Civ4 AI Survivor Season 6: Game Three Alternate Histories


Game Three 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. Game Three saw an intense struggle between Gilgamesh and Ramesses over the key city of Memphis, which eventually resolved when Augustus avenged the conquest of Ramesses by knocking Gilgamesh down into the Wildcard Game. 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:

Season Six Game Three

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.)

Game Three turned out to be the story of three leaders: Gilgamesh, Ramesses, and Augustus. All three of these leaders had a serious chance to win each repetition of the map and they monopolized the victories between themselves by taking home all 20 of the alternate histories. The other three leaders - Shaka, Asoka, and Qin - had essentially no chance to win and never found themselves in a position to claim the top spot. It was a bit crazy how thoroughly the eastern side of the map dominated over the western side of the map. In addition to the eastern three leaders winning every victory, they also scored 44 kills as compared to 7 kills for the western three AI leaders and suffered the First to Die fate a mere 4 times as compared to 16 times. Just as we watched on Livestream, the drama in this scenario came from which one of these three leaders would emerge on top while the other three leaders were a complete afterthought. It's not entirely clear to me why this was the case; perhaps the quality of land was better on the eastern side of the map? There were several rivers running through the center of the map that Gilgamesh and Ramesses tended to control and there was also that ridge of mountain tiles that may have interfered with Shaka and Asoka. Whatever the case, this was a lopsided map and only three leaders had any realistic chance to win.

The "finish points" based on first and second place finishes were extremely similar between these three leaders: 44 for Ramesses, 42 for Gilgamesh, and 38 for Augustus. Ramesses and Gilgamesh in particular were almost identical in this regard, both of them claiming eight victories apiece along with two runner up finishes for Ramesses and one for Gilgamesh. Augustus only managed four victories but made up for it with by far the most second place finishes, nine of them in total, making him the most likely leader to advance into the playoffs. Augustus also had the lowest elimination rate of any leader at only 10% which was another beneficial side effect of his isolated starting position. The main thing that differentiated these three leaders in terms of scoring was the number of kills that they claimed, with the more aggressive Gilgamesh scoring far more kills than his competitors. Even though Gilgamesh and Ramesses had the same number of victories, Gilgamesh claimed 21 kills to only 8 kills for Ramesses which accounted for the gap between their final tallies.

The crux of the map typically came down to the same border struggle that we observed in the real Game Three. Gilgamesh and Ramesses would settle their respective border cities of Eridu and Memphis in the identical spots across every replay of the map. Although other features of the settling patterns changed from game to game, these cities always remained in the same spots:

One of these cities would almost always become a Holy City but it didn't necessarily have to be Memphis as we watched in our game. Gilgamesh also founded a religion in many games and it would inevitably pop up in Eridu right there on the border with Egypt. The outcome of these games would then turn on the struggle for this disputed region along the Sumerian/Egyptian dividing line. If Gilgamesh could push through Ramesses' territory and snowball ahead through a complete conquest of Egypt, he would usually be the winner of the match. Sometimes he was able to manage this at an early date with a successful sneak attack; this resulted in two crushing Domination victories in Game #9 and Game #16. Both of them were early finishes but Game #9 was unreal by resolving with a Turn 203 (!!!) victory for Gilgamesh. I watched it happen and I can still hardly believe it. Gilgamesh could also snowball his position through an early conquest of Shaka but that was riskier because he was knocking out one of the other low peace weight leaders on the map. Sometimes this could work as in Game #13 however it could also backfire by weakening Gilgamesh and allowing more time for Ramesses and Augustus to develop in peace.

If Gilgamesh wasn't able to accelerate ahead through conquest, the games typically ended in a victory for Ramesses or Augustus. We had already seen this from Augustus on the Livestream but I was surprised to watch how strong Ramesses could be in many of these alternate histories. The Egyptian leader had a lot of space to expand and some rich territory once he could chop down the jungles. He was the most likely leader to win if Gilgamesh was diverted into long warring with Shaka or Asoka and there was time to develop the central part of the map. Ramesses turned out to be a strong leader when he wasn't getting invaded by Gilgamesh constantly. He executed perfect timing attacks to take down Gilgamesh in Game #4 and Game #17 and in several other games simply ran away from the rest of the field technologically while the others were fighting. This was a bit of a high risk / high reward sitation since Ramesses was also First to Die four times but if he could manage to stay peaceful for long enough he had an excellent position for expansion.

Augustus had the safer spot on the map amongst the high peace weight leaders, much less likely to be eliminated while also somewhat less likely to claim an outright victory. The territory immediately surrounding Augustus' starting position was weaker than what Ramesses had due to a lack of rivers and a big patch of desert to the north of Rome. This caused Augustus to get out to a slower start and his economy could be a little rough during the early turns. He also lacked easy expansion opportunities since Augustus almost never went to war with Ramesses and there were usually a few barbarian cities in the south-central portion of the map. The path to victory for Augustus was exactly what we watched on Livestream: going to war with Qin and conquering China while Ramesses blunted the aggression coming from Gilgamesh. This scenario didn't happen quite as often as the more straightforward routes open to Gilgamesh and Ramesses but it still manifested often enough for Augustus to win four times. He was the safe pick on this map and did extremely well for himself.

As for the three western leaders, none of them really did much and it was unusual for any of them to finish in a top two spot. Asoka was the most likely to come in second, not so much because of his own innate strength but because Ramesses and Augustus both won't plot war at "Pleased" relations and he could ride their coattails to an occasional runner up spot. He had exactly two good matches, Game #10 and Game #16, both of which saw Asoka kill the even weaker Qin to claim enough land for that second place spot. Shaka had a bad starting position with cramped territory and a poor diplomatic environment. He couldn't get along with the two high peace weight leaders near him but he also couldn't get along with Gilgamesh either, and Shaka almost always lost the duel with his Sumerian neighbor when they came to blows. Shaka was frequently First to Die and when he did survive it was almost always as an irrelevant civ. Finally, Qin was consistently the weakest leader for reasons that aren't entirely clear to me. While he didn't repeatedly crash his economy as we watched in the actual Game Three, he never seemed to be good at any of the key aspects of Civ4. He didn't expand well, he didn't tech well, and he lost almost every war that he fought. Qin was the most popular pick for second place in the picking contest before the game started and it's hard to believe how badly he failed on this map. Perhaps his local terrain was worse than we thought because he couldn't get out of his own way on this map.

Now for a look at the individual leaders:

Gilgamesh of Sumeria
Wars Declared: 51
Wars Declared Upon: 28
Survival Percentage: 70%
Finishes: 8 Firsts, 1 Second (42 points)
Kills: 21
Overall Score: 63 points

Gilgamesh was narrowly the highest-scoring leader on this map thanks to picking up notably more kills than Augustus and Ramesses. The three of them were close enough to be within the margin of error and more repetitions of this map easily could have shifted one of them into the top spot. As detailed above, Gilgamesh had an excellent chance of winning through military conquest of some kind, whether that was the most straightforward path through Ramesses or some kind of combination of absorbing Zulu or Indian territory. Although Gilgamesh started out the game in a cramped part of the map, the land that he did have was excellent in terms of quality, lots of floodplains tiles and luxury resources. He was almost always tops on the scoreboard at the end of the landgrab phase thanks to his Creative culture and strong local terrain. The big problem for Gilgamesh was that he absolutely had to snowball through successful military action; his land alone was too small to win peacefully and Gilgamesh's aggressive personality wouldn't allow him to sit back and try the builder path. To put this succinctly, when Gilgamesh succeeded in his wars he usually won the game and when he didn't he usually found himself outside the top two spots. He was the most interesting leader to watch in this group and achieved some shockingly fast conquests when it all came together for Sumeria.

Augustus of Rome
Wars Declared: 36
Wars Declared Upon: 12
Survival Percentage: 90%
Finishes: 4 Firsts, 9 Seconds (38 points)
Kills: 15
Overall Score: 53 points

Augustus had the most isolated start on the map, causing him to be the runaway favorite for runner up status while still having a decent chance to take first place. He was almost never targeted for attack with only 12 defensive wars faced, far lower than anyone else in the game except for Shaka (and the Zulus were rarely attacked because they started so many wars themselves). Augustus typically found himself sitting in second or third place on the scoreboard at the end of the landgrab phase, off to a slower start than Gilgamesh or Ramesses due to weaker terrain at his capital. This was balanced out by having more total land area available, however, and Augustus frequently wound up with the most cities overall as he pushed to the west and battled the barbarians camping the southern reaches of the continent. His best games occurred when Gilgamesh and Ramesses stalemated along their frontier and the Romans could achieve a midgame conquest of Qin's weak China. Long and ineffective wars elsewhere were perfect for the slow-starting but eventually-strong Roman setup on this map. Augustus wasn't as peaceful as his reputation might suggest, starting 36 offensive wars and scoring the second-most kills at 15. Now 36 offensive wars might not sound like that many as compared to the 51 offensive wars of Gilgamesh but that number should be compared to the 11 offensive wars of Asoka and the 10 offensive wars of Ramesses. The other high peace weight leaders pretty much never started a fight whereas Augustus was much more willing to throw down with his rivals. The Roman leader was the "safe" choice on this map and had great odds to take one of the two playoff spots.

Ramesses of Egypt
Wars Declared: 10
Wars Declared Upon: 47
Survival Percentage: 65%
Finishes: 8 Firsts, 2 Seconds (44 points)
Kills: 8
Overall Score: 52 points

Ramesses benefited from lush local terrain and plenty of room to spread out with a large empire. When paired with his economic traits and pacifistic tendencies, he found himself in a great position to build his way to some kind of peaceful victory condition. On the rare occasions where Ramesses didn't get stuck in early warfare, he was often able to race out to an unassailable tech lead and crush his rivals from a position of superiority. The problem for Ramesses was his northern neighbor Gilgamesh who invaded the Egyptians in almost every game. Ramesses was dragged back down to the rest of the pack over and over again as he was forced to defend against aggression from Sumeria. Sometimes Gilgamesh would prove to be too powerful and Ramesses would be knocked out of the game completely; this was responsible for his four First to Die eliminations on Turn 163, Turn 127, Turn 166, and Turn 157. More often the two of them would be stuck in a lengthy protracted struggle which could tip in either direction depending on what was going on elsewhere on the map. This was essentially what we watched in the real Game Three, Ramesses slowly losing to Gilgamesh across 200 turns but delaying the Sumerians long enough for Augustus to surge ahead in tech and win. There were better outcomes for Ramesses, however, games where he won outright victories over Gilgamesh and became completely unstoppable as in Game #11 or Game #17. Ramesses almost never started any wars but still ended up winning by Domination three times as he responded to enemy attacks with extreme prejudice. He was pretty awful at scoring kills though with only 8 kills despite 8 overall victories. This wasn't bad luck either as Ramesses was often willing to let collapsing opponents off the hook with peace treaties. Overall, Ramesses was the higher risk / higher reward version of a high peace weight leader, more likely to win outright than Augustus while also more likely be completely destroyed. Sharing a border with Gilgamesh makes for a dangerous lifestyle.

Asoka of India
Wars Declared: 11
Wars Declared Upon: 53
Survival Percentage: 35%
Finishes: 0 Firsts, 4 Seconds (8 points)
Kills: 3
Overall Score: 11 points

There was a vast gulf in terms of performance between the top three leaders and the bottom three leaders. Asoka wound up being the fourth place leader according to our scoring but he wasn't even remotely close to that top trio. He was never in a position to claim a victory and didn't even survive to the end of the game in fully two-thirds of these matches. Asoka was First to Die seven different times and found himself getting bullied by stronger leaders in game after game. He was the frequent target of Shaka or Gilgamesh or both of them together. There were quite a few games where I wrote "Asoka dies in 3 way evil dogpile" as he was pounded by all three of his low peace weight neighbors and disappeared without having a prayer of survival. It was only in the games where Asoka was left alone by his neighbors that he had any chance to spread out beyond 5 or 6 cities and that almost never happened. To his credit, Asoka had two games where he managed to eliminate the even weaker Qin. He pulled this off in Game #10 and did it again in Game #16 where Asoka also managed to conquer Shaka too. Those were two games where Asoka actually earned his second place spot. These outcomes were exceedingly unlikely, however, and in the 18 other games Asoka was never able to claim anything other than his small starting core of territory. His other two runner up finishes in Game #1 and Game #12 were purely due to Augustus smashing stronger opponents and knocking them out of the second place spot. Asoka was a weak leader on this map and didn't seem to have much of a chance even if diplomatic luck broke his way.

Shaka of the Zulus
Wars Declared: 42
Wars Declared Upon: 13
Survival Percentage: 50%
Finishes: 0 Firsts, 2 Seconds (4 points)
Kills: 3
Overall Score: 7 points

Shaka's performance in this game is what happens when an insanely aggressive warmonger gets put in a spot where his civ is too weak to conduct offensive warfare: he stomps around declaring lots of wars and huffs and puffs without actually achieving anything to show for it. Shaka's most frequent target was Asoka and he typically spent most of the early game ramming axes and swords into walled Indian cities on top of hills. Needless to say, this rarely led to much success. Shaka needed an ally of some kind to make serious progress against Asoka and he sometimes was able to get assistance from Gilgamesh. Of course, when this happened Gilgamesh was often the one taking Asoka's cities and claiming the elimination credit. Alternately, Shaka also invaded Gilgamesh pretty frequently as their border tension overrode their shared low peace weight. This also tended to go poorly for Shaka since Gilgamesh was significantly stronger and had better quality land. The only other target for Shaka's aggression was Ramesses and that wasn't much of an option given the greater distance and Egypt's powerful culture. Shaka was left with little to do aside from wildly flail about damaging one neighbor after another in random fashion. He started 42 wars that produced a grand total of 3 kills - not great! Shaka's aggression frequently backfired and he was the other favorite for First to Die, matching Asoka with seven such eliminations. Surprisingly he managed a 50% survival rate which probably overstates his actual odds of reaching the finish line on this map. Shaka had basically no chance to finish first or second and only grabbed the runner up spot when everything broke in his favor (as it did in the actual Game Three). He was a buzzing mosquito on this map, annoying for everyone else without having any real chance to be a threat.

Qin Shi Huang of China
Wars Declared: 27
Wars Declared Upon: 24
Survival Percentage: 35%
Finishes: 0 Firsts, 2 Seconds (4 points)
Kills: 1
Overall Score: 5 points

I still honestly can't figure out why things went so badly for Qin on this map but badly they went indeed across the 20 alternate histories. Qin never seemed to be able to snowball ahead into a position of strength in any of these games. Even in his two second place finishes in Game #5 and Game #18 he was unable to spread out beyond his initial starting area and simply rode the coattails of two Gilgamesh Domination victories. Qin certainly engaged in his share of wars, he simply didn't seem to win them for whatever reason. Perhaps this was due to having a weak economy as China struggled in this game to keep up in research. It was never as bad as what we watched on Livestream when Qin's teching completely collapsed but his economy wasn't exactly good either. While it's true that there was a lack of rivers in the Chinese core, Qin had plenty of room to expand and there was no reason for him to perform this poorly. He's been at least competent on other maps and this was a puzzlingly bad showing from a leader that we typically view as being mid-tier. Whatever was the case, Qin was never even remotely competitive for a top finish and he scored exactly 1 kill across 20 games. Asoka and Shaka did poorly and he was somehow even worse.


In the end, we wound up seeing a typical outcome for this map as the real Game Three went down one of the common paths in the alternate histories. We missed the "Gilgamesh smash" scenario and the "Ramesses economic dominance" scenario while instead getting the "Gilgamesh/Ramesses stalemate into Augustus tech victory" outcome. The actual Game Three was largely repeated in the alternate histories in Game #1, Game #6, and Game #12 which all looked highly similar to the Livestream performance. The finishing date, number of wars, and even the Domination result were both quite normal and in fact there were a lot more Domination outcomes that I had been expecting in the alternate histories. The oddest thing about the real Game Three was Shaka taking second place since the Zulu leader only managed that about 10% of the time. This was easily explainable, however: Gilgamesh was locked into one of the top two spots for 98% of the game and then dropped below Shaka in the final turns as he collapsed against Augustus. Change that one small detail and we would have had a completely normal result. Most of the alternate histories ended with the Gilgamesh/Augustus/Ramesses trio hogging both playoff spots and we narrowly missed that result in the game that we watched. The only people who were truly unlucky were those who backed Ramesses - there were many games where he was the top dog and Egypt was an excellent pick on this map that simply didn't work out in our single game sample size.

Thanks as always for reading, I hope you enjoyed this look back at Game Three!