Beating the Toughest AI

[checkMate]

I find that if I just set my resources pretty close to the edge of the city radius, the Nukes do very little damage. They always land it right on top of the city.

[tetley]

The catch is if you do that, something else gets it. I only make a point of placing Wonders out there, for that reason. Learned that lesson after my Colossus got fried.

[Adagio]

checkMate >>

I think we can find your message...

[rah]

And if I see that message again, I forsee a free trip to the rock for a week.

RAH

[Aginor]

Originally posted by solo

I wonder what the difference has been in our last few games, as I have had a relatively easy time of it by adding many archers for defense until siege are available. Now I make sure each city has an archer and each tower has two before adding another city, and go for attrition sooner than I used to in games when I was hurt early.

Good God man! Where do you find the Wealth? I crank out a 4-Bowmen roving defense force in Ancient, and occasionally add to that. It's a miracle if I can afford to slam up a Tower in every city but the capital (assuming protected capital by boundaries). Too busy filling 3 Universities...if I lag into Gunpowder by more than 2 minutes, that means that they have Bombards, and I don't, and that's just NOT good.

Your strategy on farm timing for the Brits was quite helpful, actually, ran it yesterday morning and it GREATLY improved early efficiency...I had no fishing and still was able to hold my own early. Got trounced by the WORST nightmare of the Brit archer force in Medieval - a swarm of Elite Javs and Knights, stopped & went to work...goofed two ways that game, built the first Tower in the wrong place AND made a big error on Mine placement by putting my first Mine AHEAD of my left flank city when I had two protected locations available...not smart.

My test on starts showed a big disparity in results after only 5 minutes, which makes me think that the gathering rates of the human player may be secretly adjusted based on things like tech acquisition, number of cities, etc. We all know that in Toughest, that the AI gathering rates have been increased by 50%, and if the game tinkers with human rates, too, this would not be noticeable without comparative tests.

Possible, though I deem it somewhat unlikely. More likely it has to do with number and quality of the Ruins bonuses you get (timing is everything on those). Fishing also has a huge impact, I find...the games you get where you have 400 Wealth at 6:00 when you upgrade to Classical run a lot faster to Gunpowder than the ones when you're poor, for instance.

[Aginor]

Two firsts:
1) Two in a row with the Brits! WOOOOOOO!

2) Won one without building either the Statue or Versailles (grumble #^&@ Egyptians grumble)

Oddly, comp didn't EVER send a single unit into my turf until Info Age in the first game...and got slaughtered for failure to disrupt the boom. Think I more than doubled up the second-highest score (final was like 55k), won a Wonder victory (up like 29 points) outside the last enemy capital with 65% of the territory, outearned all 3 comps in every resource except wood...ugly stuff.

Building additional Towers seemed to help in the second game...repulses pre-Bombard attacks pretty handily. Played around with helicopters some...they're nice since they never need to land, but VERY fragile. My feeling was if you have pop limit issues (and I always do with the Brits, Colossus or no) and more resources than you know what to do with (again, not an issue with Brits) then Helicopters on infinite queue flagged to the general rock Bombers for sheer firepower any day of the week.

[Out4Blood]

Oddly, comp didn't EVER send a single unit into my turf until Info Age in the first game...

This is why I am not a big fan of FFA games when testing validity of strategies. Half the time, the AI gets into fighting itself and the other half all 3 decide to attack you at once. The optimal strategies then become ones that avoid getting attacked based on the set of rules the AI uses. This results in bad play habits that will get you killed in multiplayer (e.g., not having to defend prior to info age!)

If you are really testing your own prowess versus the AI, to improve your multi-playing skill, I'd recommend playing ONE AI, normal speed, no pausing, with other "standard" settings.

Of course, everyone is free to make their own choices, and if you don't ever plan on playing multi, then it doesn't really matter.

[solo]

Aginor,

Some additional wealth comes from establishing trade routes earlier and by not putting so much into scholars early on. I compensate for this by getting a lot of wood and aiming for those Scientific upgrades as soon as possible. When basic defenses are in place, wealth is shifted to scholars using the infinite que.

Nice going on those wins as the Brits!

Out4Blood,

You have a very good point and it may turn out that just playing against one AI opponent is a better test of skill on Toughest, since they often end up attacking and weakening each other in many 4-player games, leaving the human player a lot more time to boom.

I have learned a bit about predicting AI behavior, making the last few games go much easier, but a key reason for this success was my policy of building up a good defense early. In games when I didn't do this I was attacked early, often and almost continously. So ignoring one's military can be equally disasterous in SP and MP games.

I'll have to try a few games against just one AI on Toughest to compare the difficulty, as I see that vs. only 1 opponent there will be more time to build a competitive military early in the game and less chance of being attacked on two fronts simultaneously. Closer proximity in 4-player games meant contact and conflict much earlier than in games vs. 1 AI.

No doubt the best test of overall skill is by playing against human opponents in MP games, and I can't help but agree that practicing vs. 1 AI on Toughest without pausing would be better way to practice for MP games. However, Toughest games vs. more than 1 opponent DO teach you to become more efficient and help you in deciphering the reasons behind AI decision making.

[Aginor]

Originally posted by Out4Blood
This results in bad play habits that will get you killed in multiplayer (e.g., not having to defend prior to info age!)

First time I've seen that in many, many games, actually - which is what makes it remarkable. Usually at least one AI sends the kitchen sink at me no later than Gunpowder. 4-man FFA is always a bit of a crapshoot...if all three players gang up on you, and they're competent, you're screwed no matter what. I fail to see how the dynamic is much different from MP with the same number of players other than the fact that you can't effectively raid the other guy circa Gunpowder (due to the fact that the AI will immediately throw the kitchen sink at you and bury you in units, which is NOT a rational and normal human response given two other rivals).

If you are really testing your own prowess versus the AI, to improve your multi-playing skill, I'd recommend playing ONE AI, normal speed, no pausing, with other "standard" settings.

Only problem is that the AI is horribly predictable in single player. Build cities in an inverted T, rush the center, fortify the front city like mad, and you're set until Industrial...adding the third and fourth player removes your ability to utilize this exploit. I've found that trying to win 4-man after mastering 1v1 excises a lot of the bad habits you can get away with in 1v1 vs. the AI. The necessity of having to capably defend more than one front levels the playing field somewhat, and there is a lot less margin for error if you expect to win consistently.

In other words, while 1v1 is a more consistent test of skill, there are a lot of lessons to learn from 4-man FFA as well.

I don't doubt that you're a better player, Out4Blood, but I also suspect that you have an RTS background...we graying TBSers understand strategy but are still getting used to the interface, control groups, and the fine art of hotkeying REALLY fast...it's not an easy transition if you're used to spending time over your decision making.

[tetley]

Hey,

I'm beating Toughest resoundingly as Mayans on Sea Map. I can't really explain why, other than I build coastal defenses and the dumb AI keeps throwing away his ships at it. Very little land military except for Spies and Archers, and the Archers get ZERO upgrades (don't need 'em--they're garrisoned). Usually a Wonder victory, but a D-Day victory works almost as well.

[Out4Blood]

Originally posted by Aginor
I fail to see how the dynamic is much different from MP with the same number of players...

First of all, most multiplayer games are n vs. n, where n >=1. FFA is less common, primarily for its unpredictability and perceived unfairness. (e.g., a good player will generally immediately be ganged up on.)

Originally posted by Aginor Only problem is that the AI is horribly predictable in single player.

It's no less predictable than in FFA. You've already determined the "attack lowest score" algorithm. And beating the AI in toughest mode 1on1 is not always that easy, since there is nothing to distract him from his goal of destorying you.

Originally posted by Aginor In
other words, while 1v1 is a more consistent test of skill, there are a lot of lessons to learn from 4-man FFA as well.

As in all games, FFA and equal teams are virtually two distinct games. What sometimes works in FFA would NEVER work in multiplayer (e.g., keeping a small strong territory to avoid attacks, NOT building in the center of the map and grabbing terrain, etc.) Unless of course you were playing multiplayer FFA, which is rare.

Originally posted by Aginor I don't doubt that you're a better player, Out4Blood, but I also suspect that you have an RTS background...

I play ALL games. I grew up on TBS games but have shifted to RTS games because they provide a better challenge. The "think forever" mindset of the TBS community is annoying to me. In fact, when playing TBS games, I always prefer using a chess clock. It vastly speeds up play and inspires a sense of urgency in the players. Real generals never have unlimited time either, and it's the requirement to make decisions quickly and with limited information that makes the game interesting to me.

[Hermann the Lombard]

Well, in between 1v1 and 4-"man" FFA, would 2v2 with a computer ally be a useful compromise? Rushing the center might not be such a good idea, as you might get pincered by the two AI enemies. Just a thought...

As for me, I don't feel up to "toughest" just yet (though I *should* be ready for it after THIS thread!) But 1v1 at "tough" isn't really a challenge anymore. There are challenges, yes, but the outcome is not seriously in doubt.

Last night I was attacking the strength-7 Incan capital in CtW in Enlightenment, and initially I thought I was in trouble, backed up against the wall with an initial AI city about where I would normally build my 2nd city, and an army already on its way to my capital. But I was fine, throwing up two stockades flanking the city and tossing damaged musketeers into garrison as needed. Then the AI continued to attack me there while I expanded laterally and put down 4 more cities. Boom for my Bantu, bust for his Incans. Wide and thin would be suicide against a human opponent, but the AI single-mindedly wasted its forces (as usual).

O4B: I've never tried MP because I'm too darned slow ("glacial" is my usual term). I'm a dyed-in-the-wool TBSer and I "think forever" in spades. It's starting to frustrate *me* because I'm hardly getting any sleep thanks to RoN.

Huh? Wha'? "I just want to take this one more city, then I'll go to bed!"

[Out4Blood]

Try using more hotkeys, waypoints, queues, and rally points. You'll start to get bored on normal speed against the AI.

[tetley]

I've confirmed, an Incan spy strategy works pretty well on Toughest/land maps, unless there's a shortage of hills. Spies cost metal and gold--something the Incans ought to have plenty of. And you get refunds when your bribed units die. It's critical to get to Classical and get mines early, because you NEED at least 5 spies by the time the AI's second wave attack comes along. Definitely get Angkor Wat.

So far I've beaten the AI Toughest multiple times with Incans, Mayans, Russians, and Bantus. Of those I have to say Mayans/sea maps are by far the easiest.

[Out4Blood]

I enjoy the British the most. Everytime I go to pick another civ, I end up picking British again.