More AI pain for me

Keep going!  Looks like the AI just needs to generate a little mana. 

AI is doing fine on mana, she was just using it.

The problem was I steamrolled them.

Question was, how did you do that?  Did you have a bigger army?  Better balanced army?  AI left cities/resources undefended?  How well were the cities developed?  How advanced was the AI tech?  Magic deployed to better ends?

When you play a better player in chess, you figure out what the other player did right against you, and then try to emulate that the next game?

I never really thought about how useful those graphs might be to a developer. For me it is just a vanity graph. 

Just want to say that not everybody love challenge. Some people hope for extreme AI, while the others just want to enjoy the game. If the AI is too difficult to the second type of people, they will consider the game as a bad game. You just need to make a standard but fun AI for Beginner, Easy and Normal AI, and give everything out to the level beyond that. Players shouldn't blame the developer for easy AI if they play in easy or normal AI; because they should go to the hardest one.

I just like the the idea of history graphs.  In final version, please ?

I'm actually curious about the difference in the graphs.  What caused your population to flatline for two large sections of years while the AI's was steadily increasing?  On the research graph does the Y axis represent research generated per turn?  If so what caused your sharp increase followed by a matching sharp drop?  Did you destroy a research generating building?  Lose an outpost?

All of these graphs make it look like an even match up between you and the AI.  If they're telling to truth and the AI matched your capabilities in each of these areas how did you steam roll them?  What caused it?  Do they just suck at tactics and you crushed them in a key battle?  Did they build bad armies?  Waste those resources (equal to yours) in areas that didn't actually increase their empire's strength? 

I guess what I'm asking is what's the story behind the graphs?

Indeed judging by those graphs the AI was playing fine. It had a scores that roughly equaled yours, even in might. However that equal might score didn't apparently actually mean the AI had a military anywhere as tough as yours, since you steamrolled it. The problem could be anything from, horrible tactical AI, bad unit design, bad strategic unit placement, scattered units, or many other things.

Brad, I know it's tempting, but refrain from touching the "iWin" button.

Dude, what's a turn-based strategy game if you don't need any strategy to play it?  It sounds like Diablo III is the game for you ... not EFE.

I wrote a hole wall ripping brainsucker a new one, but then I realized that he is probably very young or very foreign to misunderstand how difficulty will work. You can just play on novice my friend. 

Yea, the "why" did you beat the AI so soundly is the big question. The overall numbers look mostly on-par right now; so I suspect that the tactical AI is still a bit hopeless.

There is the question if attacking is too easy vs defending. If defender has not enough advantage, then AI should use steamrolling too. That makes for a dull game.

The typical situation is that you have two cities you need to defend. If the attacker can mass his forces against one of the cities, and choose the city to attack at will, then defending both cities needs 2x the army the attacker has. So, the defender should either have enough time to see the attack coming and mass his forces against the attack based on that, or defending the city should have some advantage. Halving movement speed in enemy territory could work here. A leader who can lessen this effect would be really valuable. The overwhelming maneuver speed was one of the key reasons why Genghis Khan and Napoleon were so successful in their military operations.

In middle ages the attack/defense balance was somewhat correct: a small force in a castle could defend against a much larger force for some time. This gives the defender enough time to mass a counter attack force and break the siege of the castle. If the defender doesn't do that, the siege will slowly drain the defenders supplies and thus the attacker can still conquer cities easily enough, it just takes some time.

In Civilization the balance is implemented through attack/defend values, and strong fortify and city wall effects. This made the attack/defense balance interesting. Due to how tactical battles work, it seems hard to implement a similar balance mechanism in FE.

Finding a good balance to the attack/defend ration is the key IMHO. If attacking is too easy, then the AI has little hope defending his cities. And if that is the case, attack is the best defense. This will mean short and chaotic games. On the other hand, if attacking is too hard, then that makes for a dull game, too. If you always need to siege a city for 10 turns before conquest, it will be boring...

Of course it is possible the balance is correct, and the problem is in fact the AI not building big enough army. Without playing the game it is hard to know. So, the logical conclusion is that you must give me access to the game so that we can find out what the real problem is. Right?

I am planning on making capital cities invincible until the endgame as a possible strategy. Will the AI be able to be adjusted to a defensive strategy? Some AIs should steamroll or die trying. Others will likely go for defense and try to fortify a rich piece of territory. I just want to make sure that the number of personalities is not hardcoded like it was in GalCiv2. I would like to be able to make new ones.

Stardock's done that in every single game since GalCiv I.  No worries there.

My request for AI, is that it will play a competitive game with me, as long as I play "legitimately".  In GC2, there's a few things I don't do, because I think they exploit the AI.

The question in my mind is: What is the counter to a Steamroller?

1. Perhaps withdraw all of your units in one city?
2. Sneak around the enemy to attack the cities in his rear and raze them (my favorite strategy is 1) conquer, 2) raze, 3) move on to next city and repeat)
3. Take peasants out of production and use them to fortify your cities?
4. Recruit like crazy in your home city?
5. Magic Buffs cast on your units while you are awaiting its arrival?

There needs to be a counter to the Steamroller effect otherwise games become very boring as soon as you get going (more cities conquered, means you can conquer even more cities and even more, etc.).  Maybe there needs to be a very defensible home base that takes a long time to conquer usually.

Best regards,
Steven.

I think what he meant is that he wants multiple flavors of AI with personality, rather than one very good AI that's that plays the same way regardless of the faction (which is what WoM's AI kind of seemed like). In that respect, I agree with him.

I'd rather have a bunch of pretty good factions that play very differently against me than extremely good factions that play the same way.

One sollution to this could be sieges, like in the Total War games. Unless you bring some special siege equipment with you (which is slow and potentially vulnerable), it takes a few turns to build up to the battle. This gives the defending player time to maybe marshall a relief army, or at the very least prepare other cities that might soon fall to the enemy.

Also, freshly conquered cities shouldn't just become stable and contributing to the new owner, it takes a while to quench unrest and start full production.

A very good point! In WoM all you could do was have a bigger steamroller.

Fortifying cities isn't really a counter to steamrolling unless the defensive bonuses are good enough to allow you to hold off or destroy greater numbers. If it is a stat bonus then it just means the enemy will wait a bit for a bigger steamroller. This is why I support a waiting time for armies attacking fortified cities. It would allow you to tie up blobs and strike around them. Sieges generally could take weeks or months to complete anyway.

I believe there is now max army sizes so that helps, but you can never have enough strategic military depth.

Edit: Lol Satrhan beat me to it!

Keep up the work frogboy. A good (not necessarily a extremely difficult one) AI is important for a good game, and too many games lately have had bad AI's spending all their money and graphics and fluff. I know you already know this but I just wanted to say: Keep up the good work!

This would make researching Fortress even more important. Not just a 25% bonus to defender's strength.

• Attack Fortress city with Seige Engines = immediate attack
• Attack Fortress city without Seige Engines = attacking army waits 1 turn per city level to start tactical battle.

Example:

• Attack Level 5 Fortress city without Seige Engine Unit in your army = Army waits 5 turns to attack

I love the thought of near invincible cap cities, would also be nice for minor factions and possibly wildlands.  In Europe Universalis III they use the production of different lvl's of forts to keep steamrolling to a min.= no fort- simply move on the land & take it, lvl 1 fort- you have to seige it for @ a year, lvl 2 fort - ya have to seige for @ a year & 1/2, lvl 3 fort - 2years and so on, this would help keep factions that don't rush armies/war tech around a bit longer.  

Some ideas for the attack/defense balance. This all assumes there really is a balance problem instead of just not-yet ready AI.

• Summonable guardians. These are not city-guardians. These are units which are fast moving. They are powerful inside your borders, somewhat weak outside of your borders (so you can transfer them to disconnected sections of your territory) and really weak in enemy territory. This would allow for counter attack outside of cities, too.
• Automatic city guardians, either magical units which draw their power from population (big city, powerful guardian), or they might be normal units whose level is based on the city. They are not transferable. There might be different units for different factions. Some factions get powerful guardians, and so these factions are great for the turtle strategy. Per-faction guardians could of course be implemented for the non-city guardians idea, too.
• Maybe go even further in the city-guardians idea. You could just define the wanted defense level for city (with a slider, of course). The more you put to defense, the more powerful your defenders will be. However, the more you put into defense, the more your production suffers. This way you don't have to build the defenders at all, you just define the amount of defense you want. Raising the amount of defenders takes considerable time, so you can't just put the slider to maximum when under immediate threat. The best part of this idea is that you don't have to actually build units for city defense, which is dull micro-management type work (IMO, of course).
• The siege idea above (turns to siege based on city level) is a good idea IMO.
• Advantage in tactical battle for city defenders/in your own land. Squares giving better defense, multipliers to health/attack etc.
• Spells available only when defending / in your own land. The city might have a magicians guild which allows for spell casting locally (this is against the lore, I am afraid).

It is not hard to come up with more ideas similar to the above ideas. I think some of them not only solve the balance problem, but would also allow for different per-faction strategies and make the gameplay more focused on strategic level instead of micromanagement.

I don't like the siege idea at all. It's just a patchwork on what should be a real siege system. Either do it right, or don't waste resources on it.