Civ4 AI Survivor Season 3: Game Four Alternate Histories


Introduction

Game Four Alternate Histories Spreadsheet

One of the recurring features of past seasons of AI Survivor have been our "alternate histories", running additional iterations on the same maps to see if the same events would play out again. Season Three Game Four saw Justinian absolutely torch the competition in one of the most dominant victories AI Survivor has ever seen. Was that something which would unfold in each game? This was a topic that called for more investigation with alternate history scenarios. Following the conclusion of previous seasons of AI Survivor, I had gone back and investigated some of the completed games and found that they tended to play out in the same patterns over and over again. While there was definitely some variation from game to game, and occasionally an unlikely outcome took place, for the most part the games were fairly predictable based on the personality of the AI leaders and the terrain of each particular map. Would we see the same patterns play out again and again on this particular map?

The original inspiration to run these alternate histories came from Wyatan. He decided to rerun the Season Four games 20 times each and publish the results. The objective in his words was twofold:

- See how random the prediction game actually is. There's a natural tendency when your predictions come true to go "See! Told you!", and on the contrary to dismiss the result as a mere fluke when things don't go the way you expected them to (pleading guilty there, Your Honour). Hopefully, with 20 iterations, we'll get a sense of how flukey the actual result was, and of how actually predictable each game was.

- Get a more accurate idea of each leader's performance. Over 6 seasons, we'll have a 75 game sample. That might seem a lot, but it's actually a very small sample, with each leader appearing 5-10 times only. With this much larger sample, we'll be able able to better gauge each leader's performance, in the specific context of each game. So if an AI is given a dud start, or really tough neighbours, it won't perform well. Which will only be an indication about the balance of that map, and not really about that AI's general performance. But conversely, by running the game 20 times, we'll get dumb luck out of the equation.

Wyatan did a fantastic job of putting together data for the Season Four games and I decided to use the same general format. First I'll post the resulting data and then discuss some of the findings in more detail. Keep in mind that everything we discuss in these alternate histories is map-specific: it pertains to these leaders with these starting positions in this game. As Wyatan mentioned, an AI leader could be a powerful figure on this particular map while still being a weak leader in more general terms. Now on to the results as run by Amicalola:

Season Three Game Four

Game One | Game Two | Game Three | Game Four | Game Five

Game Six | Game Seven | Game Eight | Game Nine | Game Ten

Game Eleven | Game Twelve | Game Thirteen | Game Fourteen | Game Fifteen

Game Sixteen | Game Seventeen | Game Eighteen | Game Nineteen | Game Twenty



(Note : "A" column tracks the number of war declarations initiated by the AI, "D" the number of times the AI is declared upon, "F" the points for finish ranking, and "K" the number of kills.)

Amicalola: And I'd thought Hannibal dominated the last game - sheesh! Justinian ruthlessly crushed this game over and over again, even more convincingly than Hannibal had in the last game. He cracked 100 points, averaged nearly 1.5 kills per game (including his bad ones), and could generally outfight, out-tech, and out-civ everyone else here. Now, to be fair, this was partly the result of a beautiful starting location in a fertile river valley - I suspect any leader would have done alright here. But Justinian also played to his strengths repeatedly, spreading his religion effectively to create a strong economy and diplomatic situation that allowed him to pick his battles. It took multiple opponents to bring the Byzantines down, and though that did happen occasionally, it was rare, and Justinian was clearly the deserving winner on livestream. More surprisingly, his equally clear junior partner was Saladin. Unlike Justinian, Saladin had a pretty poor starting position in terms of food, and the Arabians were rarely among the leaders coming out of the landgrab phase. Instead, Saladin seemed to simply outwork his neighbours repeatedly, solo-conquering leaders from tech parity. Louis in particular was a common target, as Saladin often killed off the French during the midgame, securing himself second place in the process. It also helped that he and Justinian shared Theocracy as a favourite civic, making them unlikely allies in a lot of games. No other leaders came close to either Justinian or Saladin, who were clearly the most likely duo to advance. Catherine came the closest though, with an uninspiring result of "one win and a random smattering of seconds and kills." She would often be a clear 'third place' leader during the games, sometimes scoring a kill or two on Darius/Roosevelt, before being eliminated by either Justinian or Saladin later - in this way, the livestream game reflected the AH results perfectly.

As for the other four leaders, in a sense they were the reason for Justinian and Saladin's strength, all performing badly. Darius was the best, securing himself two wins in atypical situations, and dogpiled into oblivion the rest of the time. His peaceweight situation combined with a central position was a recipe for disaster, and he was usually out of it by Turn 120 - that he placed fourth should tell you something about how bad the other three were. Louis was a total disaster, leveraging a fluke early conquest of Constantinople to a cultural win in Game 2, and otherwise flailing ineffectually. He had major issues with both fighting wars and running the culture slider way too early, and his main role was acting as a vehicle for Saladin to get strong. Montezuma and Roosevelt were even worse. The Americans had a nice, isolated position, but failed in the expansion phase, inevitably left in the dust by the Byzantines, Russians, and Arabians (and often the French and Persians too). They were unable to catch up through warfare or technology, leaving Roosevelt a sitting duck to be taken out in the midgame. Montezuma was the exact opposite, constantly declaring wars on anyone and everyone until he picked a fight he couldn't win and was eliminated. The only question for the Aztecs was whether that was their first war or their fourth. Both of these leaders showed exactly why they're regarded as bottom-tier for AI Survivor purposes, and I doubt anyone's opinions of them will be changed any time soon.

It's hard to create much more of a narrative for this game, because it was one of the simpler ones I've seen. Justinian killed everyone, while Saladin outfought a tough starting position to score second, Catherine was a wildcard, and everyone else generally sucked. Anything else was an outlier, and this is probably one of the clearest games to predict that we've ever had. If you picked these leaders to advance back during Season Three, you can pat yourself on the back - anything else was a major misread.

Now for a look at the individual leaders:

Leader Summaries


Justinian of the Byzantines
Wars Declared: 33
Wars Declared Upon: 45
Survival Percentage: 80%
Finishes: 14 Firsts, 4 Seconds (74 points)
Kills: 28
Overall Score: 102 points

Well, that was quite the performance. Justinian combined a strong AI personality with a disgusting starting location, and it turns out that was enough to basically dominate everybody. Even Saladin, the clear second-place, almost never came close to taking a win off this guy. Justinian had everything going for him this game: he had the most cities, the biggest cities, a religion that spread like wildfire, an excellent diplomatic situation, and the best economy. He had options: going west and crushing Roosevelt, doing the same to Montezuma (if the Aztecs hadn't already donated their land), or south to hit Darius. Justinian could also just tech up in peace for the first 200 turns, before massacring a rival or two en route to an early spaceship. The only leader who could contend for early cultural victories (Louis) was garbage here, and usually killed off. The weakest leader in the game (Montezuma) declared war on Justinian a lot, but that was usually a good thing as it meant Justinian could easily conquer more land, just as we saw on livestream. Meanwhile, the only other consistently strong leader in the game (Saladin) absolutely loved Justinian, and pretty much refused to declare war on him (they rarely fought, but when they did it was Justinian who started it). If anything, I'm surprised he didn't win even more games, but there were a few stinkers where absolutely everything went wrong. In Game 2, Louis captured Constantinople quite early with lucky combat rolls (the war alone was an outlier, let alone that result), and kept it for over 100 turns - amazingly, Justinian still managed a strong second there. In the consecutive Games 13 and 14, Montezuma and Roosevelt became the power couple the world needed, and somehow killed Justinian outright. Finally, there were a couple of games where Darius was not dogpiled, and that never ended well. However, these were all clear outlier scenarios, and in a larger sample size I expect Justinian would win even more than 70% of the time.


Saladin of Arabia
Wars Declared: 38
Wars Declared Upon: 40
Survival Percentage: 80%
Finishes: 2 Firsts, 10 Seconds (30 points)
Kills: 18
Overall Score: 48 points

In a word: consistent. Saladin didn't have many amazing games, but he also had very few bad ones, which is impressive from a central position that seemed quite weak. Saladin played a very consistent set of games, always expanding out to the same 6-7 cities and then teching up to catapults, before fighting a rival (unless one attacked him first). He only won twice, in Game 10 because of a stupid Justinian cultural push and in Game 14 because Justinian was inexplicably torn apart by Montezuma and Roosevelt. However, he came second another half the time, and was generally the second-strongest leader on the map. In particular, the Arabian economy was often surprisingly decent considering Saladin's poor traits; I think this was a combination of a lot of rivers, and a willingness to expand borders immediately. He also profited repeatedly from attacking all three of Darius, Montezuma, and Louis. In general, Saladin was the jack-of-all-trades, master of none, as he could tech and fight adequately, and he was excellent at sucking up to Justinian. Despite them often founding different religions, Saladin and Justinian often built a rock-solid alliance over a combination of Theocracy, peaceweight 4, and an inability to declare war at Pleased - they very rarely fought, although if they did it was a disaster for Sal. It left Saladin able to ride Justinian's coattails in game after game, which was the perfect situation. On the other hand, his bad games were caused by a few possibilities. Sometimes, Louis was unusually strong, which was bad news for Sal (Games 2, 7 and 11). The same could happen with Montezuma (Games 13 and 17). Finally, as mentioned Saladin occasionally got on Justinian's bad side, and that was usually no contest (Game 5). But overall, this was an impressive performance from Saladin that backed up his Season Seven breakout, and I'm a little sad he couldn't work the magic on livestream here.


Catherine of Russia
Wars Declared: 49
Wars Declared Upon: 21
Survival Percentage: 70%
Finishes: 1 First, 4 Seconds (13 points)
Kills: 13
Overall Score: 26 points

Catherine was a strange bird in this one. She wasn't weak, per se, but she also wasn't very strong. There were several games where she was able to conquer a neighbour or two (Darius and/or Roosevelt), but by that point she was simply always behind Justinian, and nearly always Saladin as well. She was very clearly the 'middling' leader in this one, who rarely died early, but who also could barely win games. Amazingly, unless she could organize a dogpile on Darius, Catherine couldn't even win that fight 1v1, such as in Games 2, 6, and 10), although the games where Darius got mauled were Cathy's best. Her only win was the flukish Game 19, where basically everything went right for the Russians - Justinian was weakened through repeated two-front wars; Saladin had the same issue; and Catherine got to slowly roll over Darius, Roosevelt, then Saladin without really facing any opposition. This was exceedingly unusual though, and I don't think Catherine got unlucky overall. She just couldn't compete with the big dogs in the lategame, and thus despite rarely dying, she could also rarely place. It was a weird one for a leader who normally snowballs or perishes, with Catherine's most likely result by far being relegation to the Wildcard Game, but I guess that's what happens when you give a leader an isolated but barren start.


Darius of Persia
Wars Declared: 23
Wars Declared Upon: 41
Survival Percentage: 25%
Finishes: 2 Firsts, 0 Seconds (10 points)
Kills: 6
Overall Score: 16 points

Sheesh, talk about feast-or-famine! Darius had exactly two good games, in which the only shared event was the other leaders failing to dogpile him. When that happened, Darius was highly competitive! In Game 6, he crushed Catherine in an early war, before barely beating Justinian in a climactic battle, while in Game 13 Darius got very lucky that the runaway leader was Montezuma and not someone more competent (Montezuma had somehow managed to get to 55% land area, while being 2.5 eras behind in tech ). But in the other 18 games, Darius pretty much got torn apart by some combination of Catherine, Saladin, Louis, and Justinian, which was reflected in his low survival rate and high defensive war count. Interestingly, it was Saladin who benefited the most from Darius' weakness rather than Catherine, which surprised me, but that was just about Darius' biggest contribution here - land to Saladin. Otherwise, it was exactly what we'd expect from Darius: he could win if he got exceedingly lucky, and otherwise he got torn apart.


Louis of France
Wars Declared: 29
Wars Declared Upon: 26
Survival Percentage: 45%
Finishes: 1 First, 2 Seconds (9 points)
Kills: 3
Overall Score: 12 points

Due to often founding one of the later religions, or converting to Justinian's, Louis spent most of the games locked in a death struggle with Saladin, and unfortunately he was clearly the weaker of the pair. He almost always got killed by Saladin in a 1v1 fight, and needed to find allies elsewhere on the map, which simply didn't happen very often. Even when they happened, Louis rarely benefited much beyond staying alive, with whatever ally he could find taking most of the spoils. For whatever reason, the French were shockingly incompetent militarily, only scoring 3 kills in this entire set - half as much as Darius! Louis also suffered from turning on the cultural slider way too early, almost always beyond Rifling, which sometimes tipped his Arabian wars from stalemates into straight losses (such as Game 14, where Louis threw away a likely victory). Louis' only win came in the highly atypical Game 2, in which he whacked Justinian very early and sniped Constantinople, before religion created an ironclad friendship between the two leaders anyway; this provided Louis with enough of a buffer zone to actually get that elusive cultural victory, though even that was a near-run thing. That was a clear outlier though, and most of the time Louis floundered ineffectively until someone (usually Saladin) put him out of his misery). This was a disappointing performance from the Season Seven Champion, who failed his cultural gambits despite an isolated starting position and low-peaceweight advantage.


Montezuma of the Aztecs
Wars Declared: 43
Wars Declared Upon: 25
Survival Percentage: 15%
Finishes: 0 Firsts, 1 Second (2 points)
Kills: 3
Overall Score: 5 points

Although he technically scored a point more than Roosevelt, Montezuma was probably the worst leader here. The only game where he was even remotely successful was the crazy Game 13, where he inexplicably performed far better than usual against Justinian, taking Constantinople by Turn 100 (I suspect wildly unlikely combat results were responsible here), and then partitioning the rest of the Byzantines with Roosevelt. He then somehow managed to eat both Roosevelt and Saladin, becoming hugely dominant in food and land (only 5% away from domination), before a pretty standard Darius who was over two eras ahead in technology unequivocally beat the Aztecs into submission. Otherwise, Montezuma was weak in every game, as incapable of fighting wars as he was at staying out of them. His weakness was a major drive in Saladin and Justinian's success, as it was usually them who were able to annex the Aztec lands from a fairly early stage. That was about as consequential as it got for old Monte, who sucked just as much as we've come to expect. I'm not sure we can call Montezuma the worst leader for AI Survivor, but he certainly must be in the conversation...


Roosevelt of the Americans
Wars Declared: 22
Wars Declared Upon: 38
Survival Percentage: 25%
Finishes: 0 Firsts, 1 Second (2 points)
Kills: 2
Overall Score: 4 points

...as must this guy. Roosevelt was pretty much the exact opposite of Montezuma in playstyle, but the same in results. He stayed in the corner not doing anything until someone stronger realized he was there and killed him. This was usually Catherine, but it could also be Justinian. Much like Montezuma, Roosevelt's only good game occurred when he and Montezuma were able to partition the Byzantine lands in Game 14, and he was then able to slowly beat Catherine in a series of 1v1 wars (that he was very lucky no one else interrupted). In all other games, he simply lacked the economic or military muscle to make much of an impact beyond being a speedbump for better leaders. The evidence continues to mount that Roosevelt is simply bad, because he performs poorly in basically every metric for AI Survivor, and this game contributes further to that narrative.

Conclusions

Much like Game 3, this one went pretty much entirely as expected, except for the leader who got second place. Justinian was by far the most likely leader to win, and Darius was a very reasonable pick for first-to-die. However, it was Saladin, not Louis, who was the king of the East, and he was quite unlucky not to at least survive until the Wildcard Game. It was nice to have a few games where the expected result actually came to pass - at least sometimes we get what's actually supposed to happen!

This was a shorter Alternate History, but I hope it still brought something interesting to your day!

Cheers ~ Amicalola