Civ4 AI Survivor Season 4: Championship Game Writeup


For the Championship Game of Season Four, I asked if anyone was willing to create a custom map and Commodore volunteered to put one together. Commodore created a far superior version of a mirrored map as compared to the Donut one that I had used for the first three seasons of AI Survivor. This was a true mirrored map with exactly identical terrain for all six leaders, looking very similar to the setups that we use at Realms Beyond for our Multiplayer games. This was one game where there would be no excuses about some AI leaders rolling better land than others. The individual capitals therefore all looked identical to one another; here's the image of the central jungle region since it fell outside of the starting positions:

Note that the terrain was again mirrored here, with the Calendar-based happiness resources in the center of the map located at the same distance from each capital city. Commodore did not run any rivers through the center of the continent and it could be some time before the AI leaders managed to tame the wilderness in there. We expected the northern and southern sides of the map to remain divided from one another during the early turns of the game.

It was fascinating to watch how the AI leaders chose to settle on this map. Faced with identical terrain and identical starting units, the six leaders mostly mirrored one another in their performances. Everyone sent the initial settler to the west out of the gate, and five of the six leaders chose the same spot to within one tile of each other. Kublai Khan was the only one who broke the pattern by settling along the coastline instead of picking a location further inland like the rest. For their third city, the first one to be built with a hand-trained settler, most of the leaders pushed further to the west beyond their second city, then returned back to the east again with their fourth city. In fact, the three northern leaders of Willem, Gilgamesh, and Darius picked out completely identical locations for their first four cities:

It was slightly eerie how those three leaders were mirroring one another thus far. I suspect this was due at least in part to the fact that all of them had started the game with Agriculture tech and were able to connect the wheat tile at their capitals immediately. The south had more variation between Kublai Khan, Charlemagne, and Stalin in terms of where the initial cities were located. Kublai had the aforementioned Beshbalik coastal spot for his second city, then picked sites to the northeast and southeast for his following cities, before becoming the first leader to push into the contested central jungle region with his fourth city. Charlemagne wasn't grabbing as much land as expected despite being the only leader in the game with the Imperialistic trait, probably because he was very late to research The Wheel tech and had no roads at all in his territory. This meant no resources and cities that couldn't grow in size; it was a bit of a pathetic spectacle. And he was still doing better than Stalin, who was struggling out of the gate for inexplicable reasons. The most likely explanation was the presence of several nearby barbarian cities, two to the east of Russia and another to the west, which may have tied up Stalin into barbarian-mode during these crucial starting turns. Russia was forced to settle a fourth city in the jungle in a very weak location and Stalin was trailing behind the rest of the field in score from an early date.

There wasn't much of a push for religion in this game. Charlemagne had an uncontested route to the first religion by virtue of being the only leader to start the game with Mysticism tech, and he established Buddhism immediately in his second city. (Commodore hadn't turned on Choose Religions in this game and therefore we ended up with the standard ones.) Darius opened the game with Mysticism into Polytheism research and he founded a competing Hinduism a few turns later. This would slow the expansion of Persia a bit in exchange for Darius also building Stonehenge and picking up the free border pops on Turn 32. That would have been a fair tradeoff in terms of expansion, but Darius would eventually find himself in a situation where he was the only one practicing Hinduism in a world where most leaders had picked up Charlemagne's Buddhism, and that was not a good sign for the long term. All of the other religions were founded too late to have any significant effect on this game.

The first 100 turns of the game were surprisingly peaceful in this match. This was probably due to the rugged terrain in the center of the map, where there was tons of jungle choking off easy movement and no river connections for additional commerce and trade route access. The rough landscape seemed to slow down the teching of the leaders in the game game somewhat, and the strength of each individual AI was largely tied to how quickly they could connect their happiness resources for more city growth. The leader who was performing the strongest in this regard was Willem, with his Creative trait popping his borders for easy access to more resources and his Financial trait carrying his economy along. The Dutch came out of the landgrab phase tied for the most cities in the game with nine, and that spelled trouble for the rest of the field. Willem's biggest weakness is his poor expansion, and if he was sitting among the leaders in city count it was almost inevitable that he would start to run away in tech. Darius was keeping pace in technology but didn't claim as much land and faced serious diplomatic issues between his unpopular religion and his high peace weight. The Persians seemed the most likely choice to be attacked first.

The non-Financial leaders doing the best were Gilgamesh and Kublai Khan. Gilgamesh came out of the landgrab phase with eight cities of his own and the potential to make further gains against his weak neighbors. Kublai Khan had the most total land of anyone in the game, largely by virtue of stealing away a spot that should have been settled by Willem, and he would be a serious threat if he could ever get his economy under control. Charlemagne had not performed particularly well given his Imperialistic trait, likely because of his terrible starting techs, and the real weakling of the game was Stalin. The Russians were still hemmed in by barbarians and had sat in last place on the scoreboard for ages. Given that Stalin was one of the favorites in the picking contest, this was a real shocker. I'm still not sure why Stalin struggled so much here - were the barbarians really that much of a problem?

When it arrived, the first war declaration was a real head-scratcher:

Gilgamesh decided to attack Willem, with whom he had "Cautious" relations, ignoring the "Annoyed" relations that Gilgamesh enjoyed with Darius. Perhaps the cultural overlap between the mutually Creative cities along their border was the trigger for this war. Gilgamesh was also temporarily practicing Judaism, a minority religion of Persia, before switching over to a permanent Buddhism a little bit later. This was a major strategic mistake by Gilgamesh and a huge swinging point in the match, as the top two leaders on the scoreboard were pitted against one another rather than working together to take down a weaker rival. Darius was last in military power at the time and Gilgamesh would have been able to cut through him rather easily. Instead, Gilgamesh and Willem fought against one another to a strategic draw, with the Dutch city of Nijmegen (under the interface above) nearly getting captured several times before barely holding out. The whole conflict was a waste of resources for two leaders who could have better spent them elsewhere.

Things took a turn for the worse for Gilgamesh when Charlemagne attacked him a little bit later on Turn 114. With Gilgamesh's civ struggling under the burden of several centuries of warfare, the Holy Roman invaders were able to capture a border city almost immediately. This prompted Gilgamesh to sign peace with Willem, followed by Gilgamesh and Charlemagne fighting to a strategic draw with no further cities changing hands. Then Stalin would also launch his own invasion of Gilgamesh, which caused the Sumerians to sign a peace treaty with Charlemagne, followed by another war fought to a draw with the Russians. This was yet another wasteful conflict for Gilgamesh, who had come out of the landgrab phase as one of the most powerful leaders, only to be dragged back into the pack with these unproductive and pointless struggles. From this point forward, Gilgamesh would be more of a mid-tier leader than a serious contender.

There were two winners from this sequence of events. The first was Willem who was able to return to teching in peace, pretty much always the main goal for a Financial leader. The other winner was Kublai Khan, who was the first leader to do the sensible thing and go after Darius:

The Persians had been keeping pace with the Dutch in terms of tech, in part because of Willem's own defensive war. However, Darius had a total glass jaw and it was obvious that he would fold as soon as he came under serious aggression. Gilgamesh passed on attacking him, Charlemagne passed on attacking him, and even Stalin chose to go after Gilgamesh due to sharing the same Hinduism with the Persians. But eventually the Mongols came calling and Kublai's attack began to demolish Persia one chunk at a time. The border city of Gordium was the first to fall, and while Darius could absorb that loss, the fall of the size 15 city of Susa was another matter entirely. Darius was still frantically building wonders even as his empire was falling apart, completing Chichen Itza (useful for defense!) and the Great Library (not as useful!) during this hot war. A long train of Mongolian elephants continued to push deeper and deeper into the Persian core with no response in sight.

The game rapidly appeared to be moving towards a two-man race between Willem and Kublai Khan. The overall situation was reminiscent of the first playoff game where Willem had the superior teching power while Kublai had a much larger amount of land area under his control. Even as the Persian war raged on, Willem decided to construct the Apostolic Palace (attuned to the Buddhism that almost everyone was practicing) as well as other useful wonders like the Mausoleum and the Statue of Zeus. Then he decided to join in on the Persian fun as well:

This was a well-timed invasion from Willem, piling onto the dying Persians after Kublai Khan had done most of the hard work. Tarsus fell easily and after that the race was on to see which leader would get the tastiest bits of the spoils. Willem had the advantage of knights as the first leader to reach Guilds tech while Kublai had far more units overall. The Mongols ended up with the biggest prize by grabbing the capital of Persepolis, which came with Stonehenge, the Great Library, the Temple of Artemis, the Colossus, and the Sistine Chapel inside - not a bad haul. Meanwhile, Willem was researching Liberalism (taking Astronomy for the free tech) and racing to capture the few remaining cities held by Darius. It ended up being Willem who delivered the killing blow:

This was a tough field for Darius given how many low peace weight leaders ended up in the Championship. His best path to victory involved making lots of friends and surviving until he could out-tech the rest of the field and reach a position of safety. As the Twitch chat pointed out during this game, the best trait pairing in Civ4 is "Financial / Alive" and Darius proved unable to maintain the second half of that pairing. He basically did everything wrong from a diplomatic perspective by founding his own unique religion and then making no effort to spread it to anyone else, ensuring that Persia would be despised by the rest of the planet. In a world of Buddhist faces, there was no room for a pacifistic Hindu civ with a ton of easy-to-conquer world wonders.

You may notice that Stalin at the bottom of the scoreboard was also sporting Hinduism as a religion. This was a deadly and potentially suicidal choice on his part, with Charlemagne predictably leading a crusade against the heathen Russians a short time later on Turn 178. With Stalin still stuck at war with Gilgamesh from their earlier pointless conflict, it seemed all but certain that the Russians were on their way out of the game next. Even as the border cities were in the process of being overrun, Stalin made an intelligent move to save his game:

He converted over to the world's dominant faith of Buddhism and therefore increased his diplomatic standing with everyone else. This was not enough to save Stalin in this particular "stop the war" Apostolic Palace vote, with both Gilgamesh and Charlemagne voting against the resolution. However, Stalin was able to get a peace treaty with Gilgamesh and once normal relations had been restored, Gilgamesh joined Willem and Kublai Khan in voting to stop the war. Therefore the following "stop the war" Apostolic Palace resolution voted on ten turns later did succeed and this was enough to keep Stalin in the game as a runt civ. He had lost his capital and about half of his territory to Charlemagne while still clinging to life with four remaining cities. Charlemagne's position in the game had improved marginally as a result of taking those Russian cities while still remaining far behind the two game leaders of Willem and Kublai Khan.

These wars involving Stalin and Charlemagne and Gilgamesh were a bit of a sideshow though. The major individual driving the action of the game forward was Willem, who was starting to race out to a frightening tech lead. He had built the Mausoleum earlier and earned the first-to-research prizes at Liberalism and Economics. He used the free Economics Great Merchant for a lengthened Golden Age, which allowed the Dutch capital to start working on the Taj Mahal:

Willem would land this wonder as well for the double Mausoleum-enhanced Golden Ages. We've seen this pattern many times before in our Multiplayer events, and if someone can line up two or three Mausoleum-boosted Golden Ages for 24 or even 36 turns of Golden Age production and research in the Renaissance era, they often blow past the rest of the field en route to an easy victory. The critical tech for Willem was Rifling, the most important military tech of the Renaissance period, opening up rifles for defensive safety and cavalry for offensive action. Willem already had Gunpowder and Military Tradition techs, and he had researched Replaceable Parts to knock out every prerequisite needed. We had multiple different people who had written in their predictions before the game started that Willem would win the Championship if he could manage to make it to Rifling tech, and that he would be eliminated if he didn't make it there. The Dutch were fully a dozen techs ahead of anyone else in the game right now, and all that Willem needed to do was pick up the one final military tech that would make him effectively invulnerable en route to his Spaceship or Cultural victory. Surely he couldn't screw this up, right?

We were about to find out. Kublai Khan launched his own attack against the Dutch on Turn 185. This was a last-gasp attempt by Kublai to rein in Willem before he became completely unstoppable, and the success of this conflict would turn entirely on how quickly Willem was able to research Rifling tech. Kublai was essentially a full era behind Willem in technology, but he was beelining Rifling as his first goal in the Renaissance, whereas Willem was ignoring Rifling while researching seemingly everything else first. Willem had already picked up the full Democracy line and Astronomy and Corporation, and now he was going after Scientific Method. He couldn't keep ignoring Rifling tech for much longer, not while in a major war with the Mongols... right?

Kublai started the war with an army composed mostly of knights, then was able to upgrade them fairly quickly to cuirassiers as he picked up Military Tradition and Gunpowder techs. From there, the Mongols began slowly making their way towards the expensive Rifling tech. Meanwhile, Willem was researching first Physics tech and then Communism tech. Yes, the Great Scientist and Great Spy were nice prizes, but he really needed to get a move on Rifling tech right now. Kublai's numerically larger armies were winning in the field, using this time to capture all of the Dutch colonies in ex-Persia. This was buying time for Willem to pick up Rifling tech but that time was starting to run out. When Communism research finished, we checked to see where Willem was headed next, and the viewers were left dumbfounded:

What insanity was this?! Willem had queued up three more techs in succession: Steam Power followed by Steel followed by Railroad. It was a testament to his ridiculous research prowess that he could knock out the expensive Steam Power tech in as little as two turns, but the whole thing was pure madness. Just look at Rifling tech and how Willem was going to tremendous lengths to avoid researching the one thing that would be his salvation in this conflict. Yes, machine guns would be useful to have against Kublai's forces, but it would be ten more turns before they were available. In the meantime, Kublai's military was making one gain after another and starting to move from the Dutch colonies into the Dutch core. The Power bar graphs told the story:

From roughly even at the start of the war to a massive advantage in favor of Kublai, all because Willem insisted on avoiding Rifling tech as though it had been poisoned. In fact, Kublai actually had the superior military tech advantage by this point in time because he had made it a priority to research Rifling. Kublai now had rifles and cavs while Willem was still defending with cuirassiers, longbows, and a smattering of machine guns. The Mongols somehow had superior units despite being close to 15 techs behind overall. This was one of the most bizarre acts of self-sabotage that we've seen in AI Survivor... and it STILL kept going onwards, as Willem finished Railroad tech and then went on to Combustion next. Yes, Combustion tech before Rifling while facing a major invasion from largely horse-based units. Willem, I literally can't even.

Finally, after finishing up with Combustion tech Willem now decided to research Rifling tech, but the damage had already been done. The Dutch core cities of Leiden and Rotterdam had fallen to the Mongols and Kublai's cavalry-based army had simply become too large to be stopped. The main Mongols stack had well over 100 units in it by now, including 90+ cavalry with lots of promotions. The fact that Willem went straight from Rifling on to Assembly Line tech was irrelevant at this point, as a dozen infantry weren't going to do anything to stop a hundred cavalry. This whole war was a fantastic demonstration of the power of cavalry in large numbers. Kublai was winning the war, and likely the game, because he had prioritized a critical military tech and built large numbers of one of the game's best units. Willem had spent too much time on economy and failed to defend himself properly. As if to add insult to injury, a random Great Engineer from Charlemagne allowed him to swipe away the Statue of Liberty a single turn before Amsterdam would have finished the wonder. That was just a kick in the teeth.

The rest of this war was largely academic. The siege of Amsterdam took about 20 turns because Kublai only had a single trebuchet on hand to remove the defenses but there was never any doubt about how it would end. Kublai's main stack had over 140 units in it by the time that the final attack was launched. Amsterdam delivered over a treasure trove of wonders to Kublai: the Kremlin, the Statue of Zeus, and the Apostolic Palace, but not the Statue of Liberty thanks to Charlemagne's last-second snipe. Afterwards, Kublai proceeded to dismantle the remaining Dutch cities largely without waiting for siege units to arrive, sacrificing sizable numbers of cavalry against infantry behind heavy cultural defenses. It was working though, and Kublai had the territory and units to spare. He was getting close to 50% land area and would need only one more successful war after Willem was gone to reach Domination. When Willem was down to his last three cities, Gilamesh opportunistically leaped into the war in the hopes of vulturing something for himself. In a totally underserved finish, Gilgamesh landed the killing blow and claimed the elimination credit:

This put Gilgamesh into the lead for the Golden Spear award with five kills on the season, one ahead of Charlemagne's four kills. It was the only city that the Sumerians had taken in the war, with everything else falling to the Mongols. Kublai Khan was now the Runaway AI of the game, and although his endless warring had put him about four or five techs behind Charlemagne for the moment, it wouldn't take long before Kublai's massive size advantage would pull him ahead. Kublai would have his choice as to whether he wanted to finish off his conquests with a Domination ending or sit back and go the economic route by heading for space. As for Willem, he had completely and utterly thrown away an easy victory by refusing to research Rifling tech. The Championship was right there for the taking and he tossed it away. Full credit goes to Kublai Khan for seeing a narrow window of opportunity and taking advantage of it. Willem had completely tossed his victory but Kublai had at least been there to pick it up.

Kublai Khan needed at least a short period of peace to develop all of his captured territories and prepare for the next stage of the game. He was denied this opportunity by Charlemagne, however, who invaded the Mongols on Turn 269. This felt like a suicidal attack from Charlemagne given the vast size of Kublai, but Holy Rome was able to take advantage of the fact that they had Assembly Line technology for factories and infantry while Kublai was still missing it. The Production bar graphs showed that this essentially left the two sides even on shield output, with the smaller number of factory-based cities in Holy Rome equalizing the larger number of non-factory cities in Mongolia. The result was a stalemate on the battlefields:

Kublai Khan was still able to capture one of the Holy Roman border cities, however that was as far as he was able to get before peace broke out. While Charlemagne had narrowly dodged a bullet in this conflict, his position looked shaky in the extreme. He had incurred a major diplomatic penalty by declaring war on Kublai, souring their relationship into a mutual "Annoyed" on both sides. It was almost inevitable that Kublai would be returning to war again before this game reached its conclusion, and the vastly larger Mongol empire would quickly surpass Holy Rome in research and production output. Charlemagne had only been able to fight to a draw in this war because he had factories + infantry while Kublai lacked both. Kublai would have to do something incredibly dumb to avoid rolling over his western neighbor and winning the game.

Speaking of incredibly dumb...

Four turns after signing peace with Charlemagne, Kublai Khan decided to turn the culture slider up to 100% and begin chasing a Cultural victory. This was indescribably stupid, fully as pants-on-head idiotic as what Willem had done earlier. What in the world would have prompted Kublai to think that going after culture was a good idea?! His cities were mostly configured for production and had poor cultural and commercial output. None of them were remotely close to hitting 50k culture and the projected finishing date for three Legendary cities was post-Turn 400. Kublai could easily finish the tech tree and launch a spaceship dozens of turns earlier, not to mention the fact that by turning off research he would find himself out-teched by vastly weaker rivals like Charlemagne and Gilgamesh. The timing of this cultural push was particularly dumb because Kublai had stopped his research just as he reached the critically important Assembly Line tech. Apparently he didn't want to have access to factories and power plants and infantry, which didn't even make sense for a Cultural game since the additional production would be highly useful for training missionaries and building cathedrals. What in the world was Kublai smoking?!?

Not much happened for dozens of turns after this. Kublai slowly limped towards his ridiculous Cultural victory while Charlemagne continued slowly researching towards the end of the tech tree and a spaceship. It was painful how slowly the game was proceeding along; Charlemagne couldn't crack much more than 1300 beakers/turn, a mark that Willem had reached well over 100 turns earlier. Charlemagne attacked Gilgamesh at one point, only for a United Nations peacekeeping resolution to stop the war after a pair of border cities were captured. Eventually Kublai Khan returned to war with Charlemagne on Turn 323:

Unfortunately, Kublai had been running the cultural slider for 40 turns by this point. Although he had slowly limped to Assembly Line over the course of 25 turns, his tech had otherwise not advanced over these intervening decades. Meanwhile Charlemagne had continued progressing down the tech tree at his mediocre pace, eventually reaching more advanced units along the way. Therefore the Mongols were once again fighting at a tech disadvantage, with Charlemagne having tanks on hand and unlocking mechanized infantry via Robotics a few turns after the war started. Kublai would win easily if he could simply fight one of these wars at tech parity, but his foolish insistence on running the cultural slider was ensuring that he was stuck a generation behind in military technology every time that a conflict began. It was incredibly frustrating to watch this farcial sequence as it played out live. Kublai was dominant in every way but insisted on throwing away his advantages by chasing after an impossible cultural victory condition. Charlemagne was completely mediocre and slowly creeping closer to a win, not due to his own efforts but through the incompetence of his rivals.

This was not how the Championship Game was supposed to play out!!!

Kublai Khan also tried his hand at the Diplomatic victory condition, and he could have won here if he could have convinced anyone else to vote for him. However, no one liked one another very much in this particular world, and the Mongols couldn't get relations with anyone else above +2 diplo points. The diplomatic option wasn't going to happen. There was one brief moment where Kublai turned research back on again and everyone got excited, but nope, he went back to the culture slider yet again after a few turns. Sigh. Charlemagne invaded Gilgamesh on Turn 358 and began carving up the southern half of Sumeria. He would have killed Gilgamesh for sure except that Kublai Khan came to the rescue with his own invasion half a dozen turns later. With the Mongols now having their own tanks, would this be the turning point? Nope, as Charlemagne countered with Holy Roman modern armor, mechs, and mobile artillery, once again remaining a generation ahead in military technology. Yet again the result was a stalemate, with the superior numbers of the Mongols countered by the superior tech of Charlemagne. This would be more legitimate if Charlemagne had been a good researcher in this game, but he had been nothing of the sort. We were looking at a Holy Roman spaceship finishing sometime shortly before Turn 400, a pathetic result that would have resulted in defeat in every other game this season. Willem would have won the game a hundred turns earlier if he had researched Rifling tech, and Kublai would have won via Domination fifty turns earlier if he had researched anything at all. What a mess.

In any event, Charlemagne launched his spaceship at the end of Turn 378, and it arrived ten turns later on Turn 388 to seal the win:

This might not be the worst ending in AI Survivor history but it was certainly the worst finish we've had in a Championship game. Charlemagne was an undeserving victor, the last AI left standing after all of the better leaders refused to take the minimal steps needed to secure a win. He won his victory in Playoff Game Three in exactly the same fashion, you may recall, after Darius tossed away his own certain victory in that game. As for Kublai Khan, we've written before that he's the ultimate second-place leader and that very much proved true to form here. His failure to close out this game will haunt him for the rest of his AI Survivor days. My overall impression from this season was that we ended up with a group of leaders who weren't especially strong in the Championship this year, and the results of this game definitely seemed to bear that out. It was hard to avoid the sense that Justinian or Huayna Capac or even Julius Caesar wouldn't have won this game instead if they were here. Sometimes in sports the best team doesn't end up winning and that felt like the case here.

Now we have to live in a world where Charlemagne - Protective/Imperialistic Charlemagne! - is an AI Survivor champion. You may be able to tell that I'm les than enthusiastic about that result. I will probably replay this game when I have some spare time to see how often Charlemagne ends up as the winner. I suspect it's not too often. Regardless, this was the finish of Season Four, which was a fantastic year in pretty much every way other than the ending of the Championship game. I'll have more thoughts in a Conclusions page shortly. In the meantime, thanks again for reading and watching along with the competition.