Deity Shadow Game

[bigbamboo]

Do you think the Chinese community's emphasis on early warfare is hurting their deity difficulty win rate?

Don't get me wrong, their approach is definitely optimal on some maps. I have seen some videos, they're impressive in their own way.

For other maps, would "our" usual strategies of curaissers and cannons be the more suitable after all?

As for the Cuirassier and Cannon Rush strategies, our community abandoned them eight years ago after our 'guru' developed the Catapult Rush. Through countless experiments, we’ve found that any map that can be beaten with a Cuirassier or Cannon Rush can also be beaten with the Catapult Rush—and in fewer turns. Of course, this applies to Pangaea maps; on more challenging Fractal or continent-split maps, we still use Cuirassiers.

On Pangaea maps, almost 50% of the time, you might only be able to build 4 or fewer cities, which could explain the inconsistent win rates. That’s why we need earlier wars to gain an advantage over the AI

 

[notslin]

I noticed in your attempt you declared on Turn 110, and even on Turn 127 when Asoka capitulated, he didn't have Monarchy(!). In my last attempt, Asoka got Feudalism around Turn 108. This made it more difficult to get his cities, and I was unable to get Feudalism in time to capitulate him. Wouldn't you say Turn 110 is a bit risky because the AI can have Feudalism at that point? Overall, it seems challenging to get a stack before Asoka has longbows. That said, I did place my cities incorrectly in my last attempt, which may have cost me or turn or so.

 

[bigbamboo]

I noticed in your attempt you declared on Turn 110, and even on Turn 127 when Asoka capitulated, he didn't have Monarchy(!). In my last attempt, Asoka got Feudalism around Turn 108. This made it more difficult to get his cities, and I was unable to get Feudalism in time to capitulate him. Wouldn't you say Turn 110 is a bit risky because the AI can have Feudalism at that point? Overall, it seems challenging to get a stack before Asoka has longbows. That said, I did place my cities incorrectly in my last attempt, which may have cost me or turn or so.

First, in 50% of cases, the 'Catapult Rush' won’t face enemies with Feudalism, and even if they do, it’s not a huge problem—especially since we have War Elephants. Unless the opponent is Protective, we’ll avoid them and target someone else.

Second, I missed a detail in the screenshots: by the time I launched my attack, my military strength was close to 90% of the world leader’s. I could have extorted almost every AI, but I didn’t need to because 4 AIs were already annoyed with Asoka, and only Elizabeth was Pleased. (Note: Elizabeth only joins wars if she’s Friendly.) Here’s the key detail: when I captured Asoka’s capital, I estimated India’s remaining land and realized it had reached the threshold for Elizabeth to vassalize them. So, I extorted 20 gold from her to force a 10-turn peace treaty, which allowed me to peacefully vassalize India

 

[bigbamboo]

The more I think about, the more I think there was an error on turn 1 in this game, after moving settler 1E.

I think settling on the coast like @bigbamboo did was better.

Consideration 1. You gain +1 C on capital tile, but lose the plantation bonus yield, which is +1 F, +2 C, which is +2 C over a farm (until biology).

In our community, when it comes to return on investment, we’re enthusiastic about investments that pay off in 20 turns, cautious about those that take 30 turns, and dismissive of anything that requires 40 turns. That’s why we prioritize faster early expansion. In this game, because I built the Great Lighthouse (due to not scouting other AIs), I only had 4 cities and 4 Workers by Turn 70. If I hadn’t built the Great Lighthouse, as an Expansive leader, I should have had at least 6 cities and 7 Workers by that point.

For example, I decisively and permanently gave up the Crabs because I needed a better spot for my first expansion to secure the Great Lighthouse earlier. I wasn’t in a rush to develop the Flood Plains—that would have delayed my Workers, who needed to chop more forests for Granaries, Settlers, and additional Workers

 

[bigbamboo]

I think it's very good and shouldn't be dismissed.
The players referenced above by Sampsa, that have excellent winrates, are part of a 1% and 3/4 of them have never offered proof of no reloads. There are others who just don't play anymore.

Drewifast is entirely right that a specific strategy (aka Construction Rush) should always be situational and part of your repertoire. Strategies should be varied.
The map should dictate the strategy. However, the earlier you can enforce a strategy, a) the greater results you can achieve, but also b) the less variance you're subject to.

I'm partial to high science strategies (or HA rushes) but it isn't clear to me they're more "reliable" than, say, a Construction rush (low science strategy).
What I can say with certainty is that achieving early map control is very safe - and a Construction rush can do that. Many have been declared war upon while researching towards Military Tradition. It is delusional to say that Lib -> Mil Trad is fool proof.
There's also : just being boxed in and out-teched is a terrible way to die. Better choose your own end

I’m glad we’ve reached a consensus on the importance of early map control. Eight years ago, we also relied heavily on Cuirassiers and Cannon Rushes. Now, we call that the 'Holding It In' strategy! Imagine you’re getting intimate with a girl, spending a long time on foreplay, and suddenly—knock, knock—her father comes home.................

We used to suffer that kind of frustration all the time eight years ago. That’s why the arrival of the 'Catapult Rush' felt like a divine light breaking through the clouds

I realized I overlooked something: another 20% of our win rate comes from the Horse Archer Rush, and 5% comes from even earlier tactics like Axeman, Swordsman, or Chariot Rushes. The actual win rate for random maps opened live by referees is around 50%

 

[notslin]

One nice thing about settling the capital on the coast -- I didn't mention this earlier -- is it gives you the worker 2 turns faster since it has access to the plains hill forest after the border pop!

 

[Snowbird]

If a so-called experts have a 90+ winrate and community average is 20... then what is non-expert winrate? Under 10% probably. That is not something to be proud off, especially given easy map script. In general, knowing mapscript alone gives player an edge. In normal play iso is so bad, because it takes time before player realizes he is isolated, for example.

I am also not sure what all the fuss and self-proclaimed expertise is about.

Does military give human edge in this game? Sure. Pangea is also a map where one can rush as much as he like. Close AI, selection of targets, no running out of trade partners.

Are war elephants good units? Sure thing. Comes at same tech (construction) as catapults, and early on can be traded for at semi-reasonable price. They also come in clusters and all the clusters would be on the same continent - since it is pangea. I think older civfanatics threads have multiple examples of beating AI with loaned elephants.

If players spend time for specific strategies, they can achieve good results since the base game is flexible enough. IIRC, obsolete had very good results with wonderspam mega capitol - as an extreme example. So if one spends a lot of time beating specific maps, sure they can be efficient. Which, IMHO, does not prove a lot in general. If someone showed a simplified strategy which worked on 80% of tectonics map or reliable way to beat deity iso, that would be interesting..

 

[bigbamboo]

If a so-called experts have a 90+ winrate and community average is 20... then what is non-expert winrate? Under 10% probably. That is not something to be proud off, especially given easy map script. In general, knowing mapscript alone gives player an edge. In normal play iso is so bad, because it takes time before player realizes he is isolated, for example.

I am also not sure what all the fuss and self-proclaimed expertise is about.

Does military give human edge in this game? Sure. Pangea is also a map where one can rush as much as he like. Close AI, selection of targets, no running out of trade partners.

Are war elephants good units? Sure thing. Comes at same tech (construction) as catapults, and early on can be traded for at semi-reasonable price. They also come in clusters and all the clusters would be on the same continent - since it is pangea. I think older civfanatics threads have multiple examples of beating AI with loaned elephants.

If players spend time for specific strategies, they can achieve good results since the base game is flexible enough. IIRC, obsolete had very good results with wonderspam mega capitol - as an extreme example. So if one spends a lot of time beating specific maps, sure they can be efficient. Which, IMHO, does not prove a lot in general. If someone showed a simplified strategy which worked on 80% of tectonics map or reliable way to beat deity iso, that would be interesting..

Pangaea isn’t a map where you can do whatever you want. When we don’t handpick maps, we often end up with strange layouts—like spawning in the middle of the map with only room for 3–4 cities and bordering at least 5 AIs. There’s no stable period for development in such cases. While IOS maps are tough, and 2-5 Fractal splits can be challenging, I don’t think other Fractal maps are harder than Pangaea.

Statistically, most Fractal maps should be 3-4 or 4-3 splits, with 5-2 and 6-1 splits being even easier. Even on Island maps or 2-5 Fractal splits, we can easily achieve a Cultural Victory (unless Cultural Victories aren’t recognized?

As for trading with more AIs, our strategy doesn’t rely heavily on tech trades. We don’t need to lead in tech; as long as we have gold, we can research key military techs ourselves or direct our vassals to do so. Most of the time, we don’t even hit the tech trade limit. On Pangaea, diplomacy can be more complicated, so we use our strength to pressure the AIs instead of relying on tech to influence them. We’ve even tried Pangaea games with all trading disabled, and it actually became easier because the AIs couldn’t trade techs with each other either

PS: You’re missing the core point. It doesn’t matter whether we use War Elephants or other units to defeat the enemy—as long as we have siege weapons, we’ll find the timing window and make every action serve that goal until we achieve it. If you’ve watched StarCraft 2 matches, you’ll understand what I mean

 

[Snowbird]

So, now you use siege for cultural victories. I see the attacko was reborn in China.

 

[Tecumseh1]

It's a good discussion but can we keep it maybe in the other thread? As Simple Machine is trying to get feedback on a shadow game here

 

[sylvanllewelyn]

In terms of feedback, I think elephants/catapults (elepults) are the strategy. I think bigbamboo illustrated the optimal approach for this map.

 

[bigbamboo]

So, now you use siege for cultural victories. I see the attacko was reborn in China.

No, no—besides the Catapult Rush, we also have the Horse Archer Rush (which we use even more often) and the 'Chariot Scratch-Off' (a strategy that uses Chariots to suppress AI development, preventing their expansion and access to military resources. If there’s no opportunity to capture cities, we pillage tiles for gold to fund research until we unlock key military techs). Of course, we also go for Cultural Victories, as our community competitions usually aim for the shortest turn count.

As Sun Tzu said: 'The supreme art of war is to subdue the enemy without fighting.' And another quote: 'Military tactics are like water; water shapes its course according to the nature of the ground, and a soldier achieves victory by adapting to the enemy.' — The Art of War

 

[bigbamboo]

In terms of feedback, I think elephants/catapults (elepults) are the strategy. I think bigbamboo illustrated the optimal approach for this map.

Yes, the 'Catapult + Elephant' combo is a variant of the Catapult Rush. We’ve even developed an extreme 3-city Catapult strategy, using an early Great Scientist to bulb Mathematics, quickly research Construction, and crush a neighbor by Turn 90 with Catapults. The same tactic has different pacing variants depending on the situation

 

[sylvanllewelyn]

Okay, you like catapults, I understand.

https://forums.civfanatics.com/ threads/sampsas-huge-epic-deity-ntt-novassals-no-fail-gold.682784/

The context is different but the idea is similar.

 

[pmarc]

@bigbamboo

Not sure if it would have helped in your case here, but if you look at the espionage spending of the first AI you meet and they don't have 4 espionage points on you, it means they met someone else.

As for going to Cuirassiers / Cannons without being attacked first, you'll find that the community here has spent a lot of time researching the working of the AI personalities and exploit that information in diplomacy (see e.g. https://forums.civfanatics.com/threads/civ-illustrated-1-know-your-enemy.478563/) Of course, if you're neighbor with a more aggressive AI or cannot expand beyond 4-5 good cities a catapult-based attack is probably still the favorite option.

 

[pmarc]

I followed the OP's initial scouting move but did not look at spoilers beyond that.

I was convinced by @Fippy's argument for starting with Pottery, and decided to try settling on the rice with first expansion on the crab PH.

The CHA-EXP combination feels very synergistic with this opening. Not only do we have cheap granaries but the extra happiness compensates a bit for the slower initial expansion from delaying bronze working.

Spoiler turn 37 :

Just finished Bronze Working, worker farmed a floodplains and cottaged 3, then build a road to connect next city, not wasting any turns. Build a warrior while growing to 4, then settler. After settler is done I will whip a worker in capital with overflow going into the workboat. I also need to chop forest for granaries in both cities, and then get 2 more settlers going ASAP.

Spoiler turn 67 :

Settled 3rd city on turn 63, a bit late but the research is going along very well (1 turn to Masonry, then on to Construction). I put the settler up in the NW on the elephants to make sure I got the city before Asoka. However this is really bad for the economy and if I was sure I had metal by then, I would have conceded the elephants in favor of an earlier catapult rush. Washington is building a settler for the pig city. Only buildings are two granaries in Washington and New York. I will build barracks as well before catapults, no time for library though.

Spoiler turn 93 :

Just finished Horseback Riding and traded both it and Construction to Brennus for Alphabet and Iron Working. Nobody would trade techs until now, Tokugawa had Alphabet earlier but he's Tokugawa (lol). I have a stack of 8 catapults north of Philadelphia (2 more almost done) and have started producing elephants.

 

[pmarc]

The same logic applies to Stables. The 'Catapult Rush' starts around Turn 110, while the 'Horse Archer Rush' begins between Turns 70 and 85. The +2 experience from Stables is crucial at this stage, as we haven’t yet unlocked Feudalism for the additional +2 experience. Mounted units with 2 promotions, such as Combat I and Shock, gain a significant advantage

That's interesting, I don't like gambling on horse archers on high difficulty because they sustain such heavy losses against cities with hills, walls or both. Maybe the promotions make a difference in some cases?

 

[Tecumseh1]

That's interesting, I don't like gambling on horse archers on high difficulty because they sustain such heavy losses against cities with hills, walls or both. Maybe the promotions make a difference in some cases?

I find you just need more numbers. That's more important than promos. Maybe 1 stable for a few super units. And yeah hills suck. Protective AI is even worse and might make me wait for catapults. But with HA you have to attack early. Hard on deity to pull off but very reliable on immortal.

 

[Earendil2024]

That's interesting, I don't like gambling on horse archers on high difficulty because they sustain such heavy losses against cities with hills, walls or both. Maybe the promotions make a difference in some cases?

I have different opinion.Horse archers performance quite better in Diety.It still has quite large rate to win against spearman in open field.
AI still try to expansive during 80-90t.If u start war with 10-15 horse arches during this time u can easily capture 1 or 2 city and eliminate their main force in open field.ai dont have enough techs to buy others or to early to get Feudalism to surrender.
if u can cut off ai's copper and iron and bypass their hill city with lots of spearmans ,u can easily capture most of other cities in quiet short turns and take large techs or money when cease fire.

 

[bigbamboo]

That's interesting, I don't like gambling on horse archers on high difficulty because they sustain such heavy losses against cities with hills, walls or both. Maybe the promotions make a difference in some cases?

The Horse Archer Rush exemplifies the fluidity of early-game warfare, mirroring strategic principles from the Art of War. For instance, when facing a mountain city with excessive AI garrison stacks, a critical mass of Horse Archers can execute flanking maneuvers to sever supply lines and split enemy forces. If a heavily fortified city proves impregnable, bypassing it becomes the optimal play — after all, the AI’s garrisons are finite. Overcommitting here leaves other fronts vulnerable.During deep raids, Horse Archers often provoke garrison rotations, forcing the AI into mobile warfare on open terrain. On flat ground, disciplined micro can achieve 3:1 kill ratios against Spearmen — quantity trivializes their anti-cavalry bonuses, especially when victorious armies heal while routed armies are obliterated.If the enemy retreats into late-stage citadels guarded by high-defense units (e.g., Pikemen, Crossbows), bait tactics using Worker decoys, Chariot skirmishers, or injured Horse Archers can lure them into ambushes. The Horse Archer Rush thrives on adaptability — I’ve salvaged seemingly doomed campaigns through meticulous tile control and priority targeting. For demonstration, If you are interested,I can share replay footage from my livestreams.

I have different opinion.Horse archers performance quite better in Diety.It still has quite large rate to win against spearman in open field.
AI still try to expansive during 80-90t.If u start war with 10-15 horse arches during this time u can easily capture 1 or 2 city and eliminate their main force in open field.ai dont have enough techs to buy others or to early to get Feudalism to surrender.
if u can cut off ai's copper and iron and bypass their hill city with lots of spearmans ,u can easily capture most of other cities in quiet short turns and take large techs or money when cease fire.

Your analysis resonates deeply — clearly, you’ve mastered the Horse Archer meta. Indeed, Sometimes, well-honed tactics can be more effective than statistical projections on paper.2-tile mobility is the linchpin: it enables rapid force concentration to create local numerical superiority (e.g., 3v1 flanking strikes), while forcing the AI into disastrous reinforcement marches across open terrain. Each time they shift garrison stacks, we ambush them mid-movement — a direct nod to the Art of War principle: "Attack where they are unprepared; appear where you are unexpected." Very impressive. You've touched upon some profound tactical thinking


我专为一，敌分为十，是以十攻其一也，则我众而敌寡 ———— 孙子兵法
Focus your entire force on one key point, while scattering the enemy's strength across ten different fronts. This turns every battle into a 10-to-1 fight – you outnumber them locally, even if they're stronger overall -—————— The Art of War