Sullla's AI Survivor Season Eight - Game 5 Alternate Histories Thread

[Eauxps I. Fourgott]

Sullla has completed and posted the "emergency" alternate histories for Game 5 of this season! The real game saw pregame favorite Darius fall flat on his face and secondary favorite Pacal recover from a weak start to barely edge out Hatty for second place, as Ramesses spent the game unmolested and cashed in with a Cultural victory. But was this "supposed" to happen?

Writeup: https://www.sullla.com/Civ4/civ4survivor8-5A.html

(No video for this set)

Short version: The Pacal truthers were correct! He was by far the most reliable AI on this map, not so much because of dominating the games himself as because his opposition was all quite weak and he thus successfully pulled in front after weak starts time and time again. Overall, the result we saw was fairly typical, except that Pacal had a somewhat weaker game than normal, and Ramesses' win was something of a fluke result that came out of nowhere.

Discuss below!

 

[Eauxps I. Fourgott]

My own thoughts:

When I first looked at the initial results table, my reaction was "wow". I thought after the real game that I thought I had read this map well; I was wrong! I thought Pacal would be a fairly weak leader here, that Ramesses had a good shot (and in particular better than Asoka's), that Darius was in good shape, and that Brennus had zero chance of winning, and all of those assumptions were incorrect.

I think I fell into the same "close, but no cigar" trap as Sullla with regards to Pacal. We both identified that he had a slow-starting position, which was correct. I also thought his diplo wasn't very good, which wasn't such a good guess. But the main mistake we both made was in thinking that his opposition would punish him for his slow starting, when in reality they were not very strong and didn't give him that hard of a time, letting him consistently build back. He did not seem to be very hampered by early wars here, or outscaled by anybody else, so yeah, in that light, his good performance makes a lot of sense. There needed to be a really aggressive unfriendly neighbor near him for him to have a really tough time here, and that didn't happen. Good job to those who picked Pacal; once again, I wrongfully doubted the judgment of the sizable minority!

Darius's performance was a "big oof" and does make it look more like he's an overall poor leader who's been carried by some excellent luck in the past. I am still overall a Darius supporter as I like his boom or bust tendencies, but much as it pains me, I have to admit that I think I may have been overestimating his skill in the past. Is he more like a luckier version of Victoria? Possibly, though I'd like to see AHs for Game 6 before committing to a response on that one.

I was also crestfallen to see Ramesses' win proven a fluke, though he was at least a good pick for second place and the second most likely to advance. (Those who care about the "correct" leaders advancing can at least rest easy for this game!) So my original optimism on his position was unwarranted, but my later assessment, made in the time immediately before this game, was correct: safe position and better odds to advance than Asoka, but not that likely to win.

Two wins for Asoka was quite the surprise and I was a bit disappointed to see that I was wrong there. Ditto for the Brennus wins.

As for Hatty, it sounds like this confirmed my existing feelings about her overall AI Survivor performance, which is neat to see. (To quote my writeup from S1 alternate histories: "Hatty is good at culturing in situations where she is left entirely alone. However, as we know, this is an extremely rare occurrence; in most cases a leader is going to have to face a hot war at some point or another. And when Hatty has to do so, she’s one of the worst leaders in the game.") Certainly her worse performance than Pacal here makes her last-minute elimination an easier pill to swallow. I wonder how much longer this "0 overall points" farce will continue. It gets more darkly humorous with each season that she remains the only scoreless leader...

 

[Thrasybulos]

I think it'd be interesting to run a few more of these games, with but a tiny edit : providing Darius with a source of Horses.

Asoka and Ramesses getting into conflict was a possibility. That it'd prove to be a feature is a tad unexpected, but partially explains Ramesses' lacklustre results I suppose : Asoka can be tough 1v1. He often builds a ton of units when on the defensive.

I believe the AI code for signing peace has been delved into and laid bare somewhere, and that one of the surprising insights it provided was that the "Aggressive AI" option made the AI more likely to sign peace.
Now, my beotian impressions from simply observing a lot of games is that AIs will often sign peace when in a war not of their choosing.
Hell, in a recent game, I had AI A plotting against AI B (using letters because I don't remember which). B was the only possible target. Then AI C declared on AI A. Although AI A won the initial fight and could easily have conquered AI C, it quickly signed peace... and in the same turn declared on B (which it proceeded to conquer).

 

[Keler]

Maybe it is not Darius himself but the starting position turns out to be weaker then our expectations. I voted Hatshepsut to win with Darius second that game. Which could have been more frustrating if I have picked Ramesses for second. I was close to do so. I just remember many conversations going around in discord that Ramesses will be invalid weeks ago. People thought his start was weak and wildcard material. Brennus winning 3 times is even a bigger shock because I really thought this was a setup where he had no chance. I think after season ends if Thrasybulos tries with that swapping around starting positions Hatshepsut will be top finisher.

Overall I still can't understand why the winner mostly ends up as outlier result that never repeats in alternative histories. Just crazy.

The fact that Brennus wins 3, Hatshepsut wins 6 yet they are at same points shows power of kill points in such rankings. If you remove kill points, Hatshepsut is not alone and some other unlucky leaders such as Napoleon,Roosevelt as well as zero points Genghis Khan is there with her.

 

[Kaitzilla]

Hatty deserves some points eventually.
My Darius pick was terrible, and I bet the farm on Hatty dying first.
Nice work to all those who picked Pacal to win!

Overall I still can't understand why the winner mostly ends up as outlier result that never repeats in alternative histories. Just crazy.

Random.org is firmly in control and will win the picking contest this season

**Edit**
Thanks Eauxps I. Fourgott!

 

[Fippy]

Ramram winning 0 but getting 2nd place so often makes me even more suspicious..
but about alternate histories, not the real game.

 

[Eauxps I. Fourgott]

Also, change the title of the post to ?

Sullla's AI Survivor Season Eight - Game 3 5 Alternate Histories Thread​

Oops, fixed!

 

[lymond]

Hatty deserves some points eventually.

Yep...the old gal has it in her

 

[danjuno]

I actually did consider Ramesses at one point, given he had a corner to himself and was well away from Brennus, but ended up going with Pacal. But at least there's some validation in the AHs.

 

[Saxo Grammaticus]

I also considered Ramesses, but the Alternate Histories got the better of me, and turns out his win was an outlier anyway

This game has helped dispel my Pacal skepticism. If you look at the Alternate Histories, he has a bad track record in the opening round, a chance of advancing between Hammurabi and Genghis Khan... I found that discrepancy with his success in the playoff rounds to be odd, perhaps the result of facing high peace weight leaders in the one and low peace weight AI in the others. In the context of all rounds these results push Pacal into the upper tier of performance or just outside of it, more or less his position in Thrasybulos's AI Jumble. His success with this field leads me to suspect map balance could be to blame for his underwhelming opening round record.

And Darius slips further into mediocrity...

 

[Eauxps I. Fourgott]

I somehow completely forgot to look for this when the set first came out last month - one of the biggest things I'd originally been thinking about for this game: Was a nobody-dies scenario realistic?

I am crestfallen now to go back and realize that no, it wasn't. I was quite sure that we'd see at least one such result in the alternate histories, but there were none to be found. Really whiffed this map!

 

[Plains-Cow]

Ramesses had the plains-cow start. Game was ogre before it even began. He's just being humble in the alternate histories, clearly.