Game Two 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 Two saw a fairly straightforward (if dominant) Gandhi spaceship victory, in what was probably the closest AI Survivor has ever come to seeing zero deaths. 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:
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: I'd been expecting this to be an exceedingly boring series of games, but thankfully it turned out that I was wrong. An average of three leaders died in each game, which is close to normal, while an average of 9.3 wars per game was also on the low side of typical. The high emphasis on religion that some of these leaders placed, and the randomness of war declarations with a lot of cautious and pleased faces, created a relatively dynamic feeling where it was difficult to know what would happen next. Indeed, wars were chaotic in this game, as every leader was torn apart by multiple rivals at least once. It turned out that Victoria was the best at navigating this chaotic environment towards a win, thanks largely to being the only leader with Financial. In games where no one snowballed (most of them), she was the best placed to grind out a spaceship win. The other main win-threatening leaders were Asoka and Gandhi, who each won four times, generally in games where Victoria was crippled in some way (either neglecting culture for too long, or being dogpiled). However, these leaders faced problems that Victoria did not, which prevented them from winning as often. In Gandhi's case, the issue was clear: his peaceweight was the furthest away from most leaders', which meant that he frequently two- or three-front wars. This is a bit counterintuitive, because we think of Gandhi as the 'most good' leader, but it turns out that in a game with zero low peaceweight meatshields, Gandhi was instead the 'least mostly-good' leader instead. In particular, Joao, Churchill, Victoria, and Washington all attacked Gandhi fairly often, and he needed to avoid these dogpiles to succeed. The problem for Asoka was also clear: he and Isabella were by far the most consistent enemies, and fought in basically every game. Asoka won most of these fights (sometimes by himself, sometimes with external assistance), but he was often left economically backwards by that point compared to Victoria and Gandhi, which meant he could struggle securing the actual win.
The other four leaders were less successful. Churchill looks like he did well, but literally all three of his wins came because he either sniped a diplo victory (from Gandhi in Game 2) or won spaceship when another leader stupidly pursued culture (Asoka and Victoria in Games 6 and 13). Additionally, several of his kills were sniped from other leaders, particularly in that early streak. In this sense, Churchill was excellent at scoring Power Ranking points, but was not often very strong himself - his true performance was probably only slightly above Washington's. Speaking of which, the American leader was also quite weak here, in this case due to a combination of a poor, jungled-bound start, and being culturally squeezed out by Gandhi repeatedly. This led Washington to begin the game-ending dogpile on poor Gandhi a lot of the time, but he rarely claimed many spoils for himself, and typically sat irrelevantly on the sidelines for the rest of the game. Joao shared the Western peninsula with Churchill, and was usually (though not always) the weaker of the pair. They fought semi-consistently as well, and Churchill usually got the better of that conflict - even when he didn't, Joao was typically way too far behind economically to compete with whoever else was left. Finally, Isabella was clearly the weakest of these leaders, despite finishing above Joao. Her inexplicable win came in Game 20, where she expanded much better than normal for reasons I couldn't discern - maybe Asoka lost a settler to the barbarians or something. Otherwise though, she almost never achieved anything of consequence beyond crippling Asoka, as she was terrible at spreading her religion which left her diplomatically isolated.
This game was interesting because as mentioned, I'd expected the lack of any genuine warmonger to create some very stagnant and bland timelines. Instead, most of the leaders were Cautious with each other due to three or four competing religions, which meant a lot of unpredictability in who got attacked each game, and how those wars went. For example, in Game 7 I watched Victoria, Joao, and Izzy all declare suicide wars that got themselves killed, before seeing Churchill kill three leaders in Game 8, and then Gandhi claim two scalps in Game 9, until Victoria stomped the next three games. What really stood out here was how differently the diplomacy could unfold in each game, and the strong ramifications this always had. Not what I'd expected from the 'lovefest' game, that's for sure.
Now for a look at the individual leaders:
Victoria of England
Wars Declared: 40
Wars Declared Upon: 19
Survival Percentage: 70%
Finishes: 8 Firsts, 1 Second (42 points)
Kills: 12
Overall Score: 53 points
Victoria scored the best here, but not because she did anything particularly impressive. Instead, Victoria would simply keep pace with everyone else for the first 150-200 turns, before pressing the Financial button and zooming off to space. I mentioned above that these game were a lot less static and predictable than I was expecting, but that was less true on Victoria's wins, since she depended on no one else snowballing before she could use her overpowered traits. Other than winning a stalemated game, Victoria could also do well by joining a dogpile on her neighbours, mostly commonly Gandhi but also occasionally Asoka or Joao. Indeed, Victoria declared the most wars of any leader at 40, though this reflects her strength as much as anything else. She tended to do well if these wars were started after she researched Construction, and badly when they were started beforehand. Victoria's bad games came when she got stuck in either unproductive 1v1 wars, or got dogpiled out of the game (e.g. Game 13), most commonly the former. Interestingly, her other big weakness was culture, and Victoria could sometimes fall behind early on and have to play catchup for the rest of the game because she didn't expand her borders for ages. That also wasn't super common though, compared to what we've seen from some of the warmongers. It's hard to create much more of a narrative than that for Victoria - she was the default 'win' leader if no one else snowballed, which was enough for a clear top score, but she certainly wasn't dominant either, playing very badly in some games (Vicky loved turning on the culture slider with 6 techs to go) - she was just the best of a bad bunch, I suppose. It is at least interesting that a leader known for being so inexplicably bad actually got quite unlucky here on livestream, and I'd be interested to see if that was the case in either of Victoria's Season One or Two games.
Asoka of India
Wars Declared: 14
Wars Declared Upon: 40
Survival Percentage: 60%
Finishes: 4 Firsts, 5 Seconds (30 points)
Kills: 7
Overall Score: 37 points
Asoka was the next-best leader, despite tying with Churchill, as he earned all of his wins and kills. Asoka's game revolved directly around Isabella's (they always founded 2/3 of the early religions), but he was clearly the stronger of the two, coming out on top of the Spanish in nearly every single game. This was partly because Asoka was simply better at Civ4 than Izzy, out-expanding and out-fighting her himself most of the time. But it was also partly because Asoka actually bothered to build missionaries and spread his religion around, while Isabella did not, which left him with a greater number of religious allies to call in. Asoka was regularly assisted by Washington, Gandhi, and Victoria in his crusade, and that was the decider in most of the initially-stalemated games. Unfortunately, by that point his economy was usually too damaged to truly compete with Gandhi or Victoria, unless they'd also both been slowed down somehow. Unlike Victoria, Asoka didn't get stuck in fruitless 1v1s very often, as he would either get dogpiled, or call-in allies himself. However, he did have a few disastrous ones, which was the case for almost all the leaders on this map (e.g. Games 2 and 9). In essence, either Asoka killed Isabella quickly and won the game, or he didn't and fell behind the other contenders, leading to either irrelevance or death.
Churchill of England
Wars Declared: 33
Wars Declared Upon: 12
Survival Percentage: 70%
Finishes: 3 Firsts, 4 Seconds (23 points)
Kills: 14
Overall Score: 37 points
Unlike Victoria, Asoka, and Gandhi, Churchill most certainly did not earn most of his points the hard way. Instead, the English sniped many kills from opponents, and all three of his wins featured a much stronger leader who bungled things in some way. In Game 2, Churchill used the United Nations to swipe victory from under Gandhi's nose, while in Games 6 and 13 he followed the Charlemagne playbook of launching a late spaceship while a stronger leader fruitlessly pursued culture. What Churchill did well was expand, as he generally claimed more land than his direct opponent Joao in the West. However, from there he was never able to compete economically with the other leaders, even when he was able to genuinely absorb someone else (usually Joao, but sometimes Asoka or Gandhi). It meant that beyond scamming wins in this way, Churchill was never going to actually 'earn' a victory. In fairness, no other leader won in this way even once, so maybe Churchill was doing something right instead of just getting lucky. I have my doubts, however, and he certainly wasn't a top performer either.
Gandhi of India
Wars Declared: 17
Wars Declared Upon: 49
Survival Percentage: 45%
Finishes: 4 Firsts, 3 Seconds (26 points)
Kills: 7
Overall Score: 33 points
Gandhi was pretty squeezed in this game, both in terms of land and diplomacy, which made it impressive when he dominated the scoreboard for the first 150 turns of each game. He expanded by far the best, and also had the best economy, virtually without exception during this period. Unfortunately, that was also where it typically fell apart for Gandhi, as his just-different-enough peaceweight made him the victim of a dogpile over, and over, and over again. This is easily seen by Gandhi's high defensive war rate, despite his early demises in a number of games. Washington was the most consistent leader to ruin Gandhi's day, followed closely by Victoria, and then there was the slight possibility of literally anyone else (except Asoka) joining as well, to create a truly unwinnable 3-or-4 vs 1 situation for poor Gandhi. As a result, he often collapsed, and this was a direct cause for Victoria's strength. However, when Gandhi wasn't dogpiled, he was able to win dominant peaceful victories, just as we saw on livestream. It was classic feast-or-famine from the Indian leader, which is pretty much exactly what we've come to expect - it just turns out that we got 'feast' on livestream.
Washington of America
Wars Declared: 25
Wars Declared Upon: 20
Survival Percentage: 65%
Finishes: 0 Firsts, 4 Seconds (8 points)
Kills: 10
Overall Score: 18 points
Washington was very weak on most of these games, which surprised me given his isolated starting position. He expanded terribly, and once I looked for the reason why I realized that his land was simply not very good, particularly his capital, and was also heavily jungle-bound. It meant that Washington got off to a slow start, and his traits certainly weren't going to help him catch up economically after that! Washington's only option was to try and snowball militarily, and to his credit he did try, attacking Gandhi very often and Isabella less so. Neither of these could be successful without help though, and even in the rare cases where Washington picked up a lot of extra land (through the dumb luck of city captures, not any particular skill) he was hopelessly far behind economically and couldn't compete in the lategame (e.g. Game 5). I don't know how many games we would have had to run to see a Washington outlier victory, but I wouldn't be surprised if it was over 50. Instead, his main consequence was ruining Gandhi's game, a fruitless proposition from which neither of them prospered.
Isabella of Spain
Wars Declared: 34
Wars Declared Upon: 31
Survival Percentage: 35%
Finishes: 1 First, 1 Second (7 points)
Kills: 5
Overall Score: 12 points
Isabella was arguably the weakest leader on this map, scoring two thirds of her points in the final game, including her only victory and over half her kills. I honestly have no idea why Isabella was so strong in this game compared to the others; I assume Asoka must have lost a settler to the barbarians, but even that doesn't really explain it. Anyway, in that game, Isabella quickly stomped all over Gandhi, then Asoka, then Washington, en route to... a passably early spaceship victory (it was bang-on average for finish dates). Meanwhile, in every other game, Isabella would flail about venting her religious frustration on someone (usually Asoka) without actually bothering to build any missionaries and spread it herself. This led her to be economically behind and diplomatically isolated, which resulted in the lowest survival rate of the entire group here. She declared stupid, suicidal wars frequently, and was attacked by other leaders regardless. Much like Washington's main consequence was to ruin Gandhi's games, Isabella's was to ruin poor Asoka's, and I'm kind of staggered that she won a game. Just another item of evidence to add to the 'Isabella is a bad leader' pile, I guess.
Joao of Portugal
Wars Declared: 24
Wars Declared Upon: 16
Survival Percentage: 60%
Finishes: 0 Firsts, 2 Seconds (4 points)
Kills: 5
Overall Score: 9 points
Joao was dealt a rough hand here, with almost no food resources that weren't jungle-bound. This meant that Churchill was able to claim most of the contested land between them, which generally left Joao as one of the weakest leaders on the board. It was basically worst-case scenario for the Portuguese, whose traits are entirely designed around snowballing early on. Joao did try to make the best of a bad situation, sometimes by attacking Churchill (unlikely to succeed), joining an anti-Gandhi coalition (occasionally succeeded, as in Game 1), or just staying out of the way and waiting for the wildcard game. He never even came close to winning, but I'm not sure that a better leader would have either from this position, so it's hard to hold that against him. As a result, it was hard to get a sense for Joao's ability as an AI here, as he was dealt such a rough hand by the map generator. At least he didn't regularly commit suicide like Isabella? That's got to count for something.
This was a weird game. It wasn't overwhelmingly predictable like I'd been expecting, with the sheer number of 'cautious' faces proving to significantly randomize the diplomacy in some games, even if there were overall trends. However, they also weren't particularly interesting, with it usually being clear who was going to win after the first unpredictable set of wars, and a set of generally passive and inert leaders. As far as how well our livestream game reflected these overall results, Gandhi wasn't the expected winner, but he clearly wasn't a total fluke either. The same is true of Asoka in second, although Churchill was pretty unlucky to die first on livestream. Perhaps the weirdest part of the livestream game was that we got very close to no one dying - that didn't happen at all in these games! Instead, I never even rolled a game where only one leader died, let alone a 'None' game like we've seen in the past. This shows that the high peaceweight leaders are certainly willing to fight and kill each other, even if it's not quite as common as the warmongers doing so.
This was a shorter Alternate History, but I hope it still brought something interesting to your day!
Cheers ~ Amicalola 
