Beating the Toughest AI

[Aginor]

Originally posted by la0tsu

If you read my thread entitled "Bantu Boom", you will see that my opening move is to research Civ1 and immediately upon completion build those two cities.

Interesting...I go with the standard build order for most civs: Civics1 + 3 Citizens, 2 Farms, 2 Citizens for 5 Woodcutters, Science 1 (ruins bonuses) once I hit 120 Wood again...Citizen. This cleans me out on both resources, and I wait a few seconds and send off two Woodcutters to build City2, replacing them as I am able.

City2 goes up fast with 2 workers, and typically I send them off immediately to City3. Waiting a bit permits me to scope out City3 in particular to make sure I get a solid location for it...in my experience as long as your capital is decent with solid woods and at least a 5-Citizen mountain you can subspecialize City2 and 3 (one wood/farms, other metal) and use City4 to cover the rest of your basic needs, which makes exploration and boundary pushing with Temples less important early.

However, by the time I'm ready to begin conquest, I have more resources than I can reasonably spend, so the added pop cap becomes the most important factor at that stage.

Agreed, but I am finding as I play with other nations that the boom really gets rolling much later, that the 2-3 minute turn advantage with Bantu is just really strong, and that I don't want to cede part of that back by building that Wonder. I want to get the Statue ASAP, as it drives my Cataphract strategy, and building the Colossus slows me down by a minute or so unless I am just rolling in money, timber, and spare citizens early.

That said I may consider starting to throw it up right after Terra Cotta in Medieval, which is an approach I've never taken. After all, at that point I've got Construction1 and a Wonder work force of 4+ laborers...

It should also be noted that my perspective is that of a perfectionist.

You will find that my tendencies lie along the same lines, but my objective is a sudden crushing military advantage rather than crushing resource advantage in the late game. I back up the initial invasion by capping out resources, but it's a #2 priority for me...the most important thing is buying as many cheap Cataphracts as possible and transforming them into 20-30 Tanks, which is just flat out devastating...the Terra Cotta Riflemen can help take care of any Anti-Tank units, and with air power a non-factor until Modern, if you supply some artillery backup then victim #1 stands no chance...

I would recommend you read the Bantu Boom thread to have a comprehensive view of my MO.

I'll take a look at it - must have missed it initially, I'll try and revive it.


As for bad habits generated by playing 'D' until Industrial, the problem with that is that you're failing to disrupt the opponent's production, which will come back to haunt you against a more competent human general that will not squander his resources on useless units. If you raid, you can often force the opposition to use more resources defending the raid than you invest attacking AND shut down the production of a city for the duration of the raid, which is pretty strong.

I'm with Grond on the necessity of a fast Commerce3 against Toughest, as well as the fact that this is ONLY smart against the comp, which just flat doesn't attack after Classical until Gunpowder in most games...but let's face it, Catapults suck . In the games where I've stolen Versailles I've found it to be VERY nice as you can just keep rolling on with the same army rather than waiting for resupply from way back in your lines, but I suspect that it is much better for a nation that has pop cap issues and simply cannot afford to field as many units.

Also, I suspect that a lot of us are breaking down the AI because we are not RTSers at heart and use 'P' quite a bit...I've never scored much over 30 on speed, and in an ugly game I'll chalk up 20-25. Learning my way around hotkeys now so I can try MP.

No games last night, it was my anniversary (3rd) and I do NOT play strategy games well with a head full of champagne. I'll try the Brits again, and I'll let you know the results.

[solo]

Happy Anniversary Aginor!

I agree on many points, and am also doing some to learn the hotkeys and start playing without so many pauses, which I tried during the last part of my East Indies game as the Brits. It's hard to remember to continue research and builds while absorbed in a big battle though! I had a lot of unspent resources by the end of the game!

I finished my Toughest game with the Brits on the East Indies map and the headstart into Industrial described in my last post was all that was needed. It was 4 minutes before any of the AI joined me, but by that time I had scouted out my first target, Tenochtitlan, the Aztec capital, while building up an invasion force on Glasgow’s island.

The attack was nothing fancy. First, a V2 was used to take out a tower. Then some helos were sent out to create a diversion, while a group of special forces trotted in next to Tenochtitlan to reduce the city’s hit points. The main army followed, protected by a carrier and some subs, crossing over to the Aztec island and arriving in time to capture and hold the Aztec capital. (I think the loooong carrier was more of a hindrance than a help, since it caused pathing problems for many of the transports)

By the time Tenochtitlan was assimilated, I had advanced to Information and had all of the latest weapons. This big technology lead assured success, and since the wonder timer was started by captured Aztec wonders, I decided to end this game early by bombing the Japanese wonders that were under construction. There was no point continuing conquest when it was going to be this easy.

The game stats had two surprises:

1) None of the AI built a fort. I thought that was odd, but maybe they just don’t do this on island maps. Anyways, this gave me a big edge.

2) I collected the most timber! I usually emphasize wood collection, but this was the first game in Toughest where I was able to outdo the AI in collecting any resource besides knowledge.

One game may not prove anything, but it appears that the AI are quite handicapped on island maps. This game was easy compared to previous ones played on land, so I will go back to a land map again as the Brits for more masochistic pleasures.

Another thing worth mentioning is that the Porcelain Tower’s 200% benefit applies to the initial amounts of food, wood or whatever is offered. For example, the PT only added 20 food to a special starting out at 10, even though I had increased the original 10 food to 15 with granary research. I also noticed that it does nothing to improve fish. Maybe this wonder should be renamed the Piddly Tower, considering its late arrival, high price and for all the good it does. I don’t see fast shipbuilding as a benefit either. Need more ships quickly? Add another dock or two, instead. It will be many moons and another game without difficulties before I build this wonder again.

Another thing I've noticed is that the last three ages at normal research speed zip by too quickly after knowledge has been maxed out. The expensive and slower speeds provide much better play during the last three ages.

Attached is the replay of the last part of my game.

[Aginor]

Thanks solo, I appreciate the congrats.

Got smoked as the Brits again, this time in Gunpowder...GOD I suck lol.

Some thoughts:
1) The vaunted Brit Taxation bonuses are not all that until Social Contract...then they are quite strong.
2) Don't try a goofy research order aimed at getting to 250 wood and selling the excess kids...it doesn't work, you lag behind in tech badly and never recover.
3) Bantu research order just won't work with the Brits. Looks to me like you need to get Medieval and Science3 ASAP for Granaries, Lumber Mills, smelters, and upgrades so you don't overbuild Farms and Woodcutters, and so you can slam up Smelters ASAP.
4) You have GOT to get decent fishing going early as the Brits to fuel Archers or you are toast.

Leaning more and more towards the Incans as a civ that can address my early poverty issues. But now I feel like I HAVE to figure out the Brits

I've outcollected the AI in Oil a few times, managed to outcollect Wealth and Oil and Knowledge once, that's about it.

The Porcelain Tower is a souped up Lighthouse, which means it still stinks...

Going to bed early, I'll look at the replay over the weekend. And MAYBE I'll manage to not get whooped again.

[Out4Blood]

Interesting...I go with the standard build order for most civs: Civics1 + 3 Citizens, 2 Farms, 2 Citizens for 5 Woodcutters, Science 1 (ruins bonuses)

My build order is different. I want to maximize the production in the beginning. IMO, as British, the critical aspect is max food, then max out wood, then max out food again.

So I start the game with VQ, then use the 2 cutters to drop 2 new farms and send all new villagers to wood. If I have crappy wood spot, I'll build a second right there. After the first 2 farms are built, I'll wait til 120 wood and research science as my first tech. There is no hurry on research, since it is not critical path. (Note: if crappy wood spot, I'll skip science research and go civic/commerce, because wood becomes more of a factor.) Food is critical path (for villagers and the civic/commerce techs) and wood is only needed for getting farms up. The science then makes subsequent techs cheaper, so might as well research it first. Dropping an early city actually sets me back and makes it slower than normal.

Then I drop a new city, and get to +90/+100 (cap is +87) on each before researching commerce. With this build order for British, I am usually able to keep pace or even lead the AI on resources - at least in the early game.

The key to British is to REALLY expand the territory. The taxation bonus is only useful if you have a lot of territory. Try to never build near the edge of the map. You cut in half the range of your city. I build out, even if it is not near resources, to maximize the area of my territory.

Some micro tricks for defeating AI armies...
Garrison a couple archers to max the building's firepower bonus, then fill the rest with slingers. When they attack, set rally point for building away from their ranged troops. Exit the slingers, who fire on the hoplites. The hoplites then attack and get chewed up by the archers. Regarrison before losing any units. Rinse and repeat. You can do the same with the archers which are even better. But archers die fast to slingers, so be careful.

[Grond]

Do counter bonuses apply to garrisonned units or do buildings have their own damage type?

[solo]

I started a new game as the Brits and was given the Old World land map, so here goes!

My scout found the Eqyptians first, who were booming, so I decided to target them with an early rush. I very much agree with Out4Blood about how the Brits need to expand their borders quickly to make the most out of double taxation.

The rushed worked, and with it came no less than 4 Egyptian cities, which when added to two of my own, gave me "control" of over 40% of the map. However, this may not last too long, since my own development was interrupted by an almost successful Korean rush of my own capital a little before I had reduced Eqypt's.

The other AI are the Germans. They are definitely booming, since they have not attacked me yet. This nixes my theory about them being an aggressive, rushing civ in Toughest. My newest theory is that either a booming or rushing script is chosen by chance for each AI at the beginning of the game. Later on, I think they are cut loose from the script, my guess being that this happens sometime after Gunpowder.

Right now the Germans are way ahead in production and were into Gunpowder before I hit Medieval. This does not augur well for my extended and feeble empire, over which the sun may be setting soon when I have a chance to continue the game. However, so far this is my most promising Brit start on a land map.

[Grond]

While I love the Brit commerce limit I think their real power lies in taxation. I find it ineffecient to max out the British commerce limit at all times. I'm usually halfway between the standard limit and the British limit. However the taxation is just huge. You can max out scholars very fast and make more scholars than other civs if you want. If you don't make liberal use of the market you are forgoing Britains biggest advantage. The key is to keep the market tab open so you can constantly monitor the market prices. Food rarely drops too far so I only buy it when really necessary however you will often see times when the price for timber and iron dips considerably. I will buy huge amounts at that time. The other thing the British can do is make tons of Heavy Knights in Gunpowder and Enlightenment. Iron often drops to 20 around that time for some reason. Then when you reach Industrial buy some Oil and upgrade all those Knights to Light Tanks. In every game where I was able to do that I dealt the AI a fatal blow shortly thereafter.

[FireWall]

I'm new to the game, and can't win at Moderate, so I really appreciate this discussion. Maybe now I will win! Thanks everyone, and especially Aginor.

[solo]

Yes, that extra wealth from taxation really lets the British take over later in the game, but it takes a while for this benefit to materialize.

I have been trying to figure out the criteria used by the AI on when and who to attack. So far it looks to me that if they are scripted for an early rush they will build a barracks immediately and then pump out a slinger and 4 hoplites. Once this team is assembled they always seem to go after the civ with the lowest ranking, which at that stage of a game is usually the human player. Proximity, prior contact, etc. do no seem to matter. This might explain why Aginor and myself have seen AI rushes coming from the civ located way across the map.

Later on, I believe a similar process is at work. Once an AI has built up an attacking force appropriate for the current age, it bases its decision to attack on who has the lowest ranking. Again, the simultaneous attacks we've experienced in Toughest games can be explained by this, since they usually occur early when we are rated weakest.

Now in games where I have not been in last place, I've not been attacked even when I felt I was vulnerable. For example, after a successful rush and assimilation, the gains in territory, buildings, commodities, etc. will often put me into 2nd or first place, which lessens my chances of being a target, even though at this point I would be quite vulnerable if one did occur.

After continuing the game described above a little further, I found myself in second place behind the Germans, but well ahead of the Koreans. I have been expecting a German attack, but they are already in Enlightment, and nothing has happened yet, which leads me to believe that their first "Gunpowder" attack was aimed at the weaker Koreans.

It will take more close observation of relative power rankings to see if this theory holds up, but it explains what I have noticed so far in Toughest games.

[FireWall]

That is a very interesting observation, and as far as my games are concerned, it appears to be entirely accurate. I am always attacked by the AI when my score is low, even if I'm well prepared and able to fend off the attack with ease. However, if my score is higher than someone else's score (which isn't often ), that person is attacked and I'm left unharmed, even if I am particularly vulnerable.

[solo]

Thanks, FireWall for that confirmation. I think there may be other things that trigger AI attacks, too, such as attacking them first or spending resources on a wonder, but when they are just sitting around wondering who to go after, I think game rankings determine who their target will be.

I continued my land map Brits game. The Germans built up quite a lead and finally attacked Whitby, my city closest to their border, soon after I researched Gunpowder. I just managed to update my siege weapons in time to hold him off.

After this there was a brief pause, allowing me to quickly take out a Korean city under construction that was threatening to encroach on my diamonds merchant.

After this was done the Germans were back for more fun and games, which continued for the rest of the game. They had a big edge, so the first few battles were defensive ones, where they were able to reduce Whitby a few times, but never were able to follow up and capture it. They took a lot of losses which allowed me to catch up quite a bit and actually get into Industrial first, saving up resources that were needed to rush the Statue of Liberty for unit upgrades.

The Germans followed quickly into Industrial, bought a bunch of tanks and went after my capital with them. Things were not looking too good, but this raid was made without supply, and attrition along with my beat up and minimal army was enough to hold them off.

During this action, the Koreans decided to attack one of my Eqyptian cities, too. There was nothing I could do about this, except fortify the other cities in the area that were not under attack yet. So the Koreans took El-Armarna, but that did not hurt as it was a small city without any improvements.

By that time, SoL was completed and my problems were solved, since I could start mass producing tanks and aircraft while the Terra Cotta kept pumping out up-to-date infantry. SoL also made quick work of the expensive job of updating all my Terra Cotta slingers (yes, they were all still slingers hiding in cities, forts and towers) into more modern infantry.

The Germans were still strong and beat me into the Modern Age by a few minutes (Info Age too, a bit later), which made them good for one more big assault on Whitby. They took out its fort and one of its towers, but I played defensively until catching up in tech and while building up my armor corps and air force.

After that it was finally on to the offensive. I attacked Hamburg first, their nearest city, using bombers and artillery, and bv the time it was reduced I finally had a good sized army and went in and captured the city. From then on it was a similar process with two more cities, during which I built Supercollider and World Government. After taking the third German city the instant assimilation triggered a victory.

Well, I've finally got a land map win as the Brits, after several previous games where I had been stomped by the AI, instead. I think I would like to try this again on another land map, but not open with a rush as I did in this game, just to compare which way of playing the opening with the Brits is best. It takes a while, but their commerce bonus pays off eventually.

Attached is a replay of the last half of this game.

[Grond]

I really think the British are the strongest Civ for taking on the AI. Once that third city is up the gold starts rolling in and your getting more resources thanks to the commerce limit. If you get a decent sized lake you can get enough fishing boats to allow you to pump some archers early and forego scholars until your third city is up and you get taxation. After that you can make scholars so fast that you will catch up in no time in research and probably get to 28 scholars first. British archers plus some garrisonned slingers will stop any early rush and those archers will get upgraded for free plus your towers and forts have extra range. The British get their other really handy unique right when I'm ready to go on the offensve which is the Highlander. By the time they are upgraded to Black Watch I'm usually handing the AI his hat. While none of those abilities by themselves is particularly devastating when added together they make for a very formidable Civ.

[tetleytea]

I've beaten Toughest with Russians, just as long as AI doesn't rush too hard. Get a quick Fortress up, go hard Spies early. Bribe like mad. Russian Spies aren't just half-price--they have a significantly higher life expectancy, so they're like way cheap.

Meanwhile, tech up to Modern. By the time you hit Modern, you rule the map, with all your UU's.

Seems to work better on water maps, since early on the AI seems to love wasting all his ships on your coastal fortresses.

[solo]

Grond,

The British are good, but I think you have to be patient for their advantages to pay off. After an “easy” win on a water map, and another win following a successful opening rush on a land map, I’ve tried a few more games on land maps without trying an opening rush, and still find it tough going in the early game. In two subsequent games where I expanded to 3 cities quickly, I was attacked early and often and ended up losing a city early and losing both games. On my third try, I played it a bit more conservatively, making sure of good defenses before expanding further. This put me further behind early on, but I hung in there and was able to manage a win later.

Tetleytea,

Nice going on your Toughest win. I have not tried the Russians, yet, but am working my way through the civs while playing on the Toughest level. I’ll look forward to using those Russian spies and some of their other UU’s. Their attrition bonus should help very much in the beginning of the game, where the human player needs the most help when playing at this level.

Once I get the hang of how to play a civ, I move on to the next. I found water maps to be much easier on Toughest than land maps. Some more details of my successful game as the Brits follow:

In this 4 player survival game in Toughest, I was given the Amazon Rainforest map and had the Germans, Spanish and the Japanese as opponents. Of the AI, the Germans turned out to be the strongest, as they have been in several other games where they appeared, so I’m looking forward to playing this civ myself later.

I played a conservative opening, and decided to research attrition earlier which meant a tower near my second city very quickly. This was followed by a barracks fairly soon, which built some archers for early defense. Before starting my third city I built a tower in that area and put an archer in it. The Germans were booming and romped way ahead, almost doubling my point score, and were the first to attack, but this was handled very easily with attrition, the tower and archers.

There was only one other attack against me in the first half of the game, coming from Spanish, who were closest to my third city. They were able to destroy my tower, but wasted themselves away while making this attack. The city was never in any real danger.

Following this, I built a fort in between my second and third cities and which combined with the Amazon terrain to give me a compact and easily defendable corner of the map. I think that turtling in a corner presented the AI with better opportunites to go after each other. The Germans and the Spanish had expanded rapidly towards the middle of the map, and this cut me off from any contact with the Japanese fairly early.

After the game was completed and when I was viewing the replay with the map revealed, I noticed that several of the Japanese attacks had diverted or weakened German or Roman troops that had been headed in my direction. Even though I was very low on points, it appears that my early policy of laying low and not expanding aggressively for space or resources, saved me from becoming the target of several potential attacks. Although the AI will usually target the weakest opponent (namely myself), they can be easily distracted from doing this by each other’s aggressive behavior.

However, one consequence of being given a lot of peace and quiet was that I fell quite a bit behind by the time I had completed Terra Cotta as my first wonder. At this time, I saved the game and quit playing, and was starting to think that I had played it too conservatively so far, and would never catch up.

The next day when I continued the game, I decided to emphasize science while building up my economy. Again, I was surprised that I was left in peace for a very long stretch, but the Germans were really racing ahead, getting to Industrial very quickly. So far in Toughest games, I’ve pinned my hopes of winning on getting to Industrial first and taking the advantage of early oil. It was quite a few minutes until I was able to catch up, but except for a few German fighters harrassing Terra Cotta, I was left pretty much all alone until I had the Statue of Liberty built.

By this time, I had science maxed out pretty well, and was able to build a modern army and start blitzkrieging the nearest Spanish cities using bombers to reduce them, and following up with armor and infantry. I pressed on towards Madrid when I should have waited. It was easy to reduce, but a strong counter-attack caused me to fall back and re-group before I could assimilate the city.

Meanwhile, Germany was still ahead a bit in research, beating me into Modern and just inching me out into Informational, when they decided to build the Supercollider, my favorite wonder. So I fired a missile and found out that they do no damage to wonders under construction. However, bombers are good for this, so they were used instead, while I started building my own copy of the wonder. After building it, Supercollider gave me a big edge, since I was able to research end game techs afterwards instantly, accumulating three of them very quickly.

With World Government, the second attack on Madrid was successful. Instant capital assimilations make quick work of conquering another civ. It took awhile to use the same trick on the Germans, who had the AI tech in time to make things wild and woolly. We had a tremendous battle along our border, and I was beaten, but then instead of marching in and threatening my cities, the Germans turned right and went after the Japanese, who had been chipping away at their other border. By keeping up constant pressure on Berlin with bombers, I was finally able to get some infantry next to the city in time to take it instantly and end the game. It took about 4 or 5 attempts, though.

This time I have included two replay files in the attached zip file, br1 and br2, which together cover the whole game.

Incidently, two things I’ve noticed while playing RoN so far:

1) The AI will not build any wonders until after I start on my first one. After that they build as many as possible. It just may be a coincidence noticed in my limited number of games, so I’m wondering whether anyone else has noticed this?

2) Likewise, this game and any previous one I’ve played in which I did not research or use nuclear weapons has led to a similar result. As with wonders, if I don’t build them, neither do the AI. Again, what’s the experience of other players on this?

[solo]

I've moved on to the Chinese, who look to me to have a pretty good advantage over other civs, as I was able to keep up with the AI with instant villagers and the 20% break on science research. I was first into Industrial and went into the lead while there, getting and keeping a very good lead in Science and Commerce. (Science VII while the AI were at Science IV). This game was on the African Watering Hole map vs. the Germans, Maya and Koreans. I was able to build Colossus in addition to the usual wonders (TC, SoL and Supercollider).

Near the end the Germans built and launched some nuclear weapons my way even though I had not researched this, so the observation in my previous post has been proven wrong. However, all AI still waited on wonder building until after I had begun my first one.

I believe that instant villager production is a HUGE advantage for the Chinese, since this allowed me to stay almost even with leading AI in the early game. With all other civs so far, I've always fallen way behind early because of that AI 50% gathering bonus in Toughest. It was almost like having that AI end game tech early.

Anyways, the Chinese are now my civ of choice, but there are still 14 others to check out, to see how their civ bonuses compare.