Civ4 AI Survivor Season 5: Game Seven Writeup

The field in Game Seven appeared to be an uninspiring lot aside from the seeded leaders. The standout in this group was Pool 1 leader Justinian, winner of Season One of AI Survivor and nearly a repeat winner at the conclusion of Season Three. Justinian combines together strong expansion with his Imperialistic trait and a tendency to form enduring alliances via shared religious faith. He would likely be opposed in the religious race by the Pool 2 seeded leader Zara Yacob, another AI leader with an unusually high emphasis on faith. Zara had a fantastic showing in Season One of AI Survivor yet struggled to make much headway in the years since then. The unseeded leaders were a mixed assortment with a little bit of everything, from Alexander's militarism to Churchill's defensive turtling to Qin's wonder building. Outside of Justinian and Zara this was a pretty lifeless group of leaders. We were curious to find out if any of them could make a name for themselves in this match.

The expectation going into the game was that Justinian and Zara would found two competing religions on opposite sides of the map. Justinian was the only leader to start with Mysticism tech, and despite opening with Fishing research to connect the seafood resources at his capital, the Byzantine leader still picked up Meditation tech before anyone else and established Christianity at an early date. Things did not work out nearly as well for the religiously-inclined Zara, who opened with Mysticism research and then began a futile chase after the same Meditation tech. The Ethiopians were beaten to the punch by Justinian and ended up with 15 turns sunk into religious techs with nothing to show for it. Rather than continue onwards into Polytheism, Zara dropped out of the religious race and allowed a surprising Churchill to found the game's other competing religion in the form of Confucianism. This would have enormous consequences that echoed down across the rest of the game.

Justinian had sent his initial settler southwest into the tundra, where its holy city culture largely ended up grabbing additional seafood resources instead of more valuable territory elsewhere. Churchill's holy city proved to be better situated, expanding borders quickly in the central jungle region and placing immediate pressure on some of Peter's nearby Russian cities. The English were looking pretty good out of the gate only to stop expansion at three cities to build first Stonehenge (arguably worth it) and then Temple of Artemis (definitely not). This deprived Churchill of the chance to claim more land and snowball ahead during the landgrab phase. Qin was another leader who was doing well in the early turns thanks to starting the game with both Agriculture and Mining techs. His problem was a failure to research Mysticism tech, leaving his cities stuck with their initial 9 tile radius and many unused resources out in the second ring of each city. And while building the Great Wall may have been thematically appropriate, it diverted Chinese resources away from expansion at a point in time when more settlers and workers would have been the better call.

As the landgrab phase began to draw to a close, the six leaders still remained tightly bunched together on the scoreboard. Justinian had been the big pregame favorite with roughly 75% of the picking contest entrants selecting him as the expected winner, and he was playing a solid game given the somewhat weak position that he was drawn into. Justinian was expanding rapidly and he would later build the Oracle for a fast Monarchy tech. The combination of a self-founded religion, cheap Spiritual temples, and Hereditary Rule happiness allowed Justinian to grow larger cities at an earlier date as compared to his rivals. The Byzantines were hampered largely by their tundra-heavy starting area along with the presence of several nearby barbarian cities. Apache popped up in a particularly annoying spot and blocked Justinian's attempts to expand further to the north. Meanwhile, Peter was doing a nice job of expanding and would eventually capturing the barbarian cities of Aryan and Vandal in his backlines. His problem was the English culture off to his east where London and York were stealing away most of the tiles from Novgorod. With room for more backfilling of cities in the icy north, Peter looked to be in a solid position for the long run.

The most consequential developments for the game were taking place off in the east. Alexander wisely opened the game with Agriculture research and expanded north towards the Ethiopians, thus allowing him to connect the farming-based food resources at his initial cities. His militaristic tech preferences also led the Greeks towards a fast Bronze Working that made use of a copper resource at the capital to train a bunch of spears and the unique unit phalanxes. By way of contrast, Zara's early game had been a complete disaster. He opened with the failed religious pursuit that didn't result in a holy city, then continued with Agriculture, Wheel, Animal Husbandry, and Pottery before heading back onto the religious track against with Polytheism and Masonry research. It looked like Zara was trying to make a run at the Monotheism religion since his faith-heavy tech preferences desperately wanted him to establish his own religion. However, Zara had left himself with two gaping holes in his military setup: he didn't have Bronze Working (or a copper resource anywhere in his territory) and he never bothered to research Archery tech. Perhaps his sheltered position free of barbarian activity led him to ignore archers, or maybe his AI tech flavors simply didn't emphasize the tech strongly enough. In any case, Zara had left himself enormously vulnerable to aggression since he couldn't train anything other than warriors and chariots.

And Alexander didn't miss this opportunity:

Typically the hyper-aggressive warmongers perform poorly in our AI Survivor games. They start their military buildups too early and hurl their forces at other leaders when much of the map is still unsettled and there's easy land to be claimed. That was very much the case here with Alex having only five cities; a barbarian settlement popped up shortly thereafter along the eastern coast in a place where the Greeks should have planted one of their own cities. In a normal scenario, Alex would be ruining both his own game and someone else's game by throwing away dozens of units to no effect, slamming axes and spears into enemy units defending in cities behind walled protection. However, this was not a normal game and Alexander had essentially lucked his way into the perfect scenario. His opponent had no axes, no spears, not even any archers. Zara didn't even have Bronze Working so he couldn't whip out additional units for defense. The only units that Ethiopia could build were warriors and chariots. Making matters even worse, the Greeks have a unique unit that specifically counters chariots! Phalanxes are axes that don't have the normal defensive weakness against chariots, thereby rendering the Ethiopians completely and utterly helpless.

As a result, we saw an unprecedentedly fast collapse of Zara's territory. Greek units walked from one city to the next and killed the defenders without suffering any losses at all. There was no last-minute whipping of defenders as we would normally expect to see since Zara had no Slavery civic to fall back upon. There were a handful of archers thanks to the four preplaced archers that each Deity AI starts the game with, and otherwise resistance was nonexistent. Ethiopian cities fell on Turn 64, Turn 70, Turn 72, Turn 76, and the finishing blow landed on Turn 80:

Just like that, Zara Yaqob was gone. This was by far the earliest elimination that we'd ever seen in an AI Survivor game, shattering the previous record of Turn 104 suffered by Pacal in Season Four. It was a result that we never would have seen before removing the extra free starting techs that the Deity AIs typically get, as Zara would have had access to Archery tech and would have been able to at least slow down the bleeding. Instead, Alex had doubled his territory at an incredibly early date, before the map had even been filled up with cities. It was all the result of an unbelievably disastrous opening sequence from Zara who had mismanaged his research to a degree that was hard to believe. How could someone lack Bronze Working *AND* Archery techs that late into the game with all of the Deity discounts the AI gets? We snapped a picture of the Ethiopian tech tree right before Zara kicked the bucket, at a time when he was frantically researching Sailing tech instead of something that might actually help him. Too much religion emphasis can be one heck of a drug.

So now Alexander was double the size of anyone else in the game, and with his extremely war-heavy personality, it was going to be his game to lose. Again, the obsessively militaristic leaders typically tend to self-destruct because they waste too much production building units that die fighting pointless wars and never get around to developing their economies. However, if the insane aggressors do manage to snowball ahead somehow, their research of military techs combined with their heavy train unit emphasis can result in some serious smackdowns of the other leaders. Alexander was already pushing for Construction tech long before anyone else would have access to catapults, and he was approaching a dozen cities capable of training units. This didn't look like it would end well for his next target.

There were other developments taking place in the game, of course. Justinian and Churchill both completed their respective shrines and began spreading their religions. Qin picked up Justinian's Christianity while Peter adopted Churchill's Confucianism. Which religion would take root over in Greece was the big question influencing how the remaining diplomacy would break down. Peter ended up liberating one of his border cities to England and saw the other one flip away outright; when combined together with Churchill's construction of the Pyramids, it looked as though the English leader was establishing himself in a strong position. It all came crashing down shortly thereafter though when Alexander singled him out as the next target:

Note that the Greeks waited for all of a dozen turns after eliminating Zara before launching the next invasion. The Alexander AI has a very one-track mind when it comes to fighting wars. His initial attacks were dashed against the walls of English cities and achieved nothing, but fortunately for Alex he was nearly finished with Construction research and catapults were on their way. Churchill was in a poor position from a diplomatic standpoint as he had the highest remaining peace weight amongst the surviving leaders. This made him a target for the other low peace weight leaders in the game, and it certainly didn't help that he had been pushing culture heavily and running up large border tensions with all of his neighbors. Qin joined the war against the English on Turn 100 and Peter jumped in as well on Turn 115. Now it wasn't a matter of whether or not the English would be eliminated but rather who would get the most from the spoils. The Greeks were the heavy favorites in this regard thanks to having the largest armies and the earliest access to catapults.

Meanwhile, Justinian was playing a legitimately strong game despite his underwhelming starting position. He suffered a setback when Peter captured the barbarian city of Apache on their mutual border, but he was compensating for this through having larger than normal cities thanks to lots of available happiness. Justinian also built the Great Lighthouse and he'd planted enough iceball fishing villages to make it a powerful wonder. The main thing that the Byzantines needed was more territory, and Justinian made a move to acquire it by attacking Qin on Turn 131. While Qin had converted to Christianity and gained a major diplomatic bonus with the Byzantines, Justinian evidently had begun plotting war before the religious conversion took place. He snapped up a couple of Chinese border cities and looked poised to escape from the southwest corner of the map. However, even as Justinian was making his move, time had run out on the English as Alex claimed his second kill of the match:

As far as the spoils shook out, Peter had picked up one city, Qin gained zero cities, and Alexander took all of the rest. Qin's invasion of England proved to be a disastrous mistake when seen in retrospect, throwing away units for no reason and softening up China for the Byzantine invasion. The Greeks were now completely and utterly out of control, having swallowed up the territory of three different civs and captured all of Churchill's juicy wonders in the process. The amazing thing was how quickly this had all happened - most games don't even have one elimination take place by Turn 136, much less two of them! A couple of the Livestream viewers compared the situation unfolding in this game to the real-world conquests of the youthful Alexander and I think it was a smart observation. In a normal game, Peter and Justinian would both be well-positioned to compete for the first place spot. It increasingly felt like they had no shot though because the Greek avalanche was already nearing the bottom of the mountain.

The next major development in the game came in the form of a Russian invasion of Byzantium. This conflict was primarily driven by the fault lines of religion, as Peter had remained staunchly Confucian ever since the early turns of the game while Justinian was tightly attached to his self-founded Christianity. In an ominous sign for the Byzantines, Alexander had converted over to Confucianism after capturing the holy city for the religion from Churchill. The AI has a preference for switching to the religion of a holy city that they control and it was difficult to see Justinian persuading Alex to come back into the Christian fold. All attention now turned to who would be the next victim of the Greeks, and with Alex able to declare war at "Pleased" relations, no one was safe from his forces. China was the unlucky target of Alex's latest conquest binge, and although a stack of 100 Greek units mindlessly stood in place for a dozen turns thanks to some bizarre AI pathfinding issue, the outcome of the conflict was never in doubt. Eventually the stack started moving and China quickly disappeared from the map:

While Alexander had been mercilessly curb-stomping poor Qin into the ground, the war between Peter and Justinian had been grinding on in the background. Although the two sides had been fairly comparable at the beginning of the war, Justinian was weakened by his ongoing struggle with the Chinese and quickly lost a pair of border cities to the Russians. This appeared to put Peter over a tipping point whereby he had too many cities and too much production for the Byzantines to prevail in the conflict. Justinian was suffering from having so many cities stuck in the deep south where they were either buried in tundra or iceball fishing villages. His situation was reminiscent of Huayna Capac from the previous game: a powerful leader stuck with mediocre terrain. However, unlikely the Incan leader who was able to break out of his starting position and snowball ahead by devouring the Aztecs, Justinian was never quite able to throw off the shackles of his tundra start. I think that in a normal game he might have been able to manage it by rolling over the Chinese and taking their territory. But in this game, the incredibly fast conquests of the Greeks threw everything off kilter and by the time that Justinian started to make his move, it was already too late. The arrival of cataphracts at Guilds tech was enough to stall out the Russian advance but not enough to turn the tide. Their war turned into a slow grinding process of attrition, with the Byzantines gradually bleeding out territory across dozens of turns.

We wondered if Alexander would choose to get involved in this war since he had nothing else to do, and we didn't have too long to wait. The Greeks launched their own invasion of the Byzantines on Turn 212 and proceeded to steamroller the helpless Justinian. Alexander had surged far ahead in technology by this point and he had reached the rifles/cavalry era of military technology while the Byzantines were stuck fielding medieval units. We all know what a slaughter that looks like in Civ4, plus Alex also had more cities under his control than the rest of the world combined. The slow, halting advances of the Russians were brushed aside as the Greeks swept away everything in front of them in a blitzkrieg campaign. It only took 16 turns for the ancient Greeks under Alexander to eliminate their medieval Byzantine counterparts and prove to be the true inheritors of Hellas:

This game had demonstrated exactly why Justinian is such a strong competitor for AI Survivor. He had expanded well, teched well, developed his cities well, and his cataphracts provided a tremendous punch when they appeared on the battlefield. However, it hadn't been enough to prevail in this game, partly due to subpar starting terrain and partly due to some bad luck with barbarian city spawns that stopped the Byzantines from claiming more of the map. But the biggest downfall of Justinian in this game was due to events taking place on the other side of the map that he had no control over. Alexander's rapid conquests snowballed the pace of the game to a ridiculous degree, such that he had already wrapped things up before the Byzantines even had a chance to gain additional territory. Midgame wars of expansion using cataphracts don't matter if a leader on the other side of the world seals things up before the midgame even arrives. Justinian was therefore a bit of an unfortunate victim of circumstance, knocked out of the competition despite playing an above-average game.

As for Alexander, he set a new record by scoring four different kills in a single game of AI Survivor, something we had never seen before. Not only that, he had accomplished all four of those kills in only 228 turns! Game Three earlier this season hadn't seen even a single leader eliminated by this date. The Greeks were now close to 60% land area and it was only a matter of time until they managed to surge over the Domination limit. Alexander had been "Friendly" with Peter in recent turns, however Peter foolishly swapped into Free Religion and lost the shared Confucian boost to diplomacy. While they remained "Pleased" with one another, it looked like only a matter of time until the Greeks went back onto the warpath again. Instead, Alex contented himself with settling more cities in the icy expanses of the South Pole, and eventually he picked up enough tiles via this method to pass over the Domination threshhold without needing to start another war:

Thanks to this ending stretch of peaceful turns, Alexander narrowly missed out on scoring the fast victor in AI Survivor history. That honor belongs to Huayna Capac's scorching Turn 252 Domination win in Season Four, a mark that Alexander could have easily surpassed if he'd been more willing to launch an assault on Peter. Regardless, this still wound up being one of the most lopsided games we've ever seen. Alex played the perfect conqueror's snowball setup from start to finish, running over one target after another while using their captured territory to launch him onto the next victim. Alexander had zero kills in the previous four seasons of AI Survivor and a single, distant, second-place finish to his name from Season One. He also set the single-game scoring record with 9 total points thanks to a first place finish and those aforementioned four kills. Whether this is repeatable in the playoffs is a different question; Alex's wild aggression is more likely to result in famine than feast in a normal game. Still, for one shining moment, he was indeed the young, brash conqueror that stormed across the Granicus, destroyed the Persian Empire, and wept that there were no more lands left for him to conquer.