Mobilization, unless I find the civic for Fascism.

Edit: I forgot to mention, I am bringing one of my field cannons and my new General over to help defend your campuses. He should have an attack of 90 if I put him on a district. Are one of those campuses on a hill?
Edit2: I have 5 ironclads and 2 frigates finishing next turn and that will be the last of my frigates before Battleships (2 turns). The good thing about using all my coal is that I could make more frigates once I run out.

[Woden]

Got stopped in my tracks by a BAD lowroll, doing only 95 damage on an attack with +35 combat strength, again well below the minimum of 97 according to the combat formula from civwiki . That attack should have had ~90% odds to kill, but so it goes... At least it happened towards the end of the turn and not right at the beginning, I guess. As a result two frigates are alive that should not be (albeit one deep in the red), and my line is messy and has one fleet out of GA range who should have been the one to kill the second frigate. So I got 7 caravel kills (including 2 fleets) but ZERO frigates, and am now about to take the full brunt of...whatever frigate stockpile Australia still has left.

Definitely predicting quite a few deaths, but at least our line is relatively safe (other than vs any direction besides the south, of course ). Most likely I will lose 4-5 caravels (which is all my caravels!) but retain the frigate core mostly intact. Hopefully I will then be able to do a "kill all the frigates" turn next turn to make up for failing to do so this turn... The one nice thing about this mess is that the open position somewhat favors us now, as they most likely do not have enough caravels left to form enough of a line to protect the frigates.

Tough luck on the bad dice roll. I actually think he is going to attack my fleet ironclad. If he does, it will pretty much take most of his ships. I attacked the his fleet caravel with a 17(?) CS advantage and I think he can get a maximum of +4 support bonus on his attack, meaning it should be a 21 CS disadvantage for him on his first attack but get smaller as the ironclad gets damaged. He will probably be able to kill it but then it allows your ships to do maximum damage again.

[Woden]

Got stopped in my tracks by a BAD lowroll, doing only 95 damage on an attack with +35 combat strength, again well below the minimum of 97 according to the combat formula from civwiki . That attack should have had ~90% odds to kill, but so it goes... At least it happened towards the end of the turn and not right at the beginning, I guess. As a result two frigates are alive that should not be (albeit one deep in the red), and my line is messy and has one fleet out of GA range who should have been the one to kill the second frigate. So I got 7 caravel kills (including 2 fleets) but ZERO frigates, and am now about to take the full brunt of...whatever frigate stockpile Australia still has left.

Definitely predicting quite a few deaths, but at least our line is relatively safe (other than vs any direction besides the south, of course ). Most likely I will lose 4-5 caravels (which is all my caravels!) but retain the frigate core mostly intact. Hopefully I will then be able to do a "kill all the frigates" turn next turn to make up for failing to do so this turn... The one nice thing about this mess is that the open position somewhat favors us now, as they most likely do not have enough caravels left to form enough of a line to protect the frigates.

Tough luck on the bad dice roll. I actually think he is going to attack my fleet ironclad. If he does, it will pretty much take most of his ships. I attacked the his fleet caravel with a 17(?) CS advantage and I think he can get a maximum of +4 support bonus on his attack, meaning it should be a 21 CS disadvantage for him on his first attack but get smaller as the ironclad gets damaged. He will probably be able to kill it but then it allows your ships to do maximum damage again.

Actually, looking at it again, I hope he rolls west and attacks your ships. That will open a path to pillage a mine along his coast. Then I can finish Refining next turn! Both ironclads can reach his coast with enough MP to pillage.

[ljubljana]

I don't know how large williams' privateer blob is, but they have added...1000 milpower in the last 6 turns . If that is privateers, it would be what, 20 privateers? They are fast too, as the card gives them +2 movement, so they can keep up with your ships. We are in real danger of meeting our end under a thrawn-style zerg rush here... Your ships will one-shot every time with the GA, even unmerged battleships, but if they can pick off your IC in small groups of 5-6 at once you might get ground down. I think you might have to bide your time for a bit until the attack force is ready and massed, even if we lose some harbors, so that they can't destroy your force piecemeal. I do think there are probably not too many privateers at your pinned location quite yet, as it is just too far from their core for a force that was produced literally last week. But I am sure they are working on it and maybe they do have an advance guard that is there already.

Of course, if even a fraction of those ships are on the Australia front, our attack there is doomed  If I can get ONE turn without a bunch of ships popping out of the fog, I can probably wipe their frigates and we will be in ok shape, but I have a horrible feeling that's not going to happen

By the way, Australia is zerg-rushing too, their privateers had +2 movement this turn. Probably both of them are making them in literally every city.

The northernmost campus is on a hill, yes

Best guess based on the last time we fought is that it will take 5 attacks from williams to kill each of our ships - last time they used 2 frigates + 2 caravels against my frontline ship and came up just short. They look to have about 17 or 18 attacks this turn, with a few fleets mixed in, so I guess they probably will kill 4 things unless they stall by not attacking with caravels or don't have enough extra frigates in the fog. One kill will surely be your fleet IC, the rest probably caravels unless they can connive a way to get at a frigate. If your other IC survives, it will be a tough call on whether to pillage or keep it safe so you have something to capture the cities with...

I should still have enough mass to do good damage this turn unless Russian ships appear...but come on, it takes just 4ish turns for newly produced ships to get here from Russia and I did telegraph this push pretty far in advance. Of course Russian ships are going to appear

[ljubljana]

On the other hand, they sure are talking a lot for someone who has a wave of privateers in the fog ready to obliterate me...

[Woden]

I did kill 2 last turn that I stumbled across by accident. One was down in the southern ocean that must have just been built and was trying to hide from my ironclad. It just happen to be where I needed to stage for the camp pillage on my turn. The other one was about 4 tiles north of (W)Odin and I pushed a fleet ironclad out to see if there were ships and ran into his ZOC. If he has a group of them up there, it might not be there when I open the save. Not the best move but what can you do.

He also has a few out of his southern ocean cities (the one with the GL) and they have nicely been giving my ironclad promotions. It is kind of weird that he is attacking that ironclad because there is fishing boats down there that I used to heal and last turn I was able to promote and this coming turn I can grab another fishing boat. I am not going to complain as the ship takes 5 hits (2 city/encampment attacks, 2 privateers, and an archer) per turn for less than half its health. I should be close to a 3rd promotion by the time he bugs out.

I didn't realize his MP grew 1,000 in 6 turns. I bet he is cranking those out every couple of turns in his most productive cities. Makes me think we should wait for destroyers before launching a massive attack but those are about 10 turns away, les if I can get some mines pillaged. We will have to watch out for subs to if he hooks up oil. We won't be able to one-shot those.

Edit: With the 7 new ships next turn plus upgrades and a few more ships the turn after, I should close, if not over, 3,000 MP, which would be about 1,0000 in 6 turns.

[ljubljana]

I don't know about waiting, it gives us time to hit destroyers but them time to hit battleships. Maybe the way to go is piecemeal - first kill the northern fleet before the frigates upgrade, then upgrade destroyers, then clear the privateers? At least they will not be able to upgrade any privateers they have wandering around your cities, unless they want to try their hand at cracking Renaissance Walls.

[Woden]

No, I don't think we should wait either. I am going to have to be careful gathering my forces and moving north. I would hate for him to bring the pack at Freya over to combine with whatever is outside of Geneva and pick off battleships as I sail them north.

We must have done something right if they started chatting soon after roland got the save and now he can't play until tomorrow.

[ljubljana]

Found a reddit post that contains what I believe is the actual damage formula. Civwiki says it is

30 * e^(D/25) * random(0.8, 1.2)

where D is the combat strength difference, but the reddit post says it is

30 * e^(D/25 * random(0.8, 1.2))

which produces a much wider range, especially at the high end. This is consistent with my observed sense that displayed one-shots from 100% health are much less trustworthy than displayed kills at half-ish health, and gives a low-end roll of 91 damage and odds to kill of ~85% (random value at or above a 0.86). With that formula, one-shot kills are actually not guaranteed until 38 CS differential!

Quick rule of thumb calculations for future reference (eg next turn):

D  |  roll  | odds
30 | 1.01 | 49%
31 | 0.98 | 56%
32 | 0.95 | 63%
33 | 0.92 | 71%
34 | 0.89 | 78%
35 | 0.86 | 85%
36 | 0.84 | 90%
37 | 0.82 | 95%
38 | 0.80 | 100%

D = strength difference, roll = minimum roll required to kill, odds = approximate probability. I assumed for simplicity that the RNG chooses uniformly with precision up to only two decimal places, but these odds should be a small enough underestimate that the effect size is swamped by rounding the percentages.

I sure wish this were less random! Lesson learned - don't make battle plans that require one-shots as cavalierly as one might with two-shots. And if you do, make sure you get them! There was a different attack order that I could have used that would have killed the second frigate (with 95% odds) before taking this shot, boosting me to +37 CS by removing a support bonus, but I rejected it because it was guaranteed to leave a frigate out of GA range. My roll was about a 0.83, which would have killed if I had gone with that plan...

For reference, my frigate fleet shots against roland's frigates are at +37 CS now, which is right in the sweet spot of almost certainly one-shotting. But with WoR and a spy up, those odds would drop all the way down to 50/50! This is another great illustration of why every single point of CS is worth scrapping for...and indeed, why that emergency was so damaging. I would have killed if not for that, too, and blocking it would have earned me another missed kill at the start of the turn too for a (theoretically-optimal) 9 ships sunk instead of the 7 that I got...

edit: My optimistic guess is they were not expecting the first strike and are debating how to respond and whether to send Russian privateers over, which would be a tough call as the fight will likely be over in just 1-2 turns so they would probably not get there in time. Or, at least, I think my Bayesian priors are shifted slightly in that direction by all the talking. But they could just as easily be discussing the minute tactical details about how to most efficiently kill us with their hidden privateer wave...

[ljubljana]

...huh! Well, I guess the precise combat formula has not fully been reverse-engineered yet? The only sourcing I can find for the civwiki one derived it as a heuristic that fits the curve well, and it looks plausible both mathematically and from a coding logic perspective...but, at least in regimes of very high and very low CS differential, neither plausible interpretation of the civwiki formula fits the data. Or my math is wrong, as is always a live possibility.

So my new table is as follows:

D   | roll             | odds                  
35  | pretty high | don't count on it
37  | 0.8             | 100%