v139 on Emperor

Korea again, and a quick overview as of t120:

Pop/Science

America 7.6/359
Greece 5.2/199
Persia 2.6/139
Denmark 2.4/130
Korea 2.2/193
China 2.2/95
Russia 2.2/140

Most out of whack is that Greece has allied with 13 CS against me. How did he even meet that many so soon, let alone afford them (and be left with a treasury of 754. with net gold 99)? Well, I noticed he had a heavily promoted caravel (huh?) on t87.

Having researched Compass that early seems surprising; the gold required for alliances, inexplicable.

The AI promotions would be a good idea in a game where the AI doesn't grow and produce gold so quickly.

I've also noticed that the AI has a generally surlier attitude, and is less likely to trade OB's or luxuries at full price. Again, this could be a good idea, creating more of a range, but doesn't work very well when the AI is raking in so much gold.

Walls seem necessary, given how easily an AI with a big army is likely to attack - and attack the city, rather than huddling around it. This is a really good thing... except again, there are too many AI units, due to all of the above.

Overall the AI produces way too much gold and too much science. My guess is that it's due to population more than anything else, but there may be other ways of tackling it rather than reducing the AI food supply (although I think that would be ideal.)

Alexander is one of only two leaders with a CS-alliance trait so I gave him a high diplomacy flavor. Where are you on the tech tree? Our turn # doesn't impact the game and depends on speed, so I rarely notice it.

Here's the changes I've made to AI bonuses in the past week (since March 22). I picked emperor because that's the difficulty you play on. Green is where I made the game easier, and red is where it's harder.

I'm on t120, and just researched Physics (Medieval).

My point is that no one should have 13 CS allies on t120 at Emperor difficulty.

Looking at your chart, why not lower the AI Growth % (assuming that's pop and not expansion)?

On t125 of a Korea game on Emperor, I managed to get the GL, HG and PT, and just entered the Renaissance with four cities. I've also been able to stay out of wars, and will now build hwachas. So I am as stable as I've been in a few games, and no one is dominating the 5 CS I've met so far. The pop/sci standings on my continent:

India 6.2/188
Siam 4.1/103 (one city)
Japan 3.0/150
France 2.2/109
Korea 1.5/147

While this game looks like it will be tough, that's okay, since nothing seems out of whack at this point. (Of course my GW's have helped unduly.)

On t250 in the game above, Korea had the GL, HG, PT and SoL, 6 cities, and steering a rampaging Japan into picking on other neighbors. I took one French city and sold another, and maintained 3 DoF's most of the game. At this point I was researching Computers, four techs behind the leaders. The pop/sci stats were:

Inca 76.7/1926
India 64.6/1337
Egypt 31.1/1730
Japan 29.3/561 (pop came from conquest)
Iroquois 23.2/1697
Korea 10.9/1380
(Siam and France were eliminated by Japan)

The Inca then won a Cultural Victory on t251.

The game was fun in that I was competitive. Happiness was tough due to no CS allies, and if I had much higher pop, would have been in the red. But 1380 beakers (1800 including RA's) is huge. Losing on t251 seems early, since I rarely win on Science in less than 250 turns. I'd say that AI pop and science are too high - just look at how many civs have lots of beakers - but that the most warped aspect of the game right now is the CS system. There is no way to compete with the AI for them. On the other hand, the AI could afford RA's plus high pop plus CS plus big armies (ranging from 232k to 331K for theAI, 63K for Korea.) All in all, Emperor seems too hard to win tall via Science right now.

Txurce said:

The Inca then won a Cultural Victory on t251.

Click to expand...

I snipped a whole bunch from your post just to comment on this: It's happened to me a couple times too, usually Egypt or Siam for me though. I'm happily pursuing a science victory and then one of the AIs pulls off a cultural. Just for era context in one game I had *just* finished rocketry and in the other I was even farther behind in tech (I *think* replaceable parts).

Might the reason for Alex being awash with gold have something to do with the bug I encountered where my city tile has a 157 gold yield from one turn to the other? That's one alliance every two turns or so...

And yes, I did drop down to King for the game today, not sure if that's not a good thing Also cultural victories won by the AI aren't a bad thing, we just need to have reaction time and a possible reaction. Starting a war is not always an option and even then, conquering one or two cities isn't enough...

Maybe add in a switch that only every 3rd or 4th buy with the gold can go to CityStates or just cap them (no influence bought when x influence points in front of next competitor)?

And I agree, science is still fast.

EDIT: Further proof that the "gold on tile" bug makes the game unplayable. Infoaddict shows a clear ~150 units spike in net gold for Gandhi from turn 32 (16) to turn 33 (168). (Btw. why can't I make a screenshot of Infoaddict? when I paste it in paint only the normal view shows, not the infoaddict window). Do you need a safe? but I don't have any near that spike since I just saw it now...

Agreed that t251 is too early for the AI to be winning. If it's consistently Culture that's doing this, it shouldn't be too hard to fix: just increase SP costs, yes?

The 3-digit gold-per-turn tile bug has been reported at http://forums.civfanatics.com/showthread.php?t=457995 , to make it easier for us to discuss the issue in a single place.

wobuffet said:

Agreed that t251 is too early for the AI to be winning. If it's consistently Culture that's doing this, it shouldn't be too hard to fix: just increase SP costs, yes?

Click to expand...

Unless I'm missing something, wouldn't that hurt the human player just as much?

That's why I keep looking for the underlying factors that supercharge the AI performance.

Ah. I guess I was assuming that culture games were going OK for the human.

If only AIs are winning Cultural Victories before t300, perhaps their Culture discount on SPs should be toned down. (It's amazing how much more competitive AIs are in VEM, even with a lot of their bonuses slashed!)

edit: I guess AIs don't even get a discount on SPs anymore!

I just played a quick partial v139.2 game on King for comparison - over 175 turns - and the competition for CS seemed like the "old normal." This is only one game (as opposed to several on Emperor), the AI may crank up more slowly on King, and the only classic CS hoarder I met so far - Greece - was wiped out. All that could be mitigating factors. But the AI had plenty of gold; in fact, on one turn France had 2915g.

AIs do not have a culture discount. The chart I posted above lists their bonuses.

Thalassicus said:

AIs do not have a culture discount. The chart I posted above lists their bonuses.

Click to expand...

I guess I missed that update. I think the only explanation is that the AI is much, much more competitive in VEM than in vanilla (a good thing!), especially now that they spend their gold now. If this persists after the gold per tile bug is fixed, I think I'll have to drop down a level, which I don't mind at all.

We could try these values for v140? The AIs will be equal to us in everything but unit bonuses, and their starting units/yields.

What I don't follow is that presently the AI is producing so much gold (and pop, etc) with seemingly modest bonuses.

If v140 made the AI identical to the human player, meaning they would have more or less the same science, gold and production potential (with an edge in worker rate and units), then can you isolate what is giving them the edge they have now?

If you could, then I'd suggest halving the difference first (since making the equal except re: units sounds like a prescription for defeat against most VEM players).

Another approach could be to put the AI in a handicap situation with regard to food. (That is possible, right?) This would allow them to keep the bonuses they have now, which allow them to stay competitive in other areas.

Txurce said:

What I don't follow is that presently the AI is producing so much gold (and pop, etc) with seemingly modest bonuses.

If v140 made the AI identical to the human player, meaning they would have more or less the same science, gold and production potential (with an edge in worker rate and units), then can you isolate what is giving them the edge they have now?

If you could, then I'd suggest halving the difference first (since making the equal except re: units sounds like a prescription for defeat against most VEM players).

Another approach could be to put the AI in a handicap situation with regard to food. (That is possible, right?) This would allow them to keep the bonuses they have now, which allow them to stay competitive in other areas.

Click to expand...

Maybe the AI is producing so much gold due to the following bug:
http://forums.civfanatics.com/showthread.php?t=457995

What I don't know is if a tile affected by this bug could produce tons of food, science or hammers instead of gold. That could be an explanation of high AI population, research, etc.

If the gold problem is as simple as what Pep has shown, v139.2 may not need any major adjustments. The population factor may not matter, given the recent AI nerfs. I'm still curious as to what causes it, though. What would happen if the AI food % were dropped to 85 or 90?

Thank you for keeping refining this game!
I tried games on Emperor and was wiped out at about T 200 v139.2
Game 1 Rome overran half the continent and had 12,000 gold
Alexander had only 300 gold each turn but had 800 plus points with EVERY city-state.

Game 2 it was Alexander who started the CS bidding, but this time Germany bid them all up to 700 plus points. His per turn gold was about 100.