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. The second Wildcard game saw Brennus run away from the rest of the field while winning a Diplomatic victory that were mere turns away from Domination. 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. This particular set of alternate histories were run by TheOneAndOnlyAtesh with assistance from Eaupxs I. Forgott - many thanks for spending so much time on this task! 
Atesh posted the resulting data from the alternate histories and then discusses some of the findings below 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:
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.)
TheOneAndOnlyAtesh: These results were noteworthy in how the leaders with the worst reputations ended up churning out the most respectable results among this field. Brennus and De Gaulle together formed the Dynamic Duo, securing the final playoff slot in 14/20 replays; a member of the tandem only failed to place in TWO games. Meanwhile, Washington was also a pleasant surprise, winning twice in dominant fashion (and nearly winning another game) despite his seemingly hopeless peaceweight situation. Conversely, the four seeded or borderline seeded leaders were either major disappointments (Sury, Willem) or complete clowns (Cyrus, Alex save for his final five games) in this setup. What gives?
First, land was quite important, both for long-term potential and short-term barbarian management. As the official game had shown, Brennus' coastal empire was barb proof, and the Celtic leader could continue working his best tiles even as barbarians were running amok in his lands. Furthermore, Brennus only had to worry about barbarians from the South, as his east was protected by the ocean while Sury and Cyrus covered his western and northern fronts respectively:
The Celtic leader's ability to keep his cities at size 6-8 for twenty more turns than the rest of his competition was instrumental in his dominance.
However, safety from Raging Barbs was not the be-all-end-all. For example, although Cyrus was virtually immune to the barbs (he was the only leader to never lose a city to the barbarians), he was still the worst performer on the map. Meanwhile, De Gaulle greatly struggled with the barbs yet was the clear cut secondary contender behind Brennus. To be frank, this was a dream diplomatic setup for both De Gaulle and Brennus, who both happen to be "demonic evil" leaders who cannot plot at Pleased. Both leaders were highly likely to found and spread religions, helping them form fruitful alliances, join favorable dogpiles, and avoid mutually destructive conflicts. Even with rival faiths, the two were predisposed to like each other from the get go, meaning that they rarely fought save for the occasional late-game showdown. This turned this map into a two-man show, while all of the other leaders had serious issues in this setup, whether it was global diplomacy, exposure to barbs, or innate personality flaws.
My notes on most of the Dynamic Duo games were some variation of "[De Gaulle/Brennus] coasts to victory while the rest of the world fights meaningless wars". The livestream game essentially followed this exact same pattern. A secondary outcome was for one member of the tandem to establish religious supremacy and take out all the infidels to claim victory. Otherwise, the only realistic way for a game to deviate from the typical outcome was for a different leader to execute a successful solo conquest. This took place in five of the six outlier games, with the other one (Alex's sole Game 20 victory) being a complete troll where Alex used the UN to snipe Brennus' playoff spot from under his nose.
This ended up being one of the bloodier (at least during the first 250 or so turns) Alternate Histories sets, with only one game having a single digit war counter. Most of these wars were protracted stalemates, and it frequently took a dogpile or a massive military edge to bring any one leader down. Due to Raging Barbarians, the commerce-poor map, and the combination of tough fighters and mediocre economic leaders in this field, these were some incredibly slow games; the average game length was 340 turns! The community was incredibly lucky to witness a quick official game, as there were too many games in this set where I was zoning out for thirty minutes while the clear game leader was taking 100 turns to go to space, Pleased-locked out of any sort of conflict. Like with the official game, it was extremely obvious who was going to win by Turn 200 or so. Perhaps that was befitting of a game in which only the winner advances to the next round.
Unsurprisingly, Spaceship was the near runaway victory condition. This was especially the case in the Dynamic Duo games; 10/14 of their games ended at Alpha Centauri. Interestingly, four of the six Domination finishes came from outlier games, while only one of the outlier games ended in space at the incredibly late date of Turn 379 - a clear sign of the land quality disparity between the Dynamic Duo and the rest of the field. Of the three Diplomatic wins, two of them were early coronations of Brennus and De Gaulle while the other one was the aforementioned Alex troll Diplo win. Because of how split the religions were and the sole culturally inclined leader being the hapless Willem, these games never came close to a Cultural finish despite their length.
Now for a look at the individual leaders:
Brennus of the Celts
Wars Declared: 50
Wars Declared Upon: 14
Survival Percentage: 80%
Finishes: 9 Firsts, 7 Seconds (59 points)
Kills: 20
Overall Score: 79 points
Truth be told, this result was less about Brennus' merits as a leader and more about how much better his starting position was compared to his rivals. Alongside his hedge against barbarian aggression and excellent diplomatic setup, Brennus' capital was located on the only major flood plains site in a world where commerce was relatively scarce. I think almost any leader would do well in his spot as long as their diplomacy was favorable. In light of this, it actually felt like Brennus underperformed here. Many of his wins reminded me of my days as an eight-year-old playing on Settler difficulty, hitting Enter before around Turn 350 I sort of just won.
Simply put, Brennus' failures in this set came from another leader just outplaying him. Maybe a leader teched better, or spread his religion better, or managed to get a more effective conquest off. Very occasionally, the barbarians hampered his game, either seizing/razing one of his settlements or making him too slow to expand. Nevertheless, the final playoff spot was handed to Brennus on a silver platter, and although he took advantage effectively, it was still pretty obvious why he is a below average leader for AI Survivor purposes. However, one can admit that 1) Brennus was a legitimately terrifying and opportunistic military leader, and 2) overall he was able to back up his legitimately strong livestream performance, playing like a mini-Justinian of sorts.
Best Performance: Easily Game 13, the only game where Brennus really teched like an above average economic leader.
Worst Performance: Flubbing a cross map war against a really weak De Gaulle so bad in Game 12 that De Gaulle was actually conquering core Celtic cities from the other side of the map.
There Can Only Be One Orange Civ Award: CONTINUOUSLY marching to the other corner of the map to attack Willem in these games, even though the much closer Washington was his Worst Enemy. Not that this was a bad thing - Willem's cities were often highly developed and full of juicy wonders, and a Dutch conquest frequently sealed Brennus' victory. In fact, conquering Willem was often key to any leader's success in this setup.
De Gaulle of France
Wars Declared: 46
Wars Declared Upon: 27
Survival Percentage: 70%
Finishes: 5 Firsts, 8 Seconds (41 points)
Kills: 14
Overall Score: 55 points
I had picked De Gaulle to win due to the following factors: I liked his numerous rivers; his land seemed to have a great balance between food, production, and commerce; his two neighbors were juicy dogpile magnets for a leader that likes to dogpile; and most of all, I liked his diplomatic situation, especially his ability to get along with most leaders despite possible religious differences. Indeed, the replays vindicated my belief for these exact reasons.
Although the stats ostensibly suggest that De Gaulle was the junior partner in the Dynamic Duo, the French leader was arguably more impressive than his Celtic counterpart. While his land was nice, it was not as good as Brennus' (especially long-term), and it was also MUCH more exposed to the barbarians due to the vast tundra region to his north. Despite this, De Gaulle's wins felt more earned than Brennus'. Furthermore, he came quite close to winning more games, especially Games 1 and 7. Place him in Brennus' spot, and I think De Gaulle likely wins at least 12 games. With Willem being essentially unviable in this setup, De Gaulle was clearly the most capable techer of the bunch.
De Gaulle's disappointing Actual Game performance should offer some clues as to how he could falter. First, the barbarians could sometimes hamper his game so much that by the time he caught up, another leader had already snowballed out of control. De Gaulle actually lost his capital to the barbs a few times in these replays:
If a non-Brennus leader also had a good game, it was not great for De Gaulle. In the livestream, that leader was Washington. Finally, the French could get overly caught up in inconclusive warring, especially against the tough-to-crack Washington. Nevertheless, De Gaulle certainly redeemed himself after his awful Alternate Histories performance in his Opening Round, and one should not underestimate him if his diplomatic position seems favorable.
Best Performance: Having a borderline Financial-esque spaceship win in Game 4, by far the best economic performance by any leader in this set.
Worst Performance: Getting absolutely abused by the barbarians in Game 10 before ultimately crumbling to Washington.
Willem's Ghost Award: De Gaulle should have won Games 15 and 16, but avoided Rifling (and Military Tradition) for WAY too long, especially in Game 16. This directly led to his elimination in those games. Amusingly, both of these were games in which he had killed Willem.
Suryavarman of the Khmer
Wars Declared: 39
Wars Declared Upon: 46
Survival Percentage: 30%
Finishes: 2 Firsts, 0 Seconds (10 points)
Kills: 12
Overall Score: 22 points
There was a steep dropoff from the Dynamic Duo into the next tier, consisting of two leaders who had two mostly spectacular (and believable) wins but who accomplished little else in most of the other games. Unlike De Gaulle, Sury did very little to repair his reputation hit from his Opening Round Alternate Histories disaster. This was not entirely his fault, as Sury had perhaps the worst spot in the entire map as far as surviving Raging Barbarians was concerned. The problem with Sury's starting position was that he had both a massive tundra region to his south AND a large jungle area to his north. In most games, Sury started so slowly that he was irrelevant by Turn 100. The Khmer leader was never even remotely as terrible as he was in the Opening Round - for example he usually researched Archery in a timely fashion and never had a "build too many warriors" ailment - but more often than not, it was too much for him to handle. Making matters worse was that Sury was Brennus' most common religious rival, and the Khmer could ill afford getting on the wrong side of the strongest leader on the map.
When Sury did hit however, he struck gold. Both his Game 14 and Game 16 victories showcased his peak potential, where he proved himself to be a fantastic military leader while just good enough economically to contend. He also came quite close to winning Game 9, nearly catching up in tech to the Dynamic Duo. Although those games showcased his sky high potential, this season has greatly exposed Sury's flaws, and I am unsure if he is truly Pool One quality. Hopefully he can regain some community goodwill in Season Nine.
Best Performance: Easily Game 14, where Sury took most of the spoils from the American partition, before conquering two foes at once en route to a respectable (for this map and circumstances) spaceship finish.
Worst Performance:
Wang Kon Award: Waiting more than fifty turns to attack a weak Cyrus in Game 16 when he easily could have won Domination, nearly blowing a victory against a Brennus when at that point I had become exasperated by seeing Brennus win games he did not seem to deserve to win.
Washington of America
Wars Declared: 18
Wars Declared Upon: 86
Survival Percentage: 20%
Finishes: 2 Firsts, 1 Second (12 points)
Kills: 7
Overall Score: 19 points
The fact that Washington score-wise was solidly middle of the pack was both a testament to how tough he was and to how embarrassingly bad the non-Dynamic Duo low peaceweights were in this setup. Despite getting attacked an unfathomable 86 times - a number admittedly inflated by last minute vulturing -, having a 40% First To Die rate, and being a peaceweight 8 leader in a central position (the other peaceweights were 0, 0, 0, 2, 3, 4), Washington performed exactly as well as Sury, matching the Khmer leader with two wins and one close call. Even their survival rates were essentially identical - Sury survived two games with just one city.
There were two main reasons why Washington had an outsider's chance. First, his diplomatic situation was not as hopeless as it might have seemed. As it turned out, Washington's proximity to Sury and Brennus was a blessing in disguise, as he frequently adopted their religions like in the Actual Game. This was never more apparent than in Game 15, where Washington and Brennus formed an unlikely Jewish partnership and partitioned the rest of the world in one of the bloodiest world wide religious conflicts I had ever witnessed in an AI game.
The other reason was that Washington excelled at expansion, almost always being the first or second leader to reach an X amount of cities no matter how much he was struggling with the barbs. This helped him be a formidable military leader, and although Washington was a prey surrounded by predators, he was less a sheep in the slaughter and more a gigantic wildebeest requiring a pack of hunting dogs to take down. A major boon to the Americans was that their land had a lot of Ivory, meaning that they had access to War Elephants at the perfect time. In Game 2, Washington SOLO CONQUERED an equally as large Cyrus (leading to the Persian leader's singular First To Die) to completely snowball out of control; he repeated this in Game 10 with De Gaulle instead being the man standing between Manifest Destiny. Meanwhile, in Game 15 Washington got most of Sury's land and leveraged that into a scoring finish. He really should have won that game, except there were some trollish shenanigans due to faraway barb city captures and the inability to sign Open Borders. I somewhat wish we got to see a Washington victory in the real thing, because it would have fit greatly with the "revenge of the nerds" theme of this season.
While Washington was extremely impressive considering his circumstances, his success ironically revealed the exact reason why his overall track record is suspect. While he is an excellent military leader, he has the unfortunate distinction of being a high peaceweight with below average economic capabilities. This means that in favorable diplomatic fields, he generally is unable to compete, while in situations where he shines, he has to climb an uphill battle due to bad diplomacy. With this set, Washington's historical mediocrity should no longer be a mystery.
Best Performance: Both his wins were equally impressive. Washington nearly had four kills in Game 2, and he won a relatively quick Domination victory in Game 10 WITHOUT EVER REACHING INFANTRY TECH. His Cavalry actually won a war against Brennus' Tanks to secure Domination!
Worst Performance: Washington handled the barbs poorly in Game 11 and eventually went out with a whimper.
Joao Award:
City of Snow Angels Award:
Alexander of Greece
Wars Declared: 58
Wars Declared Upon: 15
Survival Percentage: 50%
Finishes: 1 First, 1 Second (7 points)
Kills: 9
Overall Score: 16 points
Willem of the Netherlands
Wars Declared: 20
Wars Declared Upon: 61
Survival Percentage: 10%
Finishes: 1 First, 1 Second (7 points)
Kills: 6
Overall Score: 13 points
After the Washington/Sury tier, we have these two, who notched one extremely flukey victory each alongside one backdoor "everyone else died" 2nd place finish, while otherwise being the two scrubs of the map. The silver lining for these two was that they were obviously screwed by this game's map generation. In Alex's case, he suffered from the same issue as Sury: he was surrounded on all sides by a raging barbarian factory that made his first 100 turns into a nightmare. Making matters worse was the location of Sparta; in every replay, Alex settled his 2nd city in a tundra spot that made his empire into a BAR BAR BAR magnet. After watching the first 15 games, I believed Alex to be a completely hopeless leader in this setup. However, in the final five games, where Alex did a much better job withstanding the barbarians, I observed something noteworthy: Alex proved to be a quite significant player when he was able to launch. As it turned out, his diplomatic situation was excellent, he had two great targets as neighbors, and this was a setup that rewarded good fighters. If this were just a regular Opening Round game, I do think Alex may have made some actual noise here. Of course, he also won the final game, albeit in a troll fashion Brennus threw by building the UN when a large Alex and Cyrus were best buds. Imagine if that had been the livestream game!
Willem, meanwhile, had the worst diplomatic situation in this setup - yes, it was worse than Washington's. While Washington had a small chance of making friends due to his location, it was a lonely world for Willem, who had the double whammy of having a relatively high peaceweight and a different religion, leading to his 3:1 defensive-to-offensive war ratio. Moreover, Willem had a major barb problem of his own, except that these barbarians had green borders instead of black ones. Alex was a constant thorn in Willem's side, attacking him somewhere between Turns 70 and 100 before Willem could get any sort of Financial machine kicking off. To add salt to the wound, if Willem was finally starting to get ahead enough in tech to push back against the Greeks, BRENNUS would come barrelling in from the opposite corner of the world with a gigantic army. The most tragic example of this was in Game 1, where Willem looked like the game runaway until he was suddenly backstabbed by the rest of the world (including his religious ally Sury) and collapsed. If Gandhi was in Willem's spot, there would have been little difference in results.
This is not to say that Willem was a tragic or sympathetic figure. In fact, much of his own issues were self-inflicted. Considering how often Alex flailed around, one win (where Alex played at an all-time bad level) and a 10% survival rate - worse than Washington's - is absolutely unacceptable. Willem's deficiencies - his difficulties making allies, his issues timing his aggression well, and, of course, his propensity to avoid military techs - were on full display. What a maddeningly inconsistent season retrospective on a maddeningly inconsistent leader - on one hand, he outscores Huayna Capac despite bordering Shaka and having rough land, and on the other hand, he lays a turd like this set.
Best Performances: Other than the obvious (including Willem having the only pre-Turn 300 win in the set), Game 1 was the only game where Willem was an actual contender despite a relatively strong Alex, and Alex was quite good in Game 19, being well-positioned for 2nd until he suicided his Cavalry into Brennus' Modern Armor.
Worst Performances: See awards.
This is NOT Sparta Award:
With enough replays, I think we would have seen Alex die to the barbs.
Willy Nilly Award:
Willem lost this game. Wonder how? Here's a hint:
Cyrus of Persia
Wars Declared: 38
Wars Declared Upon: 20
Survival Percentage: 55%
Finishes: 0 Firsts, 2 Seconds (4 points)
Kills: 9
Overall Score: 13 points
Although Cyrus technically tied with Willem in score, he absolutely belongs in last place here. Cyrus did not just fail to win a game: in twenty games, there was not even ONE moment where I registered the thought that Cyrus might win. (His two 2nd places were backdoor, much like the Actual Game.) He was a complete non-factor - Sullla could have deleted him and his territory from the game and no one would have noticed. All evidence points to Cyrus being an above average leader - he has a season silver to his name and had an incredible Opening Round Alternate Histories - and he had virtually no barb issues. So why was he so utterly irrelevant?
Unfortunately, security from the barbs came with tradeoffs. While Cyrus had about eight cities worth of land all to himself, those were not exactly high quality cities, especially from a long-term scaling perspective. Cyrus was often boxed into these eight or so cities, and he had neither the land, the traits (CHA/IMP are great for war and expansion but not for econ), nor the shrines to execute a builder strategy, so he had to fight his way out. The problem? 1) His two neighbors were perhaps the two toughest fighters in this setup; 2) Other leaders benefited more from the Washington dogpile; 3) Cyrus was frequently Pleased-locked out of conflict - his diplo was almost TOO good. Thus, Cyrus was a completely inert leader here, kind of like a low peaceweight Freddie, accomplishing little of note and not even being useful as an attack dog. There were just two times he ever made an impact: the first was dying so horribly to Washington that he paved the way for an American upset, and the second was helping Alex cheat Brennus out of a playoff spot. Sorry Cyrus, but perhaps you should not have sicced your awful Immortals on your natural ally Julius Caesar in your Opening Round game.
Best Performance: Surely, you must be joking...
Worst Performance: A major reason Cyrus singlehandedly died to an equal strength Washington in Game 2 was that he was building way, way too many Immortals, which predictably got shredded by Washington's Elephants. This season made it pretty evident that these extremely mortal units are major poison pills for the Persians.
Usually, our Wildcard games are much larger, so it was interesting to see how a normally sized game would function in Wildcard settings (namely Raging Barbarians). While there has been much debate over how much Raging barbs truly affects games, I think their effects were much more pronounced with fewer leaders, as there was much less fog-busting to be done. Nevertheless, one should not overly weigh barb safety when making predictions. After all, the two worst performing leaders were arguably the two who were most sheltered. As it turned out, shelter from barbarians often meant little land, which was not at all conducive to winning.
Otherwise, this was an interesting spotlight on perhaps some underdiscussed leaders, and it was refreshing to see how leaders like Brennus, De Gaulle, and Washington could shine in AI Survivor.
