Civ4 AI Survivor Season 3: Playoff Game One Writeup

Our first playoff game for Season Three brought together a field of leaders that could be sharply divided into two groups. On one side, there was a trio of leaders who all shared high peace weight values and generally favored peaceful victory conditions of some type. Asoka was the best representative of this group, as he heavily emphasizes religion and can act nearly as pacifistically as his Indian partner Gandhi. Hammurabi often follows a similar path, with his tech preferences geared towards anything that will produce more culture. The Babylonian leader's preferred course of action is to sit in a corner of the map somewhere and build every wonder under the sun. The other high peace weight leader in this game was Wang Kon, the only Financial leader in this game and someone who has become infamous in this competition for trolling the rest of the field in creative ways. They would be facing off against three leaders with significantly higher aggression ratings at the bottom end of the peace weight scale. Justinian was the religious zealot of this particular group, someone who was all but certain to found his own religion and then war repeatedly against everyone who did not share it. Stalin was the dark counterpart to Hammurabi, as he also favors wonder-building with his Industrious trait but is much more willing to attack other civs with his Russian cossacks. Finally, this game also included Shaka after his surprising second place finish in the opening round. Shaka is pretty much willing to invade anyone else at any point in time for any reason under the sun, and the Zulus would be heavily predisposed to dislike the three leaders at the high end of the peace weight scale. It seemed likely that both of the leaders moving on to the Championship game would come from the same group, either two high peace weight leaders or two low peace weight leaders.

Our map in this game distributed the six leaders with three of them in the north and three of them in the south. The leader who appeared to be in the most danger was Asoka, with Justinian off to his east and the other two low peace weight leaders sitting to his north and northeast. If Asoka were to be attacked by all three of them his game could potentially unravel in a hurry. On the other hand, Asoka had an excellent starting location with a strong capital, and if he could expand well enough and stay out of early warfare, he might be able to get far enough ahead that it wouldn't matter. Justinian was the big pregame favorite in the community, with about 60% of the contest entrants picking him to win this match. However, he had a very weak starting capital that lacked much in the way of food bonuses or strategic resources. I decided that I would pick Stalin to finish first over the Byzantine leader for this exact reason, fearing that Justinian would have a slow start and struggle to expand. The Russians had a fair amount of jungle to contend with, but were in a good position for long term growth. Elsewhere, Hammurabi had rolled a starting position where he could do what he loved most, hang out in the far southwest corner and use a pair of nearby stone and marble resources to construct all of the wonders. Shaka was placed in the north central part of the map and was certain to do Shaka things, i.e. stirring up trouble by declaring war on anyone and everyone. Finally, Wang Kon was in the northwest corner of the map, and a good portion of the community thought that he would be the first leader eliminated due to a lack of strategic resources in his immediate vicinity. At the same time, another segment of the fan community picked Wang Kon to win the game and wanted to see more craziness coming out of the Korean leader. I had my doubts that Wang Kon would be able to pull off anything as insane as his victory in the opening round, but you never really know when it comes to the Troll King.

This was a game where three different leaders started with Mysticism tech, and the race for the first two religions was lively. It became even more interesting when Shaka popped Mysticism from a hut on the second turn of the game and began chasing after Meditation himself. That could have potentially thrown the diplomacy for this game into an unexpected whirlwind if Shaka had been the one to found an initial religion. Instead, Asoka beat out Shaka by one turn and order was preserved, with Asoka founding Buddhism and Justinian founding Christianity (via Polytheism) a turn later. Both of them immediately converted to their respective religions and this cooled relations between the Indians and Byzantines, much as we had been expecting. Hammurabi's cultural preferences in research often cause him to found his own religion (he founded BOTH early religions in his opening round game!), and he was the one to chase after the Monotheism religion, establishing Islam a little bit later on Turn 37. This gave all three of the leaders on the southern side of the map their own self-founded religion, and none of the later religions would end up playing a significant role in this game.

As far as expansion was concerned, nothing too crazy took place in this game. Perhaps the most noteworthy early development was Asoka sending his extra starting settler to the west, which created slightly more room for Justinian when the second Indian settlement became the Buddhist Holy City. Justinian's own Holy City was located to the northwest of his capital, in exactly the direction he needed to be heading to secure a strong portion of the land. Stalin probably had the worst location for a second settlement, placing it on the northern coast rather than heading inland towards the center of the continent. Over the following turns, the AI leaders continued spreading out and claiming more and more of the map. Stalin built an early Stonehenge on Turn 28, and that seemed to be a big help for a Russian civ that lacked an early religion or any other way of popping borders. Wang Kon's early construction of the Great Wall was a lot less useful, and the Koreans were struggling a bit in the early game due to the aforementioned lack of nearby strategic resources. Korea's third city was placed on the northern coast in a spot that claimed no territory at all, and some barbarian problems to the south were continuing to slow expansion. For that matter, the barbarians were going a little bit wild in terms of founding cities; we had no less than five barb cities on the map in this game! I double checked and Raging Barbarians was not turned on for this match. It was a surprising number of barbarian settlements for such a small map.

Here's the traditional overview from Turn 50. All five of the barbarian cities should be visible, and which leader claimed key spots like Aryan, Hittie, and Illinois was going to go a long way towards determining who emerged victorious on this map. Asoka seemed to be doing the best at this point, with five cities established and a lot of territory already within his borders. He had a clear score lead over the rest of the field at this point. That was no surprise to me; I had been expecting Asoka to lead the scoreboard through the whole early game, and yet also be the first leader eliminated due to his diplomatic problems with his neighbors. For the moment though, it was all working out for the Indian leader. To me, the biggest surprise was Justinian, who I thought would struggle in the early game due to his weak capital. That wasn't happening thus far, as the Byzantine leader was also out to five cities and had claimed an important copper resource at Nicaea. He would eventually capture the barbarian city in the extreme south and that would get Justinian up to seven or eight cities, more than enough to be competitive in this game. Elsewhere, Shaka had also reached five cities in a circular pattern around his capital and looked to be a power in this game. Stalin was struggling a bit but I had hopes that he would capture Hittite and emerge with the same number of cities as Justinian, to serve as a partner with the Byzantines throughout the rest of the match. The two western leaders were in weaker positions, with Hammurabi still working to expand out of the tundra surrounding much of his starting position. He was going to get about as much land as everyone else, although it didn't seem to be territory of the same quality. Finally, Wang Kon was still stuck at only three cities and looked to be falling behind in the expansion race. He was also building the Pyramids (with stone fortunately) and that was slowing down expansion further by investing production into an expensive wonder. The Korean seemed likely to be the runt of this game from a territorial perspective give how little land Wang Kon was grabbing.

Wang Kon did indeed build the Pyramids about a dozen turns later, then paired that construction with a swap into Islam. This established the Koreans and Babylonians as fast friends, although it also served to alienate them somewhat from Asoka and his Buddhist empire. Wang Kon used the Pyramids to revolt into Representation civic, and that along with his Financial trait and the presence of nearby gems and ivory allowed the Korean cities to grow to larger sizes than anyone else, packing much more commerce than any of the other leaders. We therefore ended up with the strange situation of Wang Kon leading the rest of the field in technology while also having the smallest amount of territory to his name. That almost never happens in these AI Survivor games, and it created some unusual dynamics in this particular contest. Wang Kon would go on to found both the Code of Laws religion and the Theology religion, although he wouldn't end up swapping to either. Was this another subtle form of trolling the rest of the field? I couldn't even tell at this point.

Wang Kong also had to deal with the presence of an unstable lunatic on his eastern border, and Shaka kicked off the military action in this game with the first war declaration on Turn 79. This was a key moment in the early stages of the game, as we had been waiting to see if Shaka would strike first at Asoka or Wang Kon. If he had chosen to attack the Indian leader first, then it was likely that Asoka would collapse almost immediately when Justinian and Stalin inevitably joined the war. Instead, Shaka was expending his military energies on his other neighbor, and that potentially gave Asoka a fighting chance in this match. The immediate Zulu attack against P'yongyang was a failure, which was not a surprise when invading into heavy cultural defenses. However, Shaka did manage to conquer the former barbarian city of Aryan, which Wang Kong had taken a little bit earlier. That was the Korean city due south of Zulu territory on the minimap, and by claiming this spot, Shaka was able to force Wang Kon back into his starting corner of the map, denying him access to the center of the continent. This war would continue to rage onwards for some time to come without any other cities changing hands.

At the same time, tensions had been building on the eastern side of the map. Asoka had been the one to capture the barbarian city of Illinois right along the Byzantine border, and between a mixture of border tension, religious disputes, and a large difference in peace weight, the relationship between Asoka and Justinian had become severely strained. The fighting in the east began when Stalin attacked Asoka on Turn 89, and three turns later Justinian's own separately planned invasion kicked off as well:

This was the 2 vs 1 scenario that much of the community had been predicting before the game began. Asoka was leading on the scoreboard by a wide margin, and yet that wouldn't matter very much if he proved unable to hold his cities when attacked by multiple hostile neighbors. The biggest issue here was finding a way to crack the Indian defenses. Asoka's heavy pursuit of religious and cultural stuff had left him with some very sturdy cities, and at the moment neither Justinian nor Stalin had catapults available to crack those defenses. Pataliputra was a true fortress city, up on a hill with 60% cultural defenses and with a flatground killing field stretching off to the east. Stalin spent a lot of time attacking there and had little success at this stage of the game. Justinian benefitted from Stalin pulling Indian defenders up to the north and experienced better results; he captured Illinois without much fuss on Turn 94, and then struck a more serious blow to Asoka by capturing Varanasi, the city just to the southeast of Asoka's capital, on Turn 102. It looked like Asoka was in the process of dying and would see the early exit that so much of the community (myself included) had been predicting.

But then a funny thing happened on the path to Asoka's demise. The Indian leader managed to recaptre Varanasi (which was surrounded on all sides by Asoka's culture), and then Justinian unexpectedly signed peace. The city of Illinois remained in Byzantine hands, but that was the only permanent loss. Now Asoka was only fighting a duel with Stalin, and with his powerful defenses that was enough to stabilize the situation and avoid further territorial defeats. Elsewhere, the diplomacy of the game was shifting slightly, as Islam spread to Shaka's territory and the Zulus converted. Stalin had already converted to Justinian's Christianity, leaving us with three civs following Islam in the west, two civs following Christianity in the east, and Asoka the only practitioner of Buddhism in the center. Shaka's newfound Islamic faith may have been a factor in his decision to sign peace with Wang Kon, which he did shortly after striking a major blow to the Koreans with the capture of P'yongyang:

This was surprisingly rational behavior from an AI leader, and triply so coming from nutcase Shaka. The Zulus were overextended here, and by signing peace they were able to secure their military gains. Note as well that Shaka was dead last on the scoreboard despite fighting this largely successful war against Wang Kon. That was due to the abysmal economy of the Zulus, with Shaka sitting in dead last in terms of GNP and techs researched. Doing nothing but build units the entire game makes for a good military strategy in the short run, but serves an empire poorly in the long view. This peace treaty would give Shaka an opportunity to stay peaceful for a while and work on developing his internal economy and infrastructure and LOL NO of course not, this is Shaka, he declared war on Asoka ten turns later. What else were you expecting?!

With Shaka joining the war against Asoka, the Indians were back into a dire situation once more. Their military forces were beaten down and exhausted by this point in time from so much warring, and with Shaka opening up a new front in the north, it was becoming impossible to hold the defensive lines that had stood for so long. Stalin had a major breakthrough as he finally captured the hill fortress of Pataliputra on Turn 134. This was playing out just the way that the community had expected. Now all we needed was for Justinian to join the war as well, reaping the easy benefits of the collapsing Indian empire and snowballing from there to a top finish of some kind. Justinian was "Annoyed" with Asoka and it seemed like only a matter of time. There was a huge Byzantine army moving around the map, and we watched with puzzlement as it began circling around to the north through Russian territory, and then turned west into Zulu domains. Where was this force going? Umm, Justinian, you're supposed to attack the Indians right now. What are you doing? Oh no, don't tell me...


Wang Kon's cosmic power of trolling surged to a new level as he somehow induced Justinian to march his army completely around the world to launch this new attack. Rather than participate in the straightforward and easy destruction of his nextdoor rival, Justinian was instead launching this insane crusade against the Koreans. It made absolutely no sense, as is seemingly the case for everything involving Wang Kon. Furthermore, the Koreans and their advanced research had very strong defensive capabilities, between the presence of longbows and maces and hwatchas. And of course, Wang Kon's Protective trait added extra City Garrison and Drill promotions to all of his Archery units. Combined together with the absurdly long supply lines from back in the Byzantine core, Wang Kon had little trouble defending against the much larger Byzantine forces. In fact, Justinian would make no progress at all in this war, while continuing to send giant armies to their death for dozens of turns on end. This was almost like the real Crusades, heh. It was a colossal waste of time and resources. Justinian was literally throwing away the game here - what was he doing?!

This conflict in the west was a lifeboat for the sinking fortunes of Asoka. He was about to gain further assistance from the forgotten leader in this match, as Hammurabi marched an army completely across Indian territory to declare war on Stalin. The Russian leader was caught off guard by this new threat, and the invading green forces captured the city of Vladivostok:

All across the map, the first playoff game appeared to be descending into a wild form of chaos. Justinian was off fighting and dying in Korea to no strategic purpose whatsoever. Stalin was now the recipient of a 2 vs 1 situation, desperately trying to hold off both Asoka and Hammurabi at the same time. While Stalin was able to recapture Vladivostok when the Babylonian army wandered elsewhere, he was unable to hold onto Pataliputra, which was taken by Hammurabi and then liberated back to Asoka once again. Even worse, Shaka signed peace with Asoka after taking a single Indian city. The Russians were left alone to face the storm, and it was their core cities that were now the ones under siege. It was all frustrating to watch for those of us who had picked Asoka to experience an early exit, as a game that had appeared relatively straightforward to predict was now falling apart at the seams. There had been some debate in the forum thread for AI Survivor Season Three as to just how predictable these games actually are, with several of the more frustrated viewers throwing out the notion that these games are essentially random and dice rolls would do as well as careful picks in terms of guessing the outcome. I disagreed with that notion, but this game wasn't helping my case with the crazy twists that it had been taking here in the middle stages of the action.

Slowly though, order began to restore itself to this contest. Stalin was able to secure a peace treaty with Asoka on Turn 151, and that allowed the Russian forces to concentrate on fighting against Hammurabi. This was enough to stabilize that particular conflict before Stalin lost any of his own core cities on a permanent basis. Then Shaka decided that he wanted to renew another conflict, and since he continued to share the same Islamic faith as Wang Kon and Hammurabi, Asoka was the clear target:

This attack against Asoka was more successful, with the Zulu forces capturing the city of Lahore in the west, then Aryan in the north. After that, Asoka was able to rally his forces and hold back any further territorial losses, with Shaka once again cashing out his gain with another peace treaty a dozen turns later. However, Asoka was slowly losing several key cities during these repeated attacks. Madras had fallen in a previous war to Shaka, and this latest bout had seen two more cities join the Zulu ranks. The purple Indian borders were slowly being pushed back, starting on the western side of the map and moving eastward. Given Asoka's continued poor diplomatic situation, this boded ill for his longterm chances of survival.

Meanwhile, Stalin had also managed to secure a peace agreement with Hammurabi. That was exactly what the Russians needed, as it provided a path out of a useless and destructive conflict. Hammurabi might have been better served here to keep the war going, as propping up Asoka was in his best interests. While the Babylonians had now climbed all the way into first place on the scoreboard, that was a tenuous position to be in if the other high peace weight civs were to fall. Still, Hammurabi had a solid chance at this point to find his way to a Cultural victory if he could continue to survive for another 100 turns. Justinian was still fighting his endless war against Wang Kon and making no progress; if anything, the Koreans were winning the war, somehow, in defiance of all logic. Wang Kon even built the Taj Mahal at this point for the free Golden Age, as he continued to lead the rest of the field in technology against all odds. It seemed as though Justinian was going to be trapped in a pointless war for the rest of the game.

At least, until Asoka made a huge mistake:

Perhaps encouraged by Wang Kon, Asoka took the fateful step of declaring war on Justinian. That put the Byzantines on the wrong side of a 2 vs 1 situation... except not really, since the Koreans were only defending their own territory and had no interest in crossing the continent to fight in the east. What this really did was change Justinian's AI targetting from going after Korean cities to pursuing Indian ones, and with the logistics in a much more manageable position, these were cities that he had a realistic chance of capturing. Note as well the huge mistake of declaring war on a Byzantine civ that was one turn away from finishing Guilds research and unlocking cataphracts. That's not a good fight to pick in Civ4, and Asoka was throwing himself into the arena at one of the worst possible moments.

The news kept getting worse for Asoka. Stalin piled into the conflict as well two turns later in a separate invasion, with large continents of Russian knights serving as the primary shock troops. Justinian also finally buried the hatchet with Wang Kon on Turn 189, and finally - finally! - the game's stupidest war came to a close. Now the Byzantines could focus purely on the Indian war effort. Or at least that's what he could have done, as Justinian decided to sign peace with Asoka a few turns later before capturing any cities. This was a mistake from Justinian, as everything had been set up for him to grab a bunch of cities from a collapsing civ. For the second time in this game, he refused to take the easy conquests and signed peace at a silly moment. Instead it was Stalin and Shaka who would take the spoils from this war, with Shaka declaring war yet again (the fourth time against Asoka) on Turn 199. Their combined efforts had little trouble beginning to tear apart the rapidly sinking Indian empire:

Asoka had barely been at peace since the early stages of the game. He had repeatedly been attacked by multiple opponents, and this final series of body blows was simply too much. One city after another began to fall like dominoes, and increasingly it looked like only a matter of time until he left the game. Stalin took the first couple of cities in this war, only for him to get stuck in a situation outside Bombay where there was only a single trebuchet bombarding down the city defenses, thus forcing the main Russian stack to sit in place doing nothing for almost two dozen turns while the castle walls were chipped away. This gave Shaka the chance to secure a bunch of gains for himself, and in the end, the Indian empire was effectively partitioned between the two leaders in rough halves. Stalin landed the killing blow on Turn 224:

As a result, Asoka ended up being the First to Die after all, but in a way that hadn't been expected ahead of time. Justinian's decision to waste so much time and effort attacking Wang Kon had greatly delayed the demise of the Indians, and Justinian had missed out on getting the territorial spoils that probably should have gone to him. If we were to replay this game a bunch of times, I'm sure that in most of them Justinian would have been part of the conquest of Asoka and snowballed from there into an unstoppable position. Instead it was Shaka who now topped the scoreboard, and although the Zulus had a lot of territory, they still remained badly behind in technology. Don't be thrown off by Shaka's beeline to Rifling technology; he still lacked everything in the Renaissance era that wasn't part of a direct path to the best guns possible. Amazingly, Wang Kon was still the tech leader at this point. Yes, Wang Kon - why is everything involving him so crazy?

That's not to say that the Koreans were doing well though. Justinian had renewed his war with Wang Kon yet again on Turn 217, and now that he had cataphracts in tow, he was finally able to make progress. Wang Kon was heading towards Rifling tech but still needed a good bit of time to reach it. In the meantime, it was cataphracts against longbows, and that's a very bad matchup for the defenders. The fortress of Seoul that had withstood so many previous Byzantine attacks finally fell to the invaders on Turn 223:

The loss of his capital seemed to deflate Wang Kon. He had always possessed the least territory in this game and compensated through having larger and stronger cities than his competitors. With his best two cities taken away from him at Seoul and P'yongyang, Wang Kon looked to have lost the will to continue fighting and trolling. He was reduced to a battered shell just looking to hold out as long as possible. The vultures in this game also began circling around the weakest link, with Shaka joining in the war shortly after the elimination of Asoka. The biggest Zulu death stack had about 150 units, and although a lot of them were outdated medieval stuff, that was an awful lot of military force. More Korean cities began to fall to these combined forces. And then the third member of the low peace weight trifecta decided to join in the fun, with Stalin also declaring war and bringing over his Russian cossacks. Justinian was the one who claimed most of the spoils of this war, due to a combination of fighting from an earlier date and simple luck on who attacked individual cities at the right time. However, Stalin's late war declaration was enough to allow him to sneak in at the last minute and claim the prized kill for himself:

Goodbye and good riddance, Wang Kon. I know that a lot of the viewers on Livestream enjoyed the Korean leader's antics, but all of the insanity associated with Wang Kon had been irritating to me. I still hadn't forgiven him for messing up my predictions in the opening round, when his spaceship win ended up costing me a ton of points in the contest. This game had gotten plenty weird as well, with Justinian's bizarre cross-continent invasions that seemed to throw everything out of whack. I suppose it had worked out in the end though, as the Byzantines now had a series of colonies located along the northwest coast, including the Pyramids inside Seoul that would come in handy for civic swaps. One of the viewers pointed out something interesting here: the warfare in this game had stopped being crazy as soon as Wang Kon was eliminated. And he was right! It was the Byzantine/Korean war that had thrown this game into a tailspin and threatened to unravel it completely. With Wang Kon removed, we were returning back to easily understandable patterns in how the AI behaves.

That was bad news for Hammurabi in particular. The other high peace weight leaders had both been removed from the game, and Hammurabi had a huge target on his back as a result. We might have made it 250 turns into the game, but the Babylonians still needed a lot more time if they were to reach a Cultural victory. I think Hammurabi had three cities above 15k culture at this point, which left him with another 35k still to go. All three of his competitors were "Annoyed" with Hammurabi, and all of them have aggression ratings that are well above average. It was only a matter of time until one or more of them decided to invade. Shaka was the first to cross the border with an attack on Turn 265, followed by Stalin's own separate incursion three turns later. Hammurabi had been building a bunch of Modern age wonders like Hollywood and the United Nations, and he was caught completely unprepared for these attacks. Not that there was much he could have done anyway. The first cities were already falling to the attackers when Justinian joined in to complete the trio:

The feeding frenzy was now on in earnest, with the three attackers each trying to select the tastiest morsels to devour. Given the large number of quality wonders that Hammurabi had built over the course of the game, including key targets like the Statue of Zeus and Sistine Chapel, who ended up with which cities would have important implications for the postwar situation. Stalin seemed to have the most success attacking along the western coast, as he secured all of the cities located along the water. Shaka started off a bit slow, then rallied to cut through the heart of Babylonian territory, taking four of the last five cities from Hammurabi. The Zulus ended up with the bulk of the spoils from this campaign. Justinian only picked up a single city, which was largely due to his late entry into the war. In the end it was Stalin who once again picked up the final blow, collecting his third kill for this match:

Geeze Stalin, slow down there buddy. He might not have started each war, but he certainly finished them. Combined with the three kills he picked up in the opening round, Stalin had now reached six total kills in Season Three. For point of comparison, Huayna Capac was leading the field with eight total kills across the entirety of the first two seasons of this event. And Stalin had almost equalled that in just two games. Could he pick up yet another kill in this game before it came to an end?

Now we were down to just the three low peace weight leaders remaining, Justinian and Shaka and Stalin. The Zulus had gained the most from this latest war and now stood clear of the other two leaders in overall score, with Shaka racing out to a lead of roughly 500 points. However, just having a lot of territory and population doesn't automatically translate into a win in Civ4, it only makes it more likely to happen. None of these three leaders were particularly close to a victory condition, and everything other than Culture looked to be in play here. Shaka was sitting at roughly 40% land area and population, which meant that if he could conquer one of the other two leaders he would likely have enough for a Domination victory. On the other hand, Shaka was significantly behind in technology due to his poor infrastructure, and if this game were to remain peaceful and play out to a Spaceship ending, he was unlikely to close the gap and launch before one of his two rivals. There was also a Diplomatic angle to consider; Justinian and Stalin were both "Friendly" with one another due to shared religion and mutual military struggle bonuses. If either one were to be nominated in the United Nations against Shaka, their combined votes might be enough to secure a Diplomatic win. Hammurabi had been building the UN before he was killed, but he hadn't finished it and none of the surviving civs had Mass Media tech yet. We were going to have to watch where the wonder was built closely.

The strong friendship between Stalin and Justinian also caused them to sign a Defensive Pact shortly before Turn 300. That was a very big deal indeed, as it meant Shaka couldn't attack either one of his remaining competitors without fighting both of them at once. Shaka was likely stronger than either of them individually, but taking on two opponents seemed like a risky play. Would he be aggressive enough to opt into a new attack?

This was Shaka, so the answer to that question was "yes", naturally. This was going to be quite a show, the largest civ in the game against two smaller but more technologically advanced opponents. As far as where everyone was located on the tech tree, Shaka had Railroads and was in the process of researching Assembly Line tech to unlock infantry and factories. Justinian and Stalin both already had these techs, with Stalin having recently finished Industrialism tech and Justinian now researching it. With everyone sharing the ability to train infantry and machine guns, I wondered if the tanks that the Russians and Byzantines had just unlocked would make a difference in this massive struggle.

Big endgame wars like this can be difficult to watch. They aren't like earlier wars where each side has one main stack that goes on the offensive and tries to capture cities. Instead, there are hundreds of units moving about on each side, and the struggle can best be understood in the aggregate. We tracked the ebb and flow of this war by watching which cities were captured by each side, and tracking the power rating of the combatants on the bar graphs. At first it seemed as though Shaka had gained a march on his twin opponents. The Zulus concentrated on the Russian cities in the southwest that had formerly been under Babylonian control, and Shaka took them out one at a time like a row of falling dominoes. There was less movement in the east, where all of the leaders had their core territory and making any progress in warfare was difficult. At the height of his advance, Shaka had opened up a score lead of almost 1500 points, and it seemed possible that he might roll over Stalin and Justinian en route to a Domination victory.

However, all of the fighting inevitably began to take its toll on Shaka. The key unit in this round of warfare was tanks, which Stalin and Justinian both had and Shaka did not. Since both of them had only recently researched Industrialism tech, they had to start piling up their armor corps essentially from scratch at the start of the conflict. Over time though, those tanks were individually winning their matchups and promoting while Shaka's units were losing the 1 vs 1 unit combats. This began to show up in the military power bar graphs:

Shaka's huge lead in military strength was beginning to crater. By this point he was below Justinian in power and the Zulu yellow continued to plummet closer and closer to Russian red. Stalin and Justinian both began to accumulate some highly promoted tanks, and once they were running around with City Raider III armored units, there wasn't a lot that Shaka could do to defend himself. Even the AI can capture cities easily with mobile artillery to drop cultural defenses and City Raider tanks to apply the hurting against the defenders. Shaka was only a little bit behind in military technology, but that gap between him and the eastern leaders was enough to be decisive here. Stalin led the counterattack, starting with the ex-Babylonian cities where he rallied from a single city in the extreme southwest to recapture his initial losses. Justinian was surging in two places, near his Korean colony in the northwest and along the southern edge of the map down in the tundra. This was beginning to look like a huge mistake for Shaka.

After about 30 turns of warfare, Justinian cashed out of the war with a peace treaty:

He picked up the city of P'yongyang for his troubles. Justinian was now the score leader, although he was slightly behind Stalin in tech by a very slim margin, about one or two techs only. This was Justinian's chance to use the peace interval to close the gap with Stalin and attempt to launch a spaceship first. Both of the eastern leaders were now far ahead of Shaka in military strength, as the Zulus continued their slow collapse as one city fell after another. Stalin gave no indication of slowing down, and he continued to wrestle more territory from Shaka's hands. Around Turn 260 Stalin finished the Manhattan Project and the war began to go nuclear, although not nearly to the same extent as the madness that took place in the Wildcard game. Shaka had no defense against the Russian tactical nukes, of course, and they only served to speed along the war towards its inevitable conclusion.

Justinian's power had continued to spike upwards despite this peace interlude as the Byzantines continued cranking out more military units. We suspected that he was plotting a return to the same war, and sure enough the struggle was renewed on Turn 366. Now the Byzantines began lobbing their own nuclear weapons into Zulu territory in the vanguard of their armies, and the Zulu empire continued to crumble faster and faster. Stalin would end up getting the bulk of the territorial spoils since he remained at war for longer, but Justinian did fairly well for himself in the former Korean territory of the northwest. We wondered if Stalin would be able to claim this kill as well and wrack up a new AI Survivor record with four kills in a single match:

But it was not to be. Justinian claimed the final blow against the Zulus by taking Lahore, in a photo finish with the Russian armies outside this last city. This was fair payback for Stalin swiping the kill against Wang Kon earlier in the game, and a split with three kills for Stalin against one kill for Justinian was a reasonably accurate accounting of how the warfare had taken place in this game. As for Shaka, he had been knocked out of the playoffs in the most Shaka-tastic fashion imaginable, picking an unwinnable war against two opponents at once. Shaka seems to do a little bit better in this competition than the other true lunatic warmongers like Montezuma and Temujin and Napoleon, which is probably due to his superior (in comparison) traits and the fact that the Zulus are one of the better civs in Civilization IV. However, ultimately he's a little too aggressive for his own good, and every season he seems to get eliminated in the same fashion, picking one too many wars against the wrong opponents.

With only two leaders remaining, we now knew who would be moving on to the Championship match. The real drama shifted to the picking contest, as which of these two leaders managed to emerge on top would have significant scoring implications. Stalin and Justinian were extremely close in tech, with the Russians maintaining a lead of roughly one tech over Justinian. They were also researching at almost exactly the same rate, with both of them sitting a little under 2000 beakers/turn. Justinian seemed to get a leg up in the competition when Stalin stopped to research the useless Stealth technology... only for Justinian to also waste four turns chasing after the same target. Heh. Stalin had built Cristo Redentor earlier and that stopped him from losing turns in Anarchy when doing civic swaps, and he also managed to build the Space Elevator, which could come in handy on the actual spaceship part construction. Stalin's last spaceship tech was Fusion, and he decided to launch his spaceship with only one Engine component in a reprise of Charlemagne's opening round fiasco. However, Stalin had finished the tech tree three turns before Justinian, and the Byzantine leader was only halfway done the Stasis Chamber when the Russian spaceship launched. Justinian needed several more turns to finish that last part, and when it was finally done he... didn't launch at all? Huh? I guess the AI knew that there was nothing that could be done to catch Stalin's spaceship, and decided to just sit and wait for the end to come. The Russians arrived in Alpha Centauri on Turn 403:

That was a very late finish to this game and the third-longest game in terms of number of turns that we've had in AI Survivor. Stalin and Justinian both moved on to the Championship, and the Byzantine leader will have a chance to win his second overall title. I don't think it was an accident that these were the two leaders who ended up surviving to the end of this game. They shared a rock solid alliance for virtually the whole game, based on a shared Christian religion and lots of mutual military struggle bonuses. In a game where everyone else was willing to turn on one another, having a strong ally on one flank made the difference in these two outlasting the competition. This was the biggest reason why I picked the two of them to finish in first and second place, as I expected the eastern leaders to be fast friends throughout the game and take down Asoka together. It was a longer and bumpier ride to reach that point than I expected, but ultimately it did come to pass. That was reassuring from the perspective of these games being predictable as opposed to purely random. The way that I view these AI Survivor games is a bit like batting in a baseball game. The outcome of any one at bat is essentially random, but that doesn't mean that baseball as a sport is nothing more than chance, or that it's impossible to evaluate individual players. As the sample size of AI Survivor games continues to increase, it's becoming increasingly clear to me that some of the leaders really do perform better than others; strong leaders like Mansa Musa or Huayna Capac will consistently outperform dud leaders like Sitting Bull, maybe not in one individual game, but certainly over the long haul. We're not just rolling dice here, there are understandable patterns within the back and forth of these matches. You can claim I'm seeing meaning where none exists, but I won the picking contest in Season Two and I've been in the top ten (out of 200+ entries) for all of Season Three, so I don't think I'm completely off base here.

One last thing: throughout the game, I consistently kept misspeaking and confusing "Shaka" and "Stalin" because their names were so similar. Some of the Livestream viewers began keeping track of how many times this happened, and the tally reached almost two dozen by the time that the four hour stream came to a close. I hope you guys had fun - and I hope these two leaders never show up in the same game again!