AI League 2.0

[Keler]

I think in order to have a more balanced pangea map, you need to generate map by adding one more civ in it, then remove a middle starting position where civs start closer to each other and nerf that removed capital area, removing bonuses adding 2-3 mountains in the exact removed capital spot. My switchhero Wildcard 6 map turned out somewhat balanced and I did it that way. There was a middle capital area and I removed it. Imagine removing Justinians season 8 position between Catherine and Tokugawa to make the map fairer for a 6 player map.

 

[Thrasybulos]

But did you really want a 7 player championship game?

Actually... yes.

It wouldn't be subtitled "King of the Hill Edition" for nothing.
The Championship map would feature the defending champion in the middle, with a stronger start.
So basically, he'd need to get 2v1'd to get defeated... which being in the middle would favour. And in addition, as an extra incentive, he'd get a special -1 relations with the other AIs. And should he successfully defend his title, that would be raised to -2 the next time. And so on...

I also want to feature more 7-player games, as I find those tend to be more likely to be interesting, with 2-3 strong leaders emerging ; 6-player games seem to be ever so slightly more prone to turning into one-sided affairs, where the ultimate winner is known by turn 150.

And the last advantage of that format (imo) is that it's a bit more "fair", by not mixing 6-player and 7-player games in any given round: statistically, an AI has more chances of winning a 6-player game than a 7-player game.
Now, of course, the luck of the draw (assigned start, peaceweight situation) plays a much heavier role... but still. I'd rather the tournament format not have any built-in unfairness, so to speak.

 

[Thrasybulos]

Series 13 Results

Spoiler Results :

So... we get a rather unexpected winner this time!
Not that Saladin is a dud either. But if the win doesn't go to HC, nor to Willem, nor to a top high peaceweight... you'd expect Justinian or Pacal, rather.

The thing is... this series saw all the top AI underperform. HC's final number 3 spot is flattering: he got there by winning 3 games in the very last round. But before that, he had his worst series by far, playing only two rounds in Pool 1.
But that was even worse for Justinian and Louis, who never reached Pool 1! Willem fared better, but Pool 2 was his main haunt, not Pool 1. Suleiman had a good start, but it went downhill from there. Only Pacal did fairly well... but not well enough.
And it was the same for the top high peaceweight leaders: Hatty had a decent start, but then floundered... which Mansa couldn't take advantage of as he performed even worse for most of the series (funnily, both had an excellent round 9, though: 4 wins for Mansa, 3 wins for Hatty). Darius did a bit better (even though he finishes lower, you can see he remained in the top half from round 3 onwards, contrary to Hatty and Mansa). Zara did poorly.

The only exception was Ramesses, who did his usual thing: start very strong, end up in Pool 1 in a hostile diplomatic environment, still win enough games to stay there, but not enough to have a shot at the final win.

So it was up to the second-tier leaders to shine. Hannibal had the most consistent performance, and up to round 8 looked as if he'd be the winner. Then Gilgamesh overtook him in round 9... only to concede to Saladin's 3 wins in round 10. There were also some semi-surprise guests at the top: Cathy, Qin. And an actual surprise: Mehmed.

So what happened?
It's hard to tell, but if you look at the map data section, you'll notice that quite a few of the maps used for this series were extremely unbalanced.
So it might explain why there were more games which felt like "outliers": if a leader missed out on the map's "guaranteed" win, for whatever reason, there'd be far fewer opportunities than usual to make up for it.

Spoiler Awards :

The Rorschach award is a pretty good illustration / explanation of De Gaulle's plight in this series: he may have seriously underperformed... but that's in great part because he kept drawing awful pool compositions.

Spoiler Elo Ratings :

Freddy on the rise! Beware, HC, he's coming at you!
The biggest gains this time are Mehmed who had an awesome run, Qin, Cathy, and Brennus.
Biggest losses are De Gaulle, Genghis, Churchill, and Ragnar.

But all in all, we're hitting the stage where the variations for a single series get pretty minor, and actual changes require a multi-series trend.
Which could be the case for instance in the Shaka vs Napoleon match-up: Napoleon was far ahead a few series back, they're now essentially tied.
Conversely, is Darius actually pulling ahead of Zara?

Spoiler Map Data :

Spoiler Pool 1 (S6 Game 7) :

Spoiler Pool 2 (S6 Game 1) :

Spoiler Pool 3 (S1 Playoff 1) :

Spoiler Pool 4 (S4 Game 1) :

Spoiler Pool 5 (S2 Game 6) :

Spoiler Previous Results :

Spoiler Pool 6 (S4 Game 2) :

Spoiler Previous Results :

Spoiler Pool 7 (S5 Game 4) :

Spoiler Previous Results :

Spoiler Pool 8 (S4 Game 6) :

Spoiler Previous Results :

 

[AcaMetis]

Awards

Love the fierce competition for the Kenny award there .

 

[Thrasybulos]

You know that thing about the AI getting into war mode after launching its spaceship?

That's a myth.
It never happens.

 

[Thrasybulos]

Series 14 Results

Spoiler Results :

Huyna Capac wins... again. He had a bad first round, but rallied quickly. And while he didn't crush the competition, the final outcome seemed settled by the mid-series.

Again, Ramesses performed strongly, but from round 6 onwards, he remained the sole high peaceweight in pool 1, so the deck was stacked. He had an amazing round 7 where he won 3 games, but that was his swan song: he only won one game in the remaining three rounds.
Gilgamesh, after a slow start, finished very strong, and even surprisingly threatened HC for the top spot at the very end.

De Gaulle had a diametrically opposed performance to his previous series and was the best French leader this time.
Another leader whose performance record could be described as "erratic" is Catherine: she was a top leader in the previous series, she remained in the bottom pools this time.
Kublai finishes in 16th position, which might be considered a decent result, but he spent most of his time in the bottom pools as well. The fluctuations in his results aren't nearly as wild as Cathy's, but his good runs seem to have inflated my expectations for him when where he ended up here would be actually more typical.

The worst performance of the series goes to Alexander, by quite the margin. I can remember a Sullla stream from a few months ago where a viewer expressed the opinion that Alex was a better leader than Sury... well, no. Definitely not. Sury may be grossly overestimated by the AI Survivor's community, but he's still a middle-of-the-pack AI while Alex is a bottom-feeder.

Spoiler Awards :

No new records this time, but a few eyebrow-raising results nonetheless: Gandhi not featuring in the Kenny awards, Hatty (!) getting the second-fastest kill, Mansa getting the slowest win, and I believe Augustus is the first high-peaceweight AI to feature in the Jason awards...

Spoiler Elo Ratings :

No big swings here, we're hitting dimishing returns hard.
Still, Cathy, Kublai, and Sury pay the price for their lacklustre performance, while Alex slips back below the 1,300 elo bar.
Bismarck and Stalin, in spite of a good result, remain firmly rooted in the bottom 10. De Gaulle makes up the ground lost in the previous series.
Shaka's done it: he's overtaken Napoleon! Guess I did speak too soon when I said Napoleon was clearly the best of the crazies...
Darius seems to have now opened quite the lead over Zara, while Mansa appears unable to catch up to Hatty.
And Ramesses is now within spitting distance of Justinian, which is something I wouldn't have expected!

Spoiler Map Data :

Spoiler Pool 1 (S7 Playoff 1) :

Spoiler Pool 2 (S5 Game 1) :

Spoiler Pool 3 (S2 Game 1) :

This one came as a complete surprise, as the results from @Keler and AH sets pointed at position 3 being absolutely OP... which was not the case here.

Spoiler Pool 4 (S7 Playoff 2) :

Spoiler Pool 5 (S4 Game 8) :

Spoiler Previous Results :

The results for positions 1 & 4 are swapped in the two sets. The two are certainly linked as success for whichever AI started in 4 almost invariably involved conquering the AI in 1...

Spoiler Pool 6 (S3 Game 6) :

Spoiler Previous Results :

Spoiler Pool 7 (S5 Wildcard) :

Spoiler Previous Results :

Spoiler Pool 8 (S6 Game 2) :

Spoiler Previous Results :

 

[Thrasybulos]

Spoiler :

That victory message with no name preceding "Wins" happens when the victorious AI has been wiped out.

It's a common occurence with AIs which go for an early Culture victory (as a completely random example... Gandhi? ).
Far less common, but still happening on a regular basis, is the case of an AI winning the space race, only to get removed from the map thereafter.

But for an AI to win by Domination, and then get wiped out... I must have played by now upwards of 10,000 of those AI games, and I can only remember three other occurences!

(Here it was Cyrus who reached the Domination threshold, but his subsequent DoW on a much, much more advanced Darius didn't go too well, apparently...)