v142.8 on Emperor

Txurce

Deity
Joined
Jan 4, 2002
Messages
8,285
Location
Venice, California
I just finished my first game in a while, and have some points to raise about the current state of VEM. Most of my comments are here, but some are in other threads.

For comparison, I again played Korea on Continents Plus. It was a rare overall dominant victory for me, although Japan edged me out in score. I launched on t286, at which point my 8 cities led in pop with 69M people (3 were captured, 2 were big capitals). They ranged in size from 15 to 35. Needless to say, this was a first.

My science got off to a relatively slow start (low pop), but exploded to heights I’ve never approached before: 3553 on t254, and over 4500 before goosing it up to 5200 while timing my last tech and SP. I researched Apollo in 5 turns, and built my last parts in 2 or 3.

America finished second in science with 1212, by which time it had lost 40% of its size. Persia was a powerhouse even with only three cities. AI science is strong enough, given that its population still grows quickly. The other civs didn’t favor science.

Increased AI aggressiveness (Aztecs denounced me on t22!) and military power led to more eliminations: two civs out, one reduced to an island, and the others either warring or saved by chokeholds. I puppeted the Japanese onto my continent, let them develop, then bribed them into warring against co-leader America. In a long, medieval/industrial war by which America had fighter planes toward the end, Japan won. It was a great example of the benefits of buffing the militaristic civs.

The CS algorithm has the AI back to the old normal. It’s worth considering nudging it back up - although as we know, better too low than too high.

The AI remain slow building GW’s. It’s particularly easy to pick up 2-3 key early ones. This has yet to change, despite all the recent adjustments.

It was unclear where the AI spent its gold. (It probably has less than ever!) Many were broke near the end. They could use a little more science and production at this point. I would tilt toward more science, since this helps keep them competitive militarily as well.
 
I can't say with certainty what adjustments the AI needs science-wise, and have not had the same experience with AI purchasing as you (In my latest game they seemed to purchase quite well), but I certainly agree with you on GW's.

I've found myself able to have whichever wonders I choose, and what's more, can often wait a fair few turns after enabling them to begin construction. I'm not sure why this is. Maybe it was always like this, and the improvements to the AI in other areas have made this flaw more noticeable. I can't say for certain.
 
Coincidentally I just played the first part of an Emperor game on v143 myself last night. Addressing the OP comments without quoting:

For reference I've played my last four games or so with India because the population and worker are conducive to my tall playstale (EDIT: and I wanted to keep some consistency between games).

I had hills, marble, AND stone near my capital and then got a +3 production event so I managed to bag pyramids, GL, *and* HG. I started out probably second or third in science, and once the GL was up I stayed competitive throughout.

I started near Siam and slightly farther from Japan, who took Siam's capital (but not other cities) relatively early in the game. The aggressiveness is something that actually seems well executed.

I've settled four total cities and the culture expansion seems good with tradition helping out now. I didn't have problems with leaving gaps between cities.

The AI CS purchasing seems lower again, which is better than off the charts for sure.

The AI has been really weak on wonders. In addition to the ones mentioned earlier I got Chichen, PT, ND, and am about to burn a GE on Sistene. I think there may even be another one I forget. This is a stark contrast to previous games where I got beat to pyramids and GL, which hampered my overall development.

I'm also unsure about the AI purchsing. I would definitely lean towards buffing the CS purchasing slightly and see if/how that helps science before adding a significant science buff. I definitely agree that science is a better bonus than production.
 
All of Zaldron's comments dovetail with my experience.

I had hills, marble, AND stone near my capital and then got a +3 production event so I managed to bag pyramids, GL, *and* HG.

This is close to my situation - no stone, but three stable resources. The stable can be very powerful early on.
 
All of Zaldron's comments dovetail with my experience.



This is close to my situation - no stone, but three stable resources. The stable can be very powerful early on.

This further correlates something I've been mulling over for sometime. In the early game (say the first 50-70 or so turns on standard speed), production is a significantly more powerful yield than the others. If you're working some basic tiles and have say 6 production, then create a mine and get lucky with a production opportunity, you more than double your production for working a single tile.

I know that there's always some element of luck to CiV, but different starting are production can make the difference between hitting several keys wonders and missing them by 5 turns (although the gold recompense is much nicer now). If you wind up in a very flat area with say only one cow in the way of convenient production tiles, your chances of getting wonders is reduced, while at that stage of the game the small additional growth you get may even be undesirable due to relatively low levels of happiness.

I like the early production opportunity a lot but I wonder if it should be +2 instead of +3, and if there's anything else to help equalize early production a bit. The late-game one can still be +6.

EDIT: One other thing I note, and I don't know if it's a problem: With the new 15 free gold per turn I now buy the monument on turn 11 every game because it's silly to use that early gold for anything else, or to hard build the monolith with production that could be used on scouts, warriors, workers, or wonders.
 
This further correlates something I've been mulling over for sometime. In the early game (say the first 50-70 or so turns on standard speed), production is a significantly more powerful yield than the others. If you're working some basic tiles and have say 6 production, then create a mine and get lucky with a production opportunity, you more than double your production for working a single tile.

I know that there's always some element of luck to CiV, but different starting are production can make the difference between hitting several keys wonders and missing them by 5 turns (although the gold recompense is much nicer now). If you wind up in a very flat area with say only one cow in the way of convenient production tiles, your chances of getting wonders is reduced, while at that stage of the game the small additional growth you get may even be undesirable due to relatively low levels of happiness.

I like the early production opportunity a lot but I wonder if it should be +2 instead of +3, and if there's anything else to help equalize early production a bit. The late-game one can still be +6.

EDIT: One other thing I note, and I don't know if it's a problem: With the new 15 free gold per turn I now buy the monument on turn 11 every game because it's silly to use that early gold for anything else, or to hard build the monolith with production that could be used on scouts, warriors, workers, or wonders.

Agree on both the early production and gold points. Of course it's fun, but it can really give the human player a head of steam.
 
I just finished a second, equally dominant game with Korea on v143 Emperor. What seems most obvious to me from the AI performance in these two games is that 1) it's not at all competitive pursuing GW's and 2) it seems to compete slightly less than ever for CS. By the latter I mean that only occasionally did someone bid past my status, and never countered right back. I don't know if the CS dynamic can be incrementally increased, but a moderate boosts regarding GWs would help. (In fairness I can only speak about early GWs, because after I took a big early tech lead, the rest were easy.)
 
Sounds like it's time to reintroduce the AI production/maintenance bonuses.

I agree that the free gold in combination with cheaper early buildings seems unbalanced, and can potentially speed up the early game tremendously (depending on terrain). Can the free yields scale for difficulty, Thal?
 
I can confirm Txurce's post.

One thing which surprised me in the game I played was that the AI did not try once to get one of my CS. I had the UN build and was prepared for a bidding war against the Aztecs (which had killed Washington and took some of Askia's cities) with 66k :c5gold: gold and nothing happened. Granted, my influence with all of them was between 250 and 300 but I still expected more on Emperor.

The :c5war: militaristic AIs were far behind in techs (right as I finished the UN on turn 342 (epic speed) Cathy entered the Renaissance). The :c5science: science from the patronage policy seems weak. My 16 city state allies gave me ca. 190 :c5science: beakers at the end.
 
The :c5war: militaristic AIs were far behind in techs (right as I finished the UN on turn 342 (epic speed) Cathy entered the Renaissance). The :c5science: science from the patronage policy seems weak. My 16 city state allies gave me ca. 190 :c5science: beakers at the end.

I find the science from Patronage weak as well. With regard to AI science, I think boosting their science bonus even more than production will make the game more competitive.

By the way, Thal, it's a pleasure to be back talking about helping the AI rather than the human. Congratulations on successfully digesting the AI-gold-spending elephant! VEM is going to be in by far its best shape ever heading into G&K.

I'm also looking forward to trying your improvements on whoward69's utility - particularly the Alliance turn counter. That was major.
 
I don't disagree about giving a science bonus again, but providing hammers will help with wonder competition (especially the key early ones) and maintenance reduction should provide more CS competition.

Edit - hadn't seen your post about the return of the CS issue when I posted this.:) I think giving AIs' capitals a few extra production on higher difficulty could be an alternative solution for wonder competition.
 
It's easy to adjust resource balance and placement if some resources are stronger than others for early start locations. There's a function I can modify in the "assign starting plots" file which balances start locations. :)

@Zaldron
Buying a scout would probably be a great investment for India/Korea, who don't start with any exploration units. Scouts are cheaper than monuments and can get you extra gold/culture/science from exploration. For non-India/Korea games, workers are more cost-effective to buy than monuments and only cost 30:c5gold: more. A quick early worker is typically more useful than a quick monument, if we have farm/mine tiles to improve nearby. Buying a worker for quarries or plantations usually ends up with the worker idling for a while, so in that case I buy a monument instead. Sometimes I even buy a Scout as soon as I hit 100:c5gold: for a big exploration bonus.

@pthmix
Once the Globalization tech is unlocked by any player, all AIs should start spending all their gold on citystates. Could you start a thread for this in the bug report forum, and if possible, attach a lua.log file from the time period in which they should have been purchasing influence.

@Txurce
The turn counter for alliances is very hard to do because for some unfathomable reason Firaxis did not make it a "deal" like research agreements, borders, etc. I'm researching ways to figure out the turns remaining, but it's taking more time than the others.

@pthmix
I could help militaristic AIs annex puppets to improve their science rate.
 
Buying a scout would probably be a great investment for India/Korea, who don't start with any exploration units. Scouts are cheaper than monuments and can get you extra gold/culture/science from exploration. For non-India/Korea games, workers are more cost-effective to buy than monuments and only cost 30:c5gold: more.

I would love to be able to hover and see which is a better deal to buy.
 
Isn't that what the multiplier is for (i.e. 3.1x :c5gold:)?

Of course you still have to decide if you want a fast scout or a fast worker.
 
Continuing my story from earlier in this thread (yes it really does take me that long to play a single game).

Right where we left off, Japan DoWed me. At the time I think I may have had a slight tech edge, but definitely not in military techs. My first though was to build a few units and go on the offensive. I didn't do that though - instead, I consolidated all my units and fortified in rough terrain where appropriate, bringing out the war chariots and elephants for ranged goodness.

Japan easily conquered one of my CS allies, but I Stayed the course, continuing to build NWs in my capital and science/happy buildings in my three other cities. I also set my tech to beeline to rifling (I was three techs away and Japan had just started fielding Longswordsmen [which is another issue - I couldn't tell if it was a CS gift or Japan not getting Samurai], and Japan's cities had 20-27 def so longswords would have been a tough sell defending while my trebs took down the cities.

I had started honor for the barb bonus, so while I was researching rifling I took a complete diversion from my peaceful policies and went straight down the middle of honor to the gold/upgrade policy. Simultaneously I started training ships of the line (three in total) because 4-6 of Oda's cities were prime coastal targets.

Rifling, the honor policy, and my HE/armory capital all finished about the same time so I upped all my longswords to rifles, popped a couple cannons and then switched from defense to totally obliterating his entire empire (save one pop 5 city) in somewhere around 30 turns. My cannon/rifle group took the interior cities while my ships of the line and two knights took about four coastal cities simultaneously.

This was all pretty surprising because Oda waged two really good wars against other AIs early in the game, even taking a capital. I'm eager to try the latest version that restores a small era-based science bonus.

Almost immediately afterwards my now-neighbor Ghengis DoWed me, but this was trivial to fight off.

At this point I had nine cities, probably 4 over 20 pop and the rest around 12-18. After taking Kyoto I was pretty much running away with the game, because Washington hadn't been able to expand quite enough to get his powerhouse tech machine as strong as mine.

I could have won culture vic about turn 246 or so, but I hadn't launched in ages so I didn't ever build Utopia. Instead, I continued researching and launched in....I *think* around turn 285.

All in all a very enjoyable game especially with the warlike AIs giving me a solid run for my money until I just out-teched them like crazy. I was actually hoping Khan would declare on me again at the end so I could see how fast I could take his whole empire with my tanks/mech inf vs his keshiks. :)
 
It's easy to adjust resource balance and placement if some resources are stronger than others for early start locations. There's a function I can modify in the "assign starting plots" file which balances start locations. :)

I'd prefer lowering the free gold from 15 before messing with starting plots - variety is the spice of Civilization!
 
I was addressing Zaldron's point that start locations rely a lot on luck. The per-player gold is the same for everyone, while resources are different. If we reduce the similarities, it would emphasize the differences, and make Zaldron's problem worse.

I think your question is different... that the early game goes faster? The goal is to make it both faster and more balanced across the starting locations.
 
Top Bottom