AI Survivor - Alternate Histories AI Ranking

[Thrasybulos]

1. Season 1, part one
   Season 1, part two
2. Season 2, part one (TBD)
   Season 2, part two (TBD)
3. Season 3, part one (TBD)
   Season 3, part two (TBD)
4. Season 4, part one
   Season 4, part two
5. Season 5, part one
   Season 5, part two
6. Season 6, part one
   Season 6, part two
7. Season 7, part one
   Season 7, part two
8. Season 8, part one
   Season 8, part two

---

In parallel to my AI League endeavour, I'm going to be running a set of "alternate AH" for Sullla's AI Survivor games, with the objective of ranking the AIs within the framework of AI Survivor.

Spoiler Origin of the project :

Saxo Grammaticus mentionned in Eaupx's Season 1 AH thread he was using the AH results to rank the AIs.
That spurred my interest as while that seemed a good idea (far better sample set than the live games), I could see a few issues with that:

• In many cases, the AIs which make it to the playoffs and the championship are not the AIs which the AH indicate, so using the playoffs & championship AH is at odds with the objective.
• Some game setups resulted in very unbalanced situations: using 20 iterations of such setups for a ranking system is bound to have a distortion effect.

While waiting for Saxo's response about how he handled those issues, I thought about them myself and came up with this project.

The other reason for doing this is that September/October tend to be busy months for me, so I don't have as much time available for running games manually as I've been doing for the League. The temptation was strong to start autoplaying them to keep the pace going, but that would mean losing information (not watching the games, some data becoming unavailable), so I was reluctant to proceed.
Now the League will continue at its own pace, while I can take advantage of autoplay for this project instead.

Here's how it'll work:

• I'll be running a new set of AH for each season.
• For the wildcard, playoffs, and championship games, the AIs will be selected according to the AH results, not according to the live game result.
• Each AH game result (that is, the result of the 20 iterations) will be translated on a 5/2/1* scale to attribute a Power Rating score to the AIs.

Spoiler Why run new AH ? :

I could have used the existing AH results to save time, and just run new AH when they don't exist, and for the new wilcard/playoffs/championship game compositions.
Running new sets of AH has two advantages:

• It ensures consistency. All the games will be run under the same ruleset and using the same procedure, which is not the case for the existing AH.
• It will provide a new result dataset. 40 is better than 20.

All these games will be run with the console command game.aiplay without me watching (if I were available to watch them, I'd be running League games instead).
The games will be run under my usual ruleset/procedure: S5 rules (no Deity bonus starting techs, no AP), every game run as a scenario (randomized peaceweights every time), no UN, no great spy inflitrations in the AI capitals by the observer civ.
This is going to be a pure data collection exercise: I'm not going to provide write-ups about the games.
Also, I'm not going to collect all the data which is usually collected for the AH: I'm not interested in the number of wars, so I'm not going to bother with that. And I'm only going to keep track of total kills, not the kills per game detail.

It also means that two elements of Sullla's scoring system are going to need revising:

Kill Credit
The game summary tells which civ has captured a city, not which leader. So in games where two leaders from the same civ are present, if both are at war with a leader who gets eliminated, it can be hard or even impossible to tell who got the killing blow.
Should that happen, I'll apply the following rule: the kill credit goes to the first leader who declared war on the eliminated leader (while still at war, obviously).
So, for instance, if Shaka declares on Louis, then Napoleon attacks Shaka, and Shaka gets eliminated, the kill credit will go to Napoleon.
But if Louis had declared on Shaka, with Napoleon piling in, then Louis would have got the kill credit.
(edit: Acametis has suggested a better way around it, allowing to attribute the kill credit to the actual killer).

Runner-up points
Since I won't be watching the screen at the time the win notification appears, I won't know the scores of the surviving leaders to determine who comes second.
In some cases it can easily be told from the width of the score graphs (and even more easily if there's a sole survivor, duh), but in other cases it's just impossible to tell.
So rather than use an unreliable statistics, I'm going to change it.

Spoiler Why not using autoplay instead? :

When discussing something similar in the League topic, Keler has suggested using the autoplay mod instead: since it stops the game when an AI wins, it makes getting the endgame scoreboard possible.
But after checking it out, I'm not comfortable with the fact that the observer civ isn't passive. In particular, the fact it can end up at war through Defensive Pacts has the potential to really alter some game outcomes.
So for this project, I'll keep using the console command.

Instead of the runner-up scoring 2 points, every survivor is going to score 1 point.
I had initially decided to give 2 points to a surviving AI when it was the only survivor besides the winner, but then I noticed something: 1 point for surviving and 1 point for kills makes the total points for each game constant. Killing someone is just getting their survivor point.
There's an elegance to that system which I couldn't resist, so even if an AI is the clear runner-up because everyone else is dead, it still gets only one point.

OK, so this is going to yield each AI's "score" for a given game, over 20 iterations: 5 points for winning a game, 1 point for surviving a game, 1 point/kill.

Next, this is going to be translated into the "AH Power Rating".
The winner of an AH is going to be the AI who won the most game iterations (in case of a draw, the score will be used as a tie-breaker).
The runner-up is going to be the AI with the highest score (winner excluded).
Now, kill points were a bit tricky. I haven't run exact calculations, but I would say the average for AI survivor should be around 3 kill points per game. I couldn't see a way to translate that precisely over a set of AH games, so here's what I have settled for:
The AI with the most kills gets one point for the Golden Spear award. Now, that can't be shared: it two AIs are tied for the most kill, then no one gets a point.
And then: 1 kill point is awarded for every 20 kills (rounded down). After all, it makes sense: in AI survivor, you get one point for one kill, here we're running 20 games.
Looking at the existing AH data, it should mean a lot fewer kill points attributed than in the live games (few AH sets feature AIs with 20+ kills). But since kills feature in the score, and the score determines advancement (runner-up spot, wildcard), kills are somewhat more important than in AI Survivor. So fewer points, but increased impact: it sorta balances out. Sorta.
Edit: See this post for the new system.

The number of kill points attributed for a game will depend on the total kills: total kills / 20, rounded to the nearest.
31-49 kills: 2 points
50-69 kills: 3 points
70-89 kills: 4 points

The attribution algorithmn will be as follows:
1- Sort by kills.
2- Give 1 point to top leader, decrease his total by 5.
3- If still points to attribute, go to 1.

Wilcard game
Each season will feature a number of wildcard spots, depending on the number of AIs who made it to the wildcard game in the live event.
So here, all AIs who didn't make it through a win or a runner-up spot in the opening rounds will be sorted by Score: the top ones will get into the wildcard game.

Alternate scoring
On top of the "AH Power Rating" which is an attempt at translating Sullla's system for the AH, I'll also be using for comparison a very simple system of "Championship points":
Elimination in the opening round: 0 pt.
Elimination in the wildcard game: 1 pt.
Elimination in the playoffs: 2 pts.
"Elimination" in the Championship game: 4 pts.
Champion: 8 pts.

Question for you guys

I'm pretty comfortable with what I've described, but I'd like your input about one thing: who should be the champion?
Should it be the AI with the most wins over the 20 Championship game iterations? Or the AI with the highest score?

If, for instance, we had Gandhi and Gilgamesh in the Championship.
Gandhi wins 6 games, gets eliminated 14 times (survival: 30%), and scores 5 kills.
Gilgamesh wins 5 games, gets eliminated only 3 times (survival: 85%), scores 18 kills.

Which of the two should be considered the winner of the Championship?

Most wins:
Eauxps

Highest Score:
Acametis, Keler (but without kills)

 

[Thrasybulos]

In order to illustrate this, here are the results for the first three games (S1 games 1-3).
Eauxps has already posted his AH for those games. I'll wait until he has posted the rest to post mine (don't have them yet anyway ).

I believe he intends to run AH for Season 2 next, so I'll skip it for now (might go straight for Season 5 afterwards, in case Eauxps wants to fill in the blanks for Season 3 afterwards, and I've already run AH for season 4 + Jumbled Rumble Season 1 on those maps, so no hurry to visit them once gain).

Those results differ somewhat from Eauxps'. I might be wrong, but I'd say they fall within the expected variations from one set of 20 games to another: 2 leaders were definitely weaker on this map, the other 4 fairly comparable.

When Zara won the first two games, which had never happened in Eauxps' sample, I thought the two sets were headed in very different directions... but that didn't happen. Overall, this is pretty much what Eauxps got.

Ok, now this is very different from what Eaupx got.
Isabella dominated his games, here she was just dead meat.
The different ruleset (no Deity starting techs here, no AP), and possibly a low peaceweight roll lock in Eauxps's set might explain it (depending on when the game rolls for that)?

 

[AcaMetis]

Kill Credit
The game summary tells which civ has captured a city, not which leader. So in games where two leaders from the same civ are present, if both are at war with a leader who gets eliminated, it can be hard or even impossible to tell who got the killing blow.
Should that happen, I'll apply the following rule: the kill credit goes to the first leader who declared war on the eliminated leader (while still at war, obviously).
So, for instance, if Shaka declares on Louis, then Napoleon attacks Shaka, and Shaka gets eliminated, the kill credit will go to Napoleon.
But if Louis had declared on Shaka, with Napoleon piling in, then Louis would have got the kill credit.

These maps are rolled and later edited as a worldbuilder file, correct? If so isn't it possible to give civs custom civ descriptions, so the log would say I.E. City Name (Zulu) was captured by the Churchill Empire, as opposed to the English Empire which would mean any of Churchill/Elizabeth/Victoria could have captured it?

I'm pretty comfortable with what I've described, but I'd like your input about one thing: who should be the champion?
Should it be the AI with the most wins over the 20 Championship game iterations? Or the AI with the highest score?

If, for instance, we had Gandhi and Gilgamesh in the Championship.
Gandhi wins 6 games, gets eliminated 14 times (survival: 30%), and scores 5 kills.
Gilgamesh wins 5 games, gets eliminated only 3 times (survival: 85%), scores 18 kills.

Which of the two should be considered the winner of the Championship?

I imagine highest score. If it all comes down to who wins the most games, than why even keep track of stuff like kills/points/survival rates/etc. outside of pure curiosity? I guess for tie breaks, but I feel like it should have more impact than that.

 

[Thrasybulos]

If so isn't it possible to give civs custom civ descriptions, so the log would say I.E. City Name (Zulu) was captured by the Churchill Empire, as opposed to the English Empire which would mean any of Churchill/Elizabeth/Victoria could have captured it?

Good idea!

If it all comes down to who wins the most games, than why even keep track of stuff like kills/points/survival rates/etc. outside of pure curiosity? I guess for tie breaks, but I feel like it should have more impact than that.

It's used elsewhere: determines runner-up for each game (so highest score and most wins always both move on - although in most cases that's the same AI anyway -), wildcard game participation.

But there's no runner-up for the Championship game.
I would also say highest score, but I thought getting people's input on that one was better since highest score doesn't always mean most likely to win (looking at you, Shaka!).

 

[Eauxps I. Fourgott]

In order to illustrate this, here are the results for the first three games (S1 games 1-3).
Eauxps has already posted his AH for those games. I'll wait until he has posted the rest to post mine (don't have them yet anyway ).

I believe he intends to run AH for Season 2 next, so I'll skip it for now (might go straight for Season 5 afterwards, in case Eauxps wants to fill in the blanks for Season 3 afterwards, and I've already run AH for season 4 + Jumbled Rumble Season 1 on those maps, so no hurry to visit them once gain).

I think that TheOneAndOnlyAtesh who's been doing writeups for Season 8 has caught the AH bug after running the most recently published set, so he'll be cleaning up Season 3.

View attachment 704061
Those results differ somewhat from Eauxps'. I might be wrong, but I'd say they fall within the expected variations from one set of 20 games to another: 2 leaders were definitely weaker on this map, the other 4 fairly comparable.

A somewhat surprising set, mainly Napoleon doing so much better. Perhaps having the best starting techs for his civ narrowed the gap in this case, let him grow stronger by proxy? Sury was a bit better but then tended not to do so well when he wasn't winning, it seems. Peter being weaker (and having so many survivals) is the other big difference. Don't know about that one.

View attachment 704062
When Zara won the first two games, which had never happened in Eauxps' sample, I thought the two sets were headed in very different directions... but that didn't happen. Overall, this is pretty much what Eauxps got.

I believe I did say that I thought Zara could win if enough games were run on that map! He did have a couple of near misses so that is no shock. Otherwise agreed, this map seems to have held pretty steady. Justin notably better at surviving is interesting though.

View attachment 704064
Ok, now this is very different from what Eaupx got.
Isabella dominated his games, here she was just dead meat.
The different ruleset (no Deity starting techs here, no AP), and possibly a low peaceweight roll lock in Eauxps's set might explain it (depending on when the game rolls for that)?

Wow! Talk about a completely different dynamic. I don't think PW rolls would account for Izzy's worse performance, unless perhaps I got one that made Boudica and Louis more likely to fight; Izzy was being attacked all the time in my games as well. I'd have to guess that losing the free starting techs (all four for her, right?) slowed her down enough that she wasn't ready to meet the inevitable early attacks and got crippled, unable to pull ahead in the same way? Qin's good techs and the weakening of Izzy/Joao then account for him doing better. Joao's weakness also makes sense as his better results in my games were tied to Izzy's success; with her out of the picture, he was pretty dead.

But then what I don't understand here is why Qin/Boudica were so much stronger than Louis/Alex. Louis in particular I'd have thought would be in better shape without Izzy bordering him and with one of the better starting tech pairs; how does Boudica with Hunting/Myst suddenly get so much better than his Agriculture/Wheel when she had a two tech advantage before? I would have thought the different ruleset would have made Boudy weaker, not stronger, so that's a headscratcher. Similarly, with Izzy weakened I'd have though Alex would have survived more often, not less. Though I guess there he might have had too slow of starts without starting techs, or suicided on a stronger Qin?

Very odd, in any case. Once I start posting my other results it will be interesting to see how yours diverge.

As for naming the champion, I don't really have a strong leaning either way on that, but if I had to pick one, I'd probably say the leader with the most wins.

 

[Thrasybulos]

I think that TheOneAndOnlyAtesh who's been doing writeups for Season 8 has caught the AH bug after running the most recently published set, so he'll be cleaning up Season 3.

But then what I don't understand here is why Qin/Boudica were so much stronger than Louis/Alex. Louis in particular I'd have thought would be in better shape without Izzy bordering him and with one of the better starting tech pairs; how does Boudica with Hunting/Myst suddenly get so much better than his Agriculture/Wheel when she had a two tech advantage before? I would have thought the different ruleset would have made Boudy weaker, not stronger, so that's a headscratcher. Similarly, with Izzy weakened I'd have though Alex would have survived more often, not less. Though I guess there he might have had too slow of starts without starting techs, or suicided on a stronger Qin?

Yes, Izzy tends to get carried away going after Religious techs, so missing out on all the free Deity techs must have hurt her.
Louis and Alex had central positions, so they were more exposed. From watching the replays, it seems that Boudica usually benefited more than Louis from Izzy's demise. And Qin was usually the one getting Joao's territory. Bigger + better tech: he had no trouble defeating Alex.

 

[Eauxps I. Fourgott]

A couple more thoughts came to mind about this:

First, 20 kills is pretty hard to reach in these from my experience. Perhaps a low kill count is what you want, in case that's fine. But if not, what about a lower threshold like, say, 15? That still indicates that the leader was quite active and usually getting a lot done militarily, while 20 kills is someone who was REALLY on top of it. Plus that makes it a bit more feasible that one leader could hit two kill points with a really cracked performance; 30 is a lot more attainable than 40 (although I haven't seen it yet). Looking at your charts above, this would equate to three kills in each game, much more like an average AI Survivor match (and less total kills than Season 1 had in those games).

Second, when you reach the later rounds and the fields of leaders change, can you post to show where each leader starts on those maps? Having gained good familiarity with those maps from one group of leaders, it will be fun to speculate how they'll play out differently with a different group.

 

[Thrasybulos]

Second, when you reach the later rounds and the fields of leaders change, can you post to show where each leader starts on those maps? Having gained good familiarity with those maps from one group of leaders, it will be fun to speculate how they'll play out differently with a different group.

Yes, I guess I can split the postings in Opening Round > Wilcard(s) > Playoffs > Championship. I'm keeping Sullla's team orders so where they start is a direct function of the order in which they're listed. And I'm also keeping Sullla's start attribution for the playoffs: so winner of Opening Round Game 1 gets the 1st start in Playoffs 1, etc...That said, after I've gone on about how the live game Playoffs participants were more or less random, the results here so far pretty much match Sullla's games, so not sure there'll be that much difference!

First, 20 kills is pretty hard to reach in these from my experience. Perhaps a low kill count is what you want, in case that's fine. But if not, what about a lower threshold like, say, 15? That still indicates that the leader was quite active and usually getting a lot done militarily, while 20 kills is someone who was REALLY on top of it. Plus that makes it a bit more feasible that one leader could hit two kill points with a really cracked performance; 30 is a lot more attainable than 40 (although I haven't seen it yet). Looking at your charts above, this would equate to three kills in each game, much more like an average AI Survivor match (and less total kills than Season 1 had in those games).

Hmm...
The idea was indeed to have less points distributed, to make up for the fact kill points are a component of the score: since the score allows you to move on (through runner-up spots), it indirectly gives more importance to kills than in Sullla's AI Survivor.
Plus, the "kill points" are the less consensual part of Sullla's scoring system.
But... the idea is also to "collapse" the AH as if they were a single AI Survivor game, so that a direct comparison can be made with Sullla's Power Rating rankings and those obtained here.
Which would imply that I need to try and distribute as many point kill points on average as there are in AI Survivor.

Having a lower threshold would be a solution... except one of the next games is a direct counter-example: kill distribution goes 22,20,12,10 for the top killers. That's a game were I'd want to hand out more kill points that in Game 1 (79 total kills vs 64) and it would end up giving less.
Need to think more on it...

 

[Keler]

This does not make that much sense to me, I honestly perefer if you autoplayed the entire tournament 20 times, instead of proceeding the best two of an alternative history result into play offs, that would be more exciting to me, which would need using different wildcard maps based on number of survivers every time of course, and so playoffs/finals would have 20 different group of leaders, therefore having them played once only. Well that's just me

I personally find kill points wrong and manipulating. I have done a ranking before and someone with so few victories ended up getting ranked higher.

[BTS] - AI Autoplay Ranking- testing with fair maps

ROUND 59 (no show:Churchill,Hatshepsut,Lincoln,Zara Yaqob) agg ai non agg ai

forums.civfanatics.com

I would stick to 5/2 finish rank only.
Just like elo ranking, you make conversion of AH results which is wise. I have no complain for that. Winner of champion should be considered by points but I just don't like kill points for ranking. We can also see AH results of S8G5 where Brennus with 2 victories have same points as Hatshepsut with 6 victories. Well, good luck. I think you should focus on thing at a time and AI League should be the one

 

[Thrasybulos]

I honestly perefer if you autoplayed the entire tournament 20 times

Sum of the averages vs average of the sums, hey? Nice idea, are you volunteering?

I personally find kill points wrong and manipulating.

See, Eauxps? Told you they weren't consensual!
I'm not trying here to find the best ranking system: I have the League for that (and no kill points there ).
I'm trying to adapt Sullla's system to the AH format.

I think you should focus on thing at a time and AI League should be the one

As I said, they're not competing for my time since they don't occupy the same time slots.

 

[Thrasybulos]

So, about kill points...
I think I'm going to for a different system.

The number of kill points attributed for a game will depend on the total kills: total kills / 20, rounded to the nearest.
31-49 kills: 2 points
50-69 kills: 3 points
70-89 kills: 4 points

The attribution algorithmn will be as follows:
1- Sort by kills.
2- Give 1 point to top leader, decrease his total by 5.
3- If still points to attribute, go to 1.

So for game 3 here:
4 points to distribute (76 total kills).
Qin 22, Alex 20, Boudica 18, Louis 11 => Qin gets 1
Alex 20, Boudica 18, Qin 17, Louis 11 => Alex gets 1
Boudica 18, Qin 17, Alex 15, Louis 11 => Boudica gets 1
Qin 17, Alex 15, Boudica 13, Louis 11 => Qin gets another 1
And we're done, all points handed over.

No idea how I'm going to handle ties when not enough points left, though.
I guess here too, we'll have kill steals!

 

[Keler]

Because why should we have just one champion when we can have as many as 20 different champion. Or less.

Sulla should reconsider value of a kill for his ranking.
1 kill should'nt worth 1 point
I would probably change this to
3 kill equaling to 1 point.

I don't think anything is wrong with your conversion above. And I don't have an opinion about it. This entire concept, just like your jumble rumble thread, this feels tricky and not the best way to rank ai and you know it.

I have no interest in volunteering any AHs. I want to do something else unrelated to ai survivor. I may try a ranking with default settings later. But first I need to share what I have done for last 4 months

Alright, time for me to leave this alone and watch the results.

 

[Eauxps I. Fourgott]

Well, if it helps, Thrasy, I like this series.

In the overall I agree that the kill credit scoring is not the greatest system. (Not the worst either; it's not completely without merit and is extremely easy to attribute points for.) In the context of the AHs it can occasionally do silly things (like the set I ran for S8G2 where Gilgamesh used them to tie Justinian overall, despite the latter clearly playing better in the majority of games), but it also works nicely as something easy to figure out from the replays, and I like that there's the possibility of an otherwise dud leader sneaking out a point or two that way.

Anyway, the system that you devised for it in these sounds good. I'd lean towards breaking ties in favor of whoever already has less kills assigned. If that's a tie, well then, I don't know.

 

[Thrasybulos]

Thank you.

I suppose whether Kill points are a good system or not depends on what you're ranking the AIs for.
We usually talk of the "best" AIs, but that's actually pretty meaningless.
If you mean "Best at Winning", then Win points are extremely relevant, Runner-up points are hit-or-miss (relevant only in the case of a close race), and Kill points are not (killing other AIs is one possible path to victory, whether it's effective or not is covered by the other categories).
Now, if you mean "Most Dangerous", ie the AIs most likely to defeat you as the human player, then Win points are also relevant (you lose because they won before you), Kill points are also relevant (you lose because they run you over), Runner-up points are not (unless you win by Conquest, of course an AI's gonna finish second).

I'd lean towards breaking ties in favor of whoever already has less kills assigned. If that's a tie, well then, I don't know.

I'll see when and if the case arises. Might simply go random draw, or skip the tied AIs to give the point to the next in line, making kill steal an official policy.

 

[Eauxps I. Fourgott]

Personally, I don't care that much about truly "ranking" the AIs, I'm just here for the competition (and the interesting offshoots like AHs). So the points assigned should primarily correspond to overall success in the tournament while also just being a fun system on their own.

There's two main issues with the kill credit under the current scoring system:

1) There's no peaceful counterpart. Under the current system, an AI who plays a partially successful warmongers' game (say, conquers two nations before getting mopped up and eliminated by end of game) gets points. An AI who plays a partially successful builder's game (say, leads in tech most of the game and is in position to handily win the space race before getting 2v1d and eliminated by end of game) gets nothing. Ideally, there should be points for both or for neither; why should a warmonger's game be inherently considered more points-worthy than a builder's game?

The issue with that, of course, is that it's hard to come up with a good peaceful counterpart. I don't really have any strong ideas myself.

2) As it's calculated currently, kill credits are a bit arbitrary and can lead to some silliness with kill steals, like Mehmed getting a kill in the latest game when he only vultured a single city at the last moment. I'm working right now on investigating how things would look with a different system for determining kill credits. I'll post details if it proves worthwhile.

 

[Thrasybulos]

Well, for 1) I can see at least four reasons:

1. As mentionned, if we consider that the "most dangerous" AIs are rated, it makes sense.
2. AI Survivor is also a "show": warmongers are just more fun than passive builders.
3. It's pretty hard to devise a meaningful scoring system for builders.
   One point per Legendary City? (before turn XXX?)
   One point for first to Fusion? One point for completing the tech tree? But those are only late games...
4. No such system was in place for the previous seasons, any change to the scoring would have to enable a retroactive scoring: I don't think the data's there for the old games.

As for 2) the issue is to have an objective system. The current system may be flawed, but it's objective.
It seems obvious that when an AI destroys the armies of another AI and then proceeds to capture all its cities but the last one, it "should" be attributed the kill credit. But what if the city captures are 50/50? What if an AI stays at war for 50 turns with another, slowly grinding them both down, and then a third, more powerful AI swoops in: sure, that 3rd AI is the one delivering the fatal attack, but would it have been possible without the previous 50-turn grind? Etc...
For all the egregious "kill steal" situations, there are even more situations which are handled nicely by the current rule, and which would be a headscratcher otherwise.
An objective criterium would be the civ5-style Capital city capture. But.. what if the conquering AI then signs peace?

 

[Thrasybulos]

Anyway, the system that you devised for it in these sounds good. I'd lean towards breaking ties in favor of whoever already has less kills assigned. If that's a tie, well then, I don't know.

I'll see when and if the case arises. Might simply go random draw, or skip the tied AIs to give the point to the next in line, making kill steal an official policy.

And of course, the situation wasn't long in happening!
OK, Kill Steal is the official policy.

If two or more AIs are tied, and there's not enough kill point(s) for them, the kill point(s) go to the next in line.

 

[Eauxps I. Fourgott]

Oh yes, I almost forgot about this:

one of the next games is a direct counter-example: kill distribution goes 22,20,12,10 for the top killers.

I'm guessing this is either Game 4 or Game 7.

 

[Thrasybulos]

Good guess.

 

[Thrasybulos]

So, I'm done with Season 1, just postponing publishing the results until Eauxps is done publishing his results for the opening round games of his S1 AH set.
I'm also essentially done with Season 5 (half-way through the Championship games, that'll probably be over tonight).

What should I play next (ie, are there sets which would be of more interest to you)?

There are no "official" AH for Season 2, and for parts of Season 3 and 8 (I know some games are missing from S4 too, but my set and those only truly overlap for the opening round games).
I could give the guys running those a maximum of time before I "spoil" their results (although it would seem there's a good chance my results would differ enough that there usually wouldn't be a spoiler effect), which would mean something like:
S6-S7-S3-S4-S8-S2
Or I could revert to chronological order (since Eauxps is unlikely to get to S2's AH before me anyway):
S2-S3-S4-S6-S7-S8
Or I could do the most recent first since they're the most likely to be of interest (and I must say I'm curious to test the new Championship map with a different set of leaders):
S8-S7-S6-S4-S3-S2
Or I could do a random draw supervised by a Law Officer.

Any suggestions?