SGOTM 12 - One Short Straw

[bcool]

Also, I gave myself 6 GSpies and used them all to infiltrate Roosy. The effect was to increase the chances of success from 78% to 80%, so I don't think this has a major impact on my testing.

I think there is some maximum chance of success, and so if you have a ton of espionage differential the penalty for having multiple spies appears to go away. If you only have a modest positive or negative differential, you can clearly see the penalty for mission success on multiple spies in the same city.

 

[babybluepants]

So what's your plan?

I'd like to play 3-4t tmr, until we capture Guangzhou, steal Communism, and get close to razing Lyons.

I'll post another PPP in a bit.

 

[babybluepants]

PRE-PLAY PLAN

Spying Questions
You guys are confusing a bit with the ep stuff. Are we putting multiples in Trojan Horse? Are we putting one in each chain city (all the way to Hangz)? When you say back up in St Nick, do you mean sit two spies there, increasing odds of detection, or have one ready to go in after the first one fails?
My understanding now:
7 existing spies, and I'll build 5 more in the next 10-ish turns.
2 in St Nick (existing)
1 in Hangz (existing)
1 in Kam-3S (Kamchatka)
1 in Trojan Horse (Siberia)
More in Trojan Horse (existing + Orleans) How many?
Counter-espionage? China against Roos + Churchill (where?)

Tech
Research Steel - Corporation - Medicine
Steal Communism - Guilds - Banking - Economics - Constitution
I think priorities are: Communism - Steel (cannons) - Guilds (workshops) - Banking (so we can work towards WS as needed) - Constitution/Economics - Corporation (WS) - Medicine
I.E. Focus on beakers towards Steel, then ep towards Roosevelt.

Warring
Move on Guangzhou right away, capture in T+3 or T+4.
Make a deal for a peace treaty b/w DG and Churchill; raze Lyons with trailing units.

Trading / Diplomacy
Roosevelt: switch ep
Roosevelt: Optics for 340g
Churchill: Rice for 1gpt
Churchill: Philosophy for Peace w DG + 70g
Look for resource trades/renogotiations and tech-cash every turn.
Demands: we may as well refuse everything now.

Settlers
Trojan Horse (Siberia 3S) - settler from Siberia in T+5-6
Vladivostok (Kamchatka 3S) - settler from Kamchatka in T+8 - need to raze Lyons
Sledgehammer - settler from Bahamas
Hangzhou-2W - settler from Beijing in ca. T+10 - need to raze Hangzhou by then
Mao iceball - not yet
Island cities - later, as convenient
Copper City - settler from Bahamas (maybe - can decide later)

Cities

Moscow: I still prefer only siege here, but it sounds like espionage is popular. mdy and LC said yes to IA, bcool wants only CH, Mitchum hasn't stated anything as far as I remember.

Fish: finish mace - treb - market (part) - grocer (part) - bank (part) - finish market - finish grocer - bank (3pop whip w Kremlin) - WS. We need Guilds on T190 and Banking on T193 with this. @LC, we can 1pop whip the other two, I'm cool with that. Don't need to decide yet.

GP Farm: finish GT asap; convert grass shops and one sugar into farms as needed; draft every turn for the remainder of TS (Taj GA over ca. T196, we may wanna switch back to Bureau then). Builds: barracks (quick) - National Park (slow for now - reconvert to workshops maybe later, we'll see how it goes).

Siberia: missionary - spy - settler (Trojan Horse) - missionaries for Trojan and Kam-3S - IW

Pigs: I'll hire merchants this turn - may as well take a marginal increase to odds on the next one. Then grow a couple of pop, while hiring max merchants otherwise? Market (slow).

Cuba: finish Theatre - chariot (Kamchatka) - chariot (Bananas)

Bahamas: finish Taj - settlers (Hammer - Spare for copper city - islands...)

Bermuda: finish Harbor - Caravel - Kremlin

Rheims: finish Theatre - Spy - Market
I would like to build a WB here to spy on York/London. What do you guys think of that? Doesn't need to be immediate.

Bananas: Forge - LH/Library

Kamchatka: finish CH & grow to 10 - Spy - Settler

Bcool Island: WB for The Crabbes - infrastructure

The Crabbes: WB - infrastructure

Paris: finish Galley - Forge - Barracks (or skip forge for now?)

Orleans: finish galley - spy - spy - infra

Shanghai: Granary - infrastructure (maybe theatre-lib first, since they're cheap and we might want an option on a fast 500c border pop for that extra fish?)

Beijing: settler (Hangzhou-2W) - Library (maybe also spy, if needed)

St Nick: wealth

 

[mdy]

A couple of minor points that have just occurred to me:

Perhaps it would be an idea to immediately hire 3 priests in Paris. It is the one city where we can guarantee a great prophet, which will probably b the hardest type of GP to get for a 6GP golden age. If we want one from Paris we would need to start building GP points fairly soon so it can eventually overtake the other cities.

Is there any hope of stopping Rooservelt researching democracy? If so I think we should stop trading with him. If he revolts to emancipation the happiness issues it will cause will not be worth a few hundred gold.

 

[LowtherCastle]

Espionage: Clearing Up Some Confusion

One or more spies in the city when we attempt a mission. In the spoiler, I show an example with Our-Total-EPs-Ever equal to Roosy's. Under that scenario, with one spy in Trojan Horse, our success odds are 85%. If we add a second spy, our odds drop to 67%.

Spoiler :

The chance of success for a spy mission, in percentages, is:

((25 * REP) + EIRM) * (100 + DM)) / 100

When one of your spies is on the same tile as another one of your spies, then it has a 15% increased chance of being caught.

REP = Roosy'sTotalEPsEver / (Roosy'sTotalEPsEver + OurTotalEPsEver)
EIRM = 15 (if we have two or more spies on the tile)
DM = Tech = 25

Let's assume our EPs = Roosy's EPs for simplicity's sake, then REP = .5 and we have:

((25 * .5) + 15) * (100 + 25) / 100 is the same as:

(12.5 + 15) * 125 / 100 is the same as:

(12.5 * 1.25) + (15 * 1.25) = so that tells us that the chances of us failing the mission are:

Base chances: 12.5 * 1.25 % = 15.625%
plus
2 spies on tiles: 15 * 1.25 % = 18.75%

So adding the second spy increases the odds from 15% to 33% with 1:1 spending with Roosy.

The attached save demonstrates the difference between 1 spy and two. First click on either spy and check the success odds for Steal Nationalism: 74%. Then move the spy with 0% Stationary Bonus out of the city, click on the other spy and you see that the odds have risen to: 93%. I get the same sort of results with Mitchum's save.

Note that if you move the 50% bonus spy out, then the success odds for the 0% bonus spy remain at 74%. That's because I just worldbuildered that spy in there and on the first turn a spy is able to run a mission, the exact same 15*1.25% factor is added.

 

[LowtherCastle]

BTW, is it just me are do these Christmas colors mess with your eyes...

Yes. Yesterday it annoyed me intesnely because it was hard to think straight. Today, I almost didn't notice.

So if a spy fails, we're out the 40 hammers needed to build him. If we're hammer limited, it makes sense to optimize our chances of success (notice that I didn't say maximize because at some point, it doesn't make sense to do something that only increases our odds by 1%, for example.) If we're epp limited, it makes sense to minimize our epp costs.

So the real question is, are we currently more epp limited or hammer limited? That should lead us to our answer, I think.

Good point.

For starters, you probably noticed that I subtracted the wrong column to get my differences. So they are corrected now and greater than I first reported. Your 72 is a correct calculation, of course. Namking is the cheapest, so that would preferably come in Hangzhou, for instance.

        H-2W    k-3S    Troy  Max
        1.34    1.22    1.12  EPs lost
        ----    ----    ----  --------
Guilds	 437	398	365	-72
Banking	 306	279	256	-50
Econom   612	557	511    -101
Constit	 873	795	730    -143
Corpora  699	636	584    -115
        ----    ----   ----    ----
  Total 2927    2665   2446    -481

So let's analyze this.

Investments for 4 techs stolen:
Three cities
1. Two additional ConfuMiss. Cost: 2*32bh=64bh.
2. Eight spies. 3 in 3 cities, plus 5 to "guarantee" success at 67% odds in Trojan. Since 60% of our spies are being built in non-forge cities, the cost is 32bh*2/5 + 40bh/3/5 = 36.8bh. 8*36.8bh=294bh.
3. Cost in EPs: +50+39=+89EPs for missions away from Trojan.
4. All spies must arrive at their sites 6t before the turn we want to steal the techs.
5. The ConfuMisses must spread religion the turn we want to to steal techs.

Total: 64bh + 294bh + 89EPs

One city
1. No additional ConfuMisses.
2. Twelve spies. 4 techs * 3spies to "guarantee" success at 67% odds in Trojan. 12*36.8bh=441bh.
3. Minimimal cost in EPs.
4. All spies must arrive in Trojan 6t before the turn we want to steal the techs.

Total: 441bh

Comparison:
1. 89EPs versus 83bh
2. 284bh versus 441bh spent 6t or more before steal date.

The loss in spies versus the cost of EPs seems like a wash to me. What seems to be most relevant (independent of this decision, actually) is whether we can get teh spies to their locations on time. That may be a chore for bbp.

As another point, keep in mind that we'll have to hold both Kam 2S and Hang 2W for a turn to get the 50% culture reduction if we plan to steal techs from there, which won't come for free...

I think getting the spies there early enough is far more likely to be an issue.
'

 

[babybluepants]

Ok, so 3 cities sounds better. I can figure out hammers and movement. My only question is the possibility of being caught in China, before we even capture Hangzhou.

The limiting factors seem to be the timeline for researching Steel and accumulating the EP against Roosevelt (difficult for me to estimate) and the timeline for capturing Hangz maybe (if we go for Changdu, incur some damage, etc).

 

[bcool]

Moscow: I still prefer only siege here, but it sounds like espionage is popular. mdy and LC said yes to IA, bcool wants only CH, Mitchum hasn't stated anything as far as I remember.

Only siege in your turn set I think, Courthouse later. LC changed his mind on the IA.

Siberia: missionary - spy - settler (Trojan Horse) - missionaries for Trojan and Kam-3S - IW

You might want to build more spies here (convenient getting them to Troy/Trojan Horse)

Pigs: I'll hire merchants this turn - may as well take a marginal increase to odds on the next one. Then grow a couple of pop, while hiring max merchants otherwise? Market (slow).

we are going to need the market soonish for happiness, might want to run a mine or 2.

Bcool Island: WB for The Crabbes - infrastructure

I would prioritize a courthouse

The Crabbes: WB - infrastructure

I would prioritize a courthouse after granary, lighthouse.

Paris: finish Galley - Forge - Barracks (or skip forge for now?)

Do we even need the Galley? I might start on Barracks now since DeG war will send Paris into unhappiness.

Rheims: finish Theatre - Spy - Market
I would like to build a WB here to spy on York/London. What do you guys think of that? Doesn't need to be immediate.

sounds good, but maybe a galley instead since you wanted an eventual 2 turn galley attack on London?

 

[bcool]

Espionage revisited

2. Twelve spies. 4 techs * 3spies to "guarantee" success at 67% odds in Trojan. 12*36.8bh=441bh.

The last spy has better odds I think the odds are great that we will want to steal five techs.

I bet the easiest (in terms of logistics, hammers, eps) solution would just have 1 spy in Kamchatka 3S and 9 in Trojan.

 

[bcool]

Playing the test game, I think we are going to struggle to get enough and

I'm not sure what the best way to mitigate this is, but I think we need to consider some ideas to speed up the science/espionage.

1) using a great person to help (a spy to infiltrate, a merchant to do a trade mission perhaps) risky since we probably will want this great person for Sushi or a GA

2) Running failure gold for Kremlin, building wealth.

 

[LowtherCastle]

@bbp:
1. I'm in favor of non-stop seige in Moscow. I'm okay with IA if we switch to Bureau this turn.
2. I would have 1 spy in St Nick for Communism and one ready to enter St Nick if the first one fails. The alternative would be to have 3 already there with 50% stationary bonus but we're too late for that. Whatever you do, don't add one at the last minute... ;-)
3. I would station 1 spy each at H-2W and K-3S 6t before you envision stealing Guilds and Banking, even if that means stationing a spy in Mao's culture before you capture Hangzhou.
4. Trojan Horse should probably have ~6 spies stationed in it for 6t fi we want to get all four techs in 1t. Not sure how realistic it is we'll research Steel and accumulate 2000+EPs on Roosy in 10t, though.
5. I guess you could wait to bribe Churchill to peace till next turn when you see into Lyons and see how many defenders there are.
6. The wkr S of Bermuda was heading down tohelp finish that workshop at Bahamas. Just a thought.
7. I like the idea of spying on London/York a bit.
8. GP Farm: I don't see the need for barracks anymore, since we're planning to put the hammers into NatPark rather than extra units.
9. Paris: Barracks + units makes sense here, since you feel we should have some more units.
10. Note that on Guanghzhou, after suiciding several trebs, you may get to a point where the CRIII maces have ~95% odds, while the trebs are still down at 65-70%. I took out a couple semi-healthy lbms that way, then finished red-lining the rest of the lbms at much safer odds with the trebs.

Sounds like you've got things worked out, bbp. Good luck!!!

lots of xposts!!!

 

[LowtherCastle]

Do we even need the Galley? I might start on Barracks now since DeG war will send Paris into unhappiness.

The garrison axe will help a lot immediately.

sounds good, but maybe a galley instead since you wanted an eventual 2 turn galley attack on London?

How in the world can we manage a 2-turn galley attack on London?

My only question is the possibility of being caught in China, before we even capture Hangzhou.

We need spies in the area anyway, garrisoned in Beijing, etc. They back him up.

The limiting factors seem to be the timeline for researching Steel and accumulating the EP against Roosevelt (difficult for me to estimate) and the timeline for capturing Hangz maybe (if we go for Changdu, incur some damage, etc).

1) using a great person to help (a spy to infiltrate, a merchant to do a trade mission perhaps) risky since we probably will want this great person for Sushi or a GA

2) Running failure gold for Kremlin, building wealth.

I'm okay with building wealth instead of infrastructure in many of our newer cities. Just not at the expense of:
1. EPs (work our spy specs!)
2. Tech stealing preps
3. War preps
4. circumnavigation
5. GT+Taj+GP spawning

We can also runs lots of merchants and scientists in various cities, rather than growing right now. Our cities will have plenty of food and hammers soon enough.

 

[LowtherCastle]

Moscow should work sci or eng, if at all, but not merchants, imo.

 

[babybluepants]

How in the world can we manage a 2-turn galley attack on London?

Why not? It would be fun. Keep with our galley strike theme...
Not sure it's possible.
We have all the galleys already, without much use for them anymore. I think we may be able to build up a stack of ca. 20 units, and transfer them in via 10 galleys. If it worked, it would be worth waiting some extra turns for build-up, instead of attacking earlier via Tours.

 

[Mitchum]

Sorry about focusing on espionage only for the last several days. With company in town, that was the only topic I felt I could contribute on based on my previous testing without opening the game and spending too much time on this at the expense of my visitors.

Excellent discussions, team. It looks like bbp is ready to go. I think the PPP looks good. This is the most important turnset of the game, after all...

 

[babybluepants]

Ok, I'll skip the galley in Paris. Don't really know what to do with all the existing ones anyway...

I can skip the barracks in GP, too, except I always figure it's too cheap to be worth skipping if I'm getting units there (even drafted).

I'm in favor of non-stop seige in Moscow. I'm okay with IA if we switch to Bureau this turn.

Siege it is.

I would have 1 spy in St Nick for Communism and one ready to enter St Nick if the first one fails.

Ok, that's what I figured.

Not sure how realistic it is we'll research Steel and accumulate 2000+EPs on Roosy in 10t, though.

Playing the test game, I think we are going to struggle to get enough and

Not sure what to do about it. I was gonna re-evaluate the situation after the GA starts and we switch to State Prop. Too complicated to calculate now.

Don't mind building some wealth, but don't think we can in any cities with significant hammers. I'll see what I can do with that and running some more specs...

I guess you could wait to bribe Churchill to peace till next turn when you see into Lyons and see how many defenders there are.

That was my thought, as well. Except, I don't know anything about bribe mechanics. Any chance that his price would go up, if he moves his stack into attack position?

The wkr S of Bermuda was heading down tohelp finish that workshop at Bahamas. Just a thought.

Yeah, I was gonna do that.

 

[LowtherCastle]

Why not? It would be fun. Keep with our galley strike theme...
Not sure it's possible.
We have all the galleys already, without much use for them anymore. I think we may be able to build up a stack of ca. 20 units, and transfer them in via 10 galleys. If it worked, it would be worth waiting some extra turns for build-up, instead of attacking earlier via Tours.

What I meant was, where will the galleys be before the DoW so they can reach London in 2t?

 

[mdy]

I'm in favor of non-stop seige in Moscow.

If we do this I think we are going to have far too many trebs. If we build a courthouse and intelligence agency in Moscow are treb situation after 10 turns will be as follows:

Current trebs: 16
Trebs built in Moscow 3
Trebs built elsewhere:1

Total 20.

My worldbuilder tests indicate that even if China whips an extra longbow in every city our expected treb losses in taking Guangzhou, Chengdu, Huangdcho and Lyons are about 8 (unless there is an unexpectedly large stack in Lyons).

This leaves us with 12 trebs which is enough to start the Churchill campaign without slowing down, especially as we will want to use cannons as much as possible, and we should be able to start building them after T+10.

If we delay the Churchill war slightly to enable an attack by galley (a good idea in my opinion if we can do it) on London then we would want cannons on the galleys, not trebs as much as possible-another argument against overbuilding trebs.

 

[babybluepants]

What I meant was, where will the galleys be before the DoW so they can reach London in 2t?

South of Rheims via some forts, I thought. We can do it in 2t as it is, without further English border. We could also get circumnavigation by then.

 

[bcool]

If we have too many treb, we can upgrade them for less gold (80) than we can build another cannon (100 hammers). This means that wealth building in any city other than Moscow would get us a cannon faster than building it ourselves.
I think this is worth it because we can
1) build wealth in a number of cities for a turn or 2 to upgrade them.
2) It reduces loses and therefore war weariness.
3) It reduces unit maintenance because we will need fewer units overall.
4) It saves time on the conquest because we have cannons where we need them rather than waiting for them to come from Moscow or somewhere else.

Lastly I think the numbers LC used to compare cannon to treb chance of losing against protective longbow on flat ground were wrong or at least misleading. We will occasionally face CGIII longbows and longbows without all the fortification bonus and in those cases the cannon has a much better survival chance.