Civ4 AI Survivor Season 6: Fantasy Retrospective

Fantasy Summary Excel Spreadsheet Download

One of the new features added for Season Six of Civ4 AI Survivor was the creation of a fantasy football style competition using the AI leaders to score points. Since the leader draw for the opening round took place two weeks before the first match of the season, everyone had the chance to study the upcoming games and decide which leaders they liked the most. One week before the start of the season, we held an auction draft on Livestream with seven total people (including myself) as we bid on which leaders we wanted to fill out our fantasy AI Survivor teams. Everyone was given 200 gold to spend however they wished drafting as many or as few AI leaders as they wanted. There were no "positions" to worry about or a limit to the number of AI leaders that could fit on a team; whether to have a small handful of favorites or a large group of underdogs was a choice that each participant had to make. Scoring was based on our standard AI Survivor points system: 5 points for a first place finish, 2 points for a second place finish, and 1 point for each elimination of another leader. The further that the leaders on each team made it, the more chances they had to score points. Unlike the normal AI Survivor points system, players also scored 2 points for having a "First to Die" leader on their team to make the bidding more interesting for weaker leaders.

The fantasy draft took place on Friday, 30 April 2021 between Sullla (myself), walloper, antisocialmunky, Kjotleik, Eauxps I Forgott, ljubljana, and Xinitiao. Here's a link to the Livestream recording for anyone curious about how this played out in real time. None of us really knew what we were doing and it was fascinating to watch the auction economy develop over the course of two hours of bidding. Eventually it became clear that the most popular leaders demanded a price in the 60-70 gold range, about a third of the 200 gold total alloted to each fantasy team. Initially some of the least popular leaders were snapped up at single digit gold totals, such as Xinitiao getting Sitting Bull for 5 gold, but over time it emerged that the going rate for these leaders was roughly 10-15 gold. Several of the fantasy participants wound up with unspent gold at the end of the auction which they could have used earlier, but of course there was also the real fear of everyone spending all their income and leaving someone to snap up a bunch of leaders at 1 or 2 gold apiece. In the end, most everyone wound up with seven or eight AI leaders filling out their team which was just what I was expecting to see. At the end of the season, we tallied up the results and, uh, I was the winner:

I invested heavily in Huayna Capac and Willem which payed off in the former case and backfired terribly in the latter case. After a rocky opening round game, Huayna Capac won the Wildcard game and his playoff match to deliver 13 total points, the second-most of the whole competition. I also landed massive scores from the two German leaders, Frederick and Bismarck, who both won their opening round games and then conveniently exited as First to Die in their playoff matches to lock down two additional points. This narrowly boosted me past Xinitiao, who had gone with a broad strategy of trying to pick up as many leaders as possible, and antisocialmunky, who had done the exact opposite in picking a smaller team of favorites and scored the coup of getting Season Six Champion Mansa Musa onto his squad. The other fantasy teams didn't have as much good luck as the trio that finished at the top of the scoring, with pricey leaders like Justinian (Eauxps), Suryavarman (Kjotleik), and Gandhi (walloper) delivering disappointing results. We know that there's a major degree of randomness to AI Survivor and sometimes an excellent pick doesn't wind up hitting due to dumb luck.

After the season concluded and I started running the alternate histories, I wanted to go back and revist the fantasy contest to take a fresh look at which draft picks turned out to be the best choices in light of running those replays of each map. For example, Mansa Musa was a huge winner for antisocialmunky while his other expensive bid on Julius Caesar turned into a dud. Were these the most likely outcomes for those picks or lucky/unlucky breaks of fortune? We're therefore going to take a look at each of the opening round games in the specific context of the fantasy draft using the additional data from the alternate histories. Note that this analysis can *ONLY* be done on the opening round games because naturally the playoff matches were dependent on what happened in the first eight games. If a different leader had been victorious in Game One then there would have been different leaders in Playoff Game One, and so on. Similarly, this analysis will naturally undervalue heavy bidding on big favorites because it only counts the scoring from the opening round games. Dropping a heavy investment into Huayna Capac or Justinian is all about getting value from more than a single game and it can absolutely be worthwhile to overbid with the expectation of paying for future playoff performances. So with those caveats in mind, let's take a look at the opening round games starting with Game One:

Game One

The first game of the season wound up with a highly atypical result. Pericles was the giant runaway favorite on this map, winning 15 times in 20 games, and Pericles bizarrely didn't even claim a Top Two finish in the actual Game One. This was a terrible break of fortune for walloper who had made an amazing pickup of Pericles for only 34 gold and should have expected to tally 5 points for the fantasy contest plus a chance to score more points in the playoffs. Instead, Pericles went to the Wildcard game and gained only 1 point from a random kill before finishing in fourth place - ouch! None of the other leaders scored particularly well in the alternate histories and it was largely random who happened to finish in second place (or first place when Pericles didn't win). The alternate histories confirmed that Caesar and Gandhi both had not been worth large bids on this map, with Caesar unable to go on one of his killing rampages and Gandhi largely confined to First to Die status. That was worth a modest bid but not the 46 gold that walloper paid. The two lucky dogs from Game One were myself and Xinitiao since we both took points from Frederick and Sitting Bull who were two of the weaker scoring leaders on this map. Frederick in particular scored me 6 points for a win and a kill despite amassing a total of 20 points across *ALL* of the alternate histories. That was absurdly good fortune and certainly much better than I deserved to do on this map. Better lucky than good and all that.

Game Two

Game Two was especially important for my fantasy team since I had three leaders from this match including the heavy investment into Huayna Capac. The alternate histories revealed that the Incan leader had not been a good choice for this particular starting position, with only four wins in twenty games and a surprisingly high elimination rate of 40 percent. I was fortunate to see Huayna Capac survive long enough to reach the Wildcard game where he subsequently went on a big scoring spree. However, my fantasy team benefited from landing the actual top two scoring leaders on this map in the French duo of Napoleon and Louis. Napoleon was in fact the top scoring leader in the alternate histories from ALL of the opening round games with an insane 107 points and Louis piled on another 53 points due to a slew of Runner Up finishes. At a bid of only 9 gold Napoleon proved to be the best value of any leader in the whole fantasy draft, at least as far as the opening round was concerned. (This was somewhat good luck as I spotted Napoleon going for a cheap price early in the fantasy auction and snapped him up; Louis was not a case of good luck as he was on my list of leaders to target in the bidding.) So while I was fortunate to have Huayna Capac survive into the Wildcard round, I was actually unlucky to have Napoleon finish in second place and Louis wind up in the Wildcard game. This was much worse than their typical first/second finish and I easily could have scored 10 points from the two of them in this game alone. As much as the RNG had favored me in Game One it kind of went against me here.

Amongst the remaining leaders, ljubljana made a sizable investment into Pacal and it paid off when Pacal won the actual Game Two and then advanced all the way to the championship. However, it turned out that this was an extreme outlier result: Pacal scored ZERO first or second place finishes and zero kills in the alternate histories! His points came entirely from six First to Die results as Pacal was repeatedly killed by Napoleon and Louis. This was one of the wildest results of the whole season and it had massive effects on the rest of the fantasy contest. Pacal wound up carrying ljubljana's fantasy team when his most likely outcome would have been an immediate elimination. No one else was heavily invested in the leaders from this game although Kjotleik made an excellent value bid on Saladin at only 15 gold which paid off when Saladin wound up being First to Die in the actual Game Two. Eauxps and walloper took similar fliers on Suleiman and Hatshepsut that had about the same odds without hitting.

Game Three

Game Three was the story of three strong leaders in Gilgamesh, Ramesses, and Augustus. This trio of leaders hogged the scoring in the alternate histories and left table scraps for the other three AI nations. The alternate histories revealed that Xinitiao's large bid of 52 gold was completely justified for Gilgamesh as he had 8 wins on this map and an impressive 21 kills leading to many alternate histories where he ran away with the scoring. Unfortunately for Xinitiao the actual Game Three saw Gilgamesh getting sent to the Wildcard game with only a single point for a kill, though this was offset somewhat by Shaka taking a very unlikely second place finish to claim some points for Xinitiao's team. The best value bets in this game were Ramesses for 16 gold by walloper and Augustus Caesar for 17 gold by antisocialmunky. Both of these were strong leaders with high scores in the alternate histories, with Ramesses winning 8 times and Augustus having a ton of second place finishes. The actual Game Three wound up being kind to antisocial and cruel to walloper, as Augustus took home the first place finish while Ramesses was eliminated without even having the courtesy to be First to Die. This was the second time that walloper made an excellent pick in the fantasy draft (Pericles and Ramesses) only to have the RNG of the matches run against his team.

Game Four

Game Four was where the favorites in the alternate histories started to get back on track after a bunch of wacky results in the first few games. Catherine was the winner of the actual Game Four and the alternate histories confirmed that this was not a fluke as she won another nine times on the same map. Xinitiao made a modest bid of 26 gold for Cathy and it paid off enormously as she would score 11 total points by making a run to the championship. This was the best pick of Xinitiao's whole fantasy team and with slightly better luck elsewhere it would have won the fantasy contest. Eauxps invested a fairly large sum of 40 gold into Stalin and the alternate histories proved that this was a justified bid as Stalin was the second-highest scoring leader in the scenario. Sadly for Eauxps he watched Stalin suffer the First to Die fate which proved to be highly atypical in the alternate histories, only happening once in twenty games. Ljubljana had the other large bid with 40 gold for Charlemagne and this proved to be an unwise choice both in the actual Game Four as well as in the alternate histories. Charlemagne was hemmed in by his neighbors and couldn't stop fighting with Elizabeth, scoring no points in the fantasy contest and doing little in the alternate histories. Finally, antisocial and Xinitiao took some longshot odds on Brennus, Alexander, and Hammurabi which were very much worthwhile at the miniscule prices those leaders commanded. Brennus and Hammurabi were misses but Alexander wound up being a huge hit as he landed a second place finish and ultimately 4 kills in total between Game Four and his playoff game. Xinitiao was fortunate to get that kind of outcome from Alexander which was well above expected value; then again, Xinitiao also had the most total AI leaders on his fantasy team so he was very likely to have a few of his longshot leaders pay off somewhere.

Game Five

Game Five was another game where I was heavily invested thanks to spending so much gold on Willem. When scouting the opening round games, I liked Willem's position the most out of any leader and thought that he was a huge favorite to win his initial game and then continue onwards into the playoffs. This proved to be a disastrous misread as Willem only won a single game in the alternate histories and picked up most of his points from being First to Die six times - urp! Willem scored zero points in the actual Game Five and I essentially threw away 60 gold on one of the weakest leaders in the whole competition. My fantasy team was saved due to the fact that I picked up Bismarck with my final 13 gold in the fantasy auction; I distinctly remember that we were getting down to the last seven or eight AI leaders in the auction and I thought it was worthwhile to spend all of my remaining income at once on Bismarck so that I would get at least one more leader on my team. I did like Bismarck's starting position and thought he was worth the minimal investment. This was another major misread but one that worked in my favor: Bismarck was by far the strongest AI leader on this map with 12 victories, 29 kills, and 99 total points. It was the second-highest score in the alternate histories of the opening round behind Napoleon... who was also on my fantasy team, heh. It was wild how I largely missed on Huayna Capac and Willem only to find myself with these unexpected juggernauts in Napoleon and Bismarck. The German leader won the actual Game Five while scoring two kills and then amazingly managed to be First to Die at a very late date in the playoffs to reward me with 9 total points. Obviously this was lucky but it wasn't totally unearned since I did make a smart move in the fantasy auction to pick up Bismarck, without which my fantasy team would have been dead in the water.

Amongst the other leaders, Kjotleik spent heavily on Suryavarman who proved to be a solid choice if perhaps not worth the full 50 gold investment. The alternate histories confirmed that Suryavarman's second place finish was his most likely outcome from the game but unfortunately that was all he would achieve as he failed to score any points in the playoffs. Definitely not a bad pick but a little bit underwhelming. No one else spent much on this game which made the minimal bids on Washington (12 gold), Ragnar (15 gold), and Lincoln (7 gold) good value picks. The alternate histories showed that Washington scored almost as many points as Suryavarman and ljubljana picked him up at a fraction of the price. Washington scored two points for being First to Die and Ragnar grabbed a random kill to reward those wise purchases.

Game Six

Game Six was kind of a crapshoot and as a result it tended to reward the fantasy teams that didn't make a big investment into any of these leaders. Eauxps had by far the best bid by landing Churchill, the winner of Game Six who was confirmed as the top leader on the map by the alternate histories. Churchill inexplicably made it all the way to the championship where he actually had a chance to win the whole tournament, scoring 11 points in total for Eauxps across three games. Although Churchill wasn't as dominant in this game as leaders like Pericles and Napoleon and Bismarck in their respective games, he was still by far the best pick and a total steal at 17 gold cost. The other leaders were kind of a muddled bunch which the alternate histories showed could randomly be either strong or weak depending on the vagaries of each game. Cyrus at 40 gold and Kublai Khan at 47 gold don't appear to have been worth their expense given that they had no better chance to win than Boudica or Tokugawa who commanded significantly lower sums. Kjotleik was fortunate to have Cyrus finish in second place and then make his own run to the championship game since the alternate histories showed that Cyrus only had 4 Top Two finishes across 20 map replays. Kublai Khan was unable to score any points in Game Six which was the biggest miss for Xinitiao's fantasy team. This was a bit unlucky but not terribly so as Kublai wasn't particularly impressive in the alternate histories.

Game Seven

I was so interested in the outcome of the alternate histories for Game Seven that I ran them while the AI Survivor season was still taking place. They confirmed what we had seen on the Livestream: Game Seven was the Mansa Musa show as he won 11 more times and completely justified antisocialmunky's major spending on the Malinese leader. Mansa would end up winning the overall competition for Season Six and proved to be the top scoring leader for fantasy purposes with 17 points. In retrospect, he was probably undervalued at 65 gold given Mansa's strength on this map plus playoff scoring potential. Antisocial also picked up Mao who performed basically exactly as expected with a second place finish and a kill for 3 fantasy points. Mao was likely worth a bid of middle value and that's exactly where he ended up at 34 gold - the free market working efficiently in action! The alternate histories showed that Isabella was the second-strongest leader on this map and Kjotleik made an excellent bid to pick her up for 17 gold. This failed to work out in the real Game Seven since Isabella died without scoring any points but it was a smart investment that simply didn't pan out. Peter also proved to be a solid pickup for 22 gold as it turned out he was very likely to be First to Die (which is exactly what happened in Game Seven). By contrast, Roosevelt was so weak at scoring on this map that he wasn't even worth the 13 gold that I bid and predictably turned into zero points. He couldn't even manage to be First to Die more than twice in the alternate histories - bah!

Game Eight

Game Eight delivered another "what if" scenario of a path not traveled in the game we watched on Livestream. The repeated playthroughs of the map revealed that pregame favorite Justinian had indeed been the top scoring leader for this scenario, winning 9 times and picking up 24 kills along the way. Eauxps made the highest bid of the whole fantasy auction at 77 gold and saw exactly zero fantasy points for his investment, a complete tragedy given how powerful Justinian had been in the alternate histories. Along with Churchill's surprisingly strong showing, this should have made Eauxps' fantasy team one of the contenders to win the contest but instead it finished in the middle of the pack. Eauxps and walloper had the two fantasy teams that suffered the most from bad RNG luck working against some very good picks.

The other strong leader on this map was Darius who ljubljana was able to acquire in another tremendous bargain for only 20 gold. Ljubljana was a bit unfortunate here even though Darius finished in second place during Game Eight since the Persian leader had seven first place finishes in the alternate histories. He easily could have scored double those fantasy points or more. At the other end of the spectrum was walloper's pickup of Victoria, which was good value at only 14 gold cost and then became incredible value when Victoria went on to win Game Eight. The alternate histories showed that Victoria was an unlikely winner on this map and walloper's good luck here partially offset his bad luck earlier in the season with Pericles and Ramesses. None of the other leaders attracted too much in the way of bidding aside from Hannibal going for 30 gold from Kjotleik. Hannibal was the runaway favorite for First to Die on this map (which he failed to achieve in an unfortunately blow for Kjotleik) but that probably wasn't worth the investment of 30 gold. Leaders who were likely to be First to Die seemed to be worth roughly 10-15 gold investment since even hitting on the feat only delivered 2 fantasy points with no potential for further scoring. I'm sure that everyone will have a better sense of where and how to bid if we run this exercise again for the next season.

Here's the combined fantasy summary from the alternate histories of all eight opening round games:

I included the cost to purchase, the actual points scored across all of Season Six, and then the alternate history points scored to get a better sense of the true value of each leader in the opening round games. Again, this analysis necessarily does not include any results from the playoff rounds and therefore undervalues teams of AI favorites like the one put together by antisocialmunky so please do keep that in mind. Nevertheless, I still think that the exercise is interesting as it helps showcase which fantasy teams experienced good luck and bad luck throughout the opening round of games. I ended up with the highest score in the actual fantasy contest and then also had the highest score in this look at the alternate histories. Once again, my misses with Huayna Capac and Willem were counterbalanced by having such massive hits with Napoleon and Bismarck and Louis. It was a true case of feast or famine as I had both the most cost-efficient and least cost-efficient leaders on my fantasy team. Willem was the second-worst purchase of the whole auction at 60 gold for only 19 points in the alternate histories (0.32 points per gold) while Napoleon was the best value at 9 gold for 107 points (11.89 points per gold) and then Bismarck was the second-best value in the fantasy draft at 12 gold for 99 points (8.25 points per gold). Leaving aside the limitations of this analysis, I was pleased to find that my fantasy team's performance could not be chalked up solely to dumb luck in a tiny sample size of games. I drafted three very strong leaders who carried my team to the top score in the compiled alternate histories of 20 games apiece on each map.

Xinitiao's team came close to outscoring mine in the real fantasy season and that was the case in looking at the alternate histories as well. Catherine was the core of the team once again in the alternate histories and unlike in the actual Season Six she was joined by Gilgamesh who had a very strong showing in the alternate versions of Game Three. The rest of Xinitiao's leaders largely balanced out one another in comparison to the matches that we watched: Alexander and Shaka and Zara all outscored their average alternate history performances but Kublai Khan and Lincoln were misses that easily could have delivered more points instead of nothing. Xinitiao's fantasy team likely ended up right where it was supposed to be in terms of scoring. By contrast, antisocial's team was the one most shortchanged by this look at the opening round since it couldn't predict how Mansa and Mao and Augustus (all of whom had excellent odds to advance) would perform in the playoffs. Based solely on the opening round, antisocial's pick of Julius Caesar was a mistake but the other four leaders were all excellent choices, especially Mansa Musa. Brennus for 13 gold and Augustus for 17 gold were both strong value picks as they each had plenty of success in the alternate histories. Augustus wound up hitting while Brennus missed in the actual Season Six but they both had very good efficiency for their bidding cost (each was about 3 alternate history points per gold spent, very cheap).

The alternate histories conclusively showed that two teams had been victims of poor RNG in the opening round matches: Eauxps and walloper. Eauxps had the terrible bad luck to see both of his big purchase fail to live up to expectations, with Justinian and Stalin combining for a mere 2 points thanks to Stalin's First to Die status. However, these were not bad picks at all as the alternate histories had the two of them combining for 129 points which was a solid if not spectacular conversion rate (117 gold for 129 points = 1.10 points per gold). When combined together with Churchill's stellar season and amazing value (17 gold for 67 points = 3.94 points per gold, third-best behind Napoleon and Bismarck), Eauxps would have been right there challenging for the fantasy lead if he'd had better luck with Justinian and Stalin. Walloper was a similar victim of bad luck as Pericles and Ramesses scored 143 points in the alternate histories compared to a grand total of 1 point in the real Season Six. That was terribly unfortunate and even having Victoria claim an unexpected win in Game Eight wasn't enough to turn things around. Walloper also had a fairly large bid on Gandhi which didn't work out in practice as the Indian leader failed to be worth his cost. This fantasy team clearly should have scored significantly more than 11 fantasy points though; I imagine that this was the real world equivalent of a team derailed by injuries.

Unfortunately some of the teams inevitably had to come up short and those proved to be ljubljana's and Kjotleik's teams which coincidentally finished at an identical 236 alternate history fantasy points. The core of ljubljana's team in the real fantasy competition was Pacal but the alternate histories demonstrated that this had been supremely lucky. At a cost of 40 gold for only 12 alternate history points, Pacal was the worst theoretical value of the whole auction at a monstrously expensive 0.30 alternate history points per gold. It's completely crazy that what was mathematically the worst pick of the draft wound up scoring 11 fantasy points for ljubljana - Lady Luck can work in strange ways. Darius was a much better pick at only 20 gold for 60 alternate history points that wound up underdelivering in the games on Livestream. The other picks on ljubljana's team simply didn't work out, both in the real games and also in the alternate histories, suggesting that these leaders innately weren't that strong. Kjotleik's best picks mathemtically were Isabella and Saladin, neither of which performed as well in their respective opening round games as they did in the alternate histories. But Kjotleik wasn't particularly unlucky in terms of his fantasy team as no one proved to be especially dominant in their opening round games. Suryavarman was a disappointment for his cost and the rest of the team didn't have any true heavyweights. Given that Cyrus outperformed his alternate histories result, Kjotleik's team is another one that looks to have ended up about where it should have been.


So what conclusions can we draw from this initial season of fantasy AI Survivor? Obviously getting lucky in the actual games had a massive effect on the standings and randomness will always play a huge role in the outcome. But it's also not true that everything comes down to pure RNG either as there were plenty of bargains out there in the auction for those who could claim them. Bismarck, Napoleon, Churchill, and Catherine were all cheaply available and delivered enormous value for their respective teams. In some cases it was worthwhile to spend heavily on favorites due to their later playoff scoring potential, as proven by Mansa Musa and Huayna Capac. Or it could also be worthwhile to spread out the bidding over lots of less popular leaders in the hopes of hitting on some of the longshot situations; Xinitiao did this better than anyone by scoring a combined 12 points from Sitting Bull, Hammurabi, Shaka, and Alexander. At the same time, the alternate histories showed that this wasn't necessarily an easy recipe for success either. Xinitiao ended up with one of the highest scores by drafting a broad pool of nine leaders... but ljubljana also had one of the lowest scores by drafting the same large group of nine leaders. More leaders did not automatically translate into success.

The secret to success was of course the ability to maximize the value of each leader, drafting the most cost-efficient team by acquiring cheap talent. This is always the case in fantasy drafts as well as in putting together rosters in real athletic competitions. My team had the best season because I managed to pick up Napoleon/Louis/Bismarck through a mixture of smart drafting and total dumb luck and then combine them together with performances from Huayna Capac and especially Frederick that the alternate histories said were well above the norm. Drafting Bismarck (the Tom Brady of Season Six AI Survivor) yielded so much value that it was sufficient to offset my total analysis failure in paying out the nose for Willem (the Ryan Leaf or JaMarcus Russell of this year's draft). Nothing carried these AI fantasy teams more than having an inexpensive Churchill on the roster just as nothing sunk it worse than emptying the wallet for a relative dud like Julius Caesar or Gandhi. Much of this whole fantasy competition came down to AI randomness but it sure was a lot of fun watching everything play out!

I'm hoping that we can do this again with eight total participants the next time that we run a season of AI Survivor. I'd also be happy to act as a moderator for eight individuals to allow for more impartiality if there happens to be enough interest on the part of the community. Until then, I hope this was a fun look back at our first fantasy competition. Thanks as always for reading along!