Please correct me if I'm wrong, but isn't Tough the last AI level before the difficulty factors shift from limits on AI ability to bonuses for the AI?

Actually, I'm no longer sure what level is what anymore. There's a 100% level where the AI runs at full potential but gets no bonus, a 105% level (5% bonus), a 125% level and the 200% level. AFAIK these have been changed with v1.31 and there are more levels between 125% and 200%.

Again AFAIK the 200% level means that the AI get 2x *everything*, economics, military production, social production, research, miniaturization, diplomacy, sensors, range, yada, yada, *everything*.

From a "fairness" point of view I think the 125% level is probably the most reasonable place to play. The extra 25% makes up for the intrinsic advantage the human has without being a ridiculous advantage.

But what do you do once you beat this level, and can do so as a matter of course? They only real choice is to up the difficulty. You continue to up the difficulty until winning is no longer guaranteed and you find yourself playing against the AI with a truly ridiculous advantage.

The biggest part about winning against a 200% bonus is in your own mind. You can’t imagine that you can do it and so you can’t. I remember the first suicidal game I played, it really wasn’t that long ago. I got a little past the colony rush and I was so far behind in absolutely *everything* that I felt I had no chance. I even posted a question asking how to beat the suicidal AI. For yuks here's the thread.

How can you compete with the suicidal AI?

I got a few responses but the answers really didn't help that much. Anyway, I went back to the game determined just to fight it out, but convinced that I would lose. To my surprise, I won. The funny thing about this game was that although I selected the incredible (200%) level for all 9 AI's, the game only counted as Masochistic because I had changed the starting relationship with the AI's to friendly. Whatever, from then on I was hooked. Yes, I get to a point in the game that I know I'm going to win and it becomes a mop up exercise. But I haven't had a suicidal game yet where at least at one point or another I was sure I would lose. That's what makes it fun. Hope this helps you make the transistion.

The problem would be the game probably wouldn't be much fun. It would be fair, but not fun... (and that's probably why the AI doesn't use speed)

I doubt that the reason the AI doesn't use speed is because of any concern about fun. It's just not programmed for speed.

But as far as not being any fun, I disagree. I used to play a game called Reach for the Stars. Old game on an Apple IIe. It was a totally different game than GC2, much more simplistic, but for whatever the reason it wasn't practical to try to defend planets with fleets of ships. You had planet defense bases that would damage an attacking fleet somewhat but really couldn't protect the planet, just slowed up your opponent a bit. The key to winning that game was to take your opponents planets faster than he could take yours. Also you had a better chance of protecting conquered planets than planets that you originally colonized. So part of the strategy was to allow your opponent to colonize the better planets and take the 2nd class planets for yourself. This could backfire if his production base was a lot better than yours but otherwise was a successful strategy.

Anyway, the game where your racing to take more of his planets than he takes of yours, or trying to take selected important planets that cripple his production base while protecting yours, can be a very exciting game. I think any game that you have a realistic chance of losing is exciting. If you *know* you're going to win from the start, why bother.

I think any game that you have a realistic chance of losing is exciting. If you *know* you're going to win from the start, why bother.

Amen.

defense in this game needs an advantage Absolutely true. Right now there's basically no viable defensive strategies other than military starbases.

I agree with this. Actually, I disagree that military starbases are a viable defense. They're expensive to research, expensive to build (especially if you're not evil), and they really aren't any better for defense than they are for offense. Military starbases don't move, but constructors can move just fine and they can turn into a fully upgraded starbase instantly. If you're going to build military starbases, you might as well build them in enemy territory, use them to help conquer the planets within their reach, and then use them to defend your new planets.

Defense needs to be stronger than offense, otherwise offense is the best possible strategy. If offense has the advantage, then when two opponents are evenly matched, the first one to attack will win. After he has won, he has a larger production base than he had before, and a larger production base than the other players in the game. If he attacks, he will win easily. It's fair, because all players can use this strategy, but it isn't fun because the only way to win is to build the strongest military possible and attack.

But, if defense has the advantage, then other methods of winning are possible, because defense alone can't win the game. You can build a cheap defensive military that can protect you from an enemy with a stronger, more expensive military, and then use to resources you didn't spend on military to pursue a different method of victory. Non-aggressive strategies are only possible if defense is stronger than offense.

*yawn* This thread made me tired from reading it, was there an argument? Since as always this site is filled with educated people, something most sites lack.

In several threads, developers have flat out said this is the case. In fact, it would make no sense that the algorithm would stop working on a software level at a certain number. No sense at all.

Unless everyone lines up to buy an outranged Cray-2,,there is no way for the current game to be playable with speeds reaching 80 and above.


Now lets see what updating the hardware can do:
http://preview.tinyurl.com/y9ck4a
This leads you to Tomshardware.com
Updating from an Athlon64 2800+ to the topnotch of Core2Duo cpu systems increases the computing capacity by factor ~4.

For going from a Pentium3 to the Athlon,,the factor is ~6.
(No link for this,,so either you have to make your own comparisons or just take my word for it)

Assuming that the P3 is good for 6 moves,,how many more moves would you get from multiplying the computing power by factor 24?


Frogboy made the comparison between the AI and the chesscomputers.
Until DeepBlue came along,,no computer was able to beat a chess master other than occasionaly.
The reason they can do that is that you have a very limited area,,a set number of pieces who's movement are predefined.
Still,,just the first move by each side leaves (8x2+2x2)^2 permutations.
The next has between (7x2+1x4)^2 x 20^2 and (7x2+1x1+2x2+3x6x4)^2 x 20^2 possible permutations.

IN GalCiv,,the area is much larger,,the number of pieces and their movements are unknown.
Care to make a quick estimation of the number of possible permutations that has to be calculated before the AI can pick it's choise?

In short,,that means that the workload raises far more rapidly than the suggested hardware upgrade can compensate for.

In short,,that means that the workload raises far more rapidly than the suggested hardware upgrade can compensate for.

You could even look at it this way: Most computer chess programs can easily beat a human opponent (chess masters not included)

On the other hand, in Galciv 2, even the suicidal AI, with its *massive* artificial bonuses, is easily beaten once the player discovers it's weaknesses.

On the other hand, in Galciv 2, even the suicidal AI, with its *massive* artificial bonuses, is easily beaten once the player discovers it's weaknesses.

You may be beating the suicidal AI easily, but I'm not. I'm beating it, but by no means is it easy. Perhaps over time it will become easier. Also, it's by no means *everyone* that beats the suicidal AI. It may be a little presumptuous, but perhaps those beating the AI at suicidal are indeed the equivalent of “chess masters”. Of course, it’s Brad’s opinion that those beating the suicidal AI are using cheese and exploits to do so. The reality of the situation is probably somewhere between these two extremes.

I tried suicidal, just for kicks the other day. WOW   It looked like rainbow(tact view) of high speed colony ships going through my space. I got thirteen planets while the lowest AI had 28, and the Yor had 56. I hung in there for good long while, but I was too far behind in everything. Used the mil. base array to get a respectful military rating, trying to keep the arceans off my back. I saw them loading up the ships near my SBs and around me, so it was only a matter of time. When they finally came for me, it was a slaughter. I had been playing painful and doing ok, but my colonizing strat was blown out the water and to make matters worse, when I realized my mistakes tried to change strat mid game. I just call it: A Lesson Learned. I'll try it again some time, but will probably just stay going up the difficulties one at a time.  

You may be beating the suicidal AI easily, but I'm not. I'm beating it, but by no means is it easy.

Good point. It works the other way around too. You find it a lot easier to make tons of points, while I don't. I think it's due to my style of play. I have a system of beating the AI, but it doesn't really maximize points. I'm going to have to try larger galaxies even though I hate micromanagement, because on smaller galaxies it's a lot harder to specialize planets (since you have so many of them)

Anyway, the secret is, IMHO, diplomacy. Give your race as high of a diplomatic bonus as possible, always be neutral (so that you that you have less alignment penalties), research as much diplomacy and trade techs as possible, and get Galactic Showcase ASAP.

One shortcoming of the suicidal AI is that it doesn't prioritize diplomatic techs, or the most important tech in the game, Xeno Ethics. With Galactic Showcase you never need to worry about being behind in tech, as you can catch up quickly by trading with each race. In fact, even if you are doing the least research you can easily be the most technologically advanced (unfortunately this results in a lower score than researching them yourself)

Then, the secret is to play the AIs against each other using your diplomatic advantage. Always get the AIs to fight amongst themselves. It works really well when you have the weaker races to gang up on the strongest race. When an AI becomes weak enough, swoop down for the kill. Rinse and repeat.

I'm not sure if that's cheese or not. Probably half way in between like you said

I tried suicidal, just for kicks the other day. WOW It looked like rainbow(tact view) of high speed colony ships going through my space. I got thirteen planets while the lowest AI had 28, and the Yor had 56.

If that's the case, you're better off to restart. Given the huge advantages suicidal AIs get you can't let them have that huge of an advantage. Play a race that gets Ion engines from the get-go and immediately make several new colony ship designs. Make some with higher range, others with higher speed, and produce the mix that you need. Skip lower PQ planets in your territory as you can flip them and the AI does a better job developing them. Buy as many colony ships as you can, even if it might mean paying interest. Pray to god that you get some +2500BC anomalies so you can buy more colony ships. Keep your moral at 100% to get double reproduction rate. Perhaps try a smaller galaxy; it will be easier to colonize, but the AI will be able to handle the open travel system better. If you get a crappy start, ctrl-n is your friend.

Anyway, the secret is, IMHO, diplomacy.

Careful here, you're asking Brad to nerf it.   

Xeno Ethics. With Galactic Showcase you never need to worry about being behind in tech, as you can catch up quickly by trading with each race.

Good advice, though I've always played with tech trading off. But I do go for Xeno Ethics ASAP, whether I'm evil or neutral. I want to get the alignment bonuses ASAP. I've never been good, never seen the benefit, but that's a different discussion.

(unfortunately this results in a lower score than researching them yourself)

From what I have seen this is not true. Research score is based on research production not total amount of techs researched. There appears to be no penality score wise for tech trading.

From what I have seen this is not true. Research score is based on research production not total amount of techs researched. There appears to be no penality score wise for tech trading.

Sorry, you misunderstood me. What I meant was that was that you get less points by relying on the AI to trade tech with you than when you build enough research infrastructure to do it yourself.

Careful here, you're asking Brad to nerf it.

Yeah, I thought about that. Oh well...

Sorry, you misunderstood me. What I meant was that was that you get less points by relying on the AI to trade tech with you than when you build enough research infrastructure to do it yourself.

Ahh. Well then. Correct indeed.

After reading the title of this thread over a few days (contents, too, but it's the title that repeatedly gets my attention) and seeing the recent spate of really good ideas being put forward, I can't help seeing a picture of the development team in my head.

It's like the scene in Animal House where D-Day and Bluto are digging through the trash to find the answers to an upcoming test. Then the camers cuts to the two Omega guys who are watching all of this unseen, smoking a pipe and nodding knowingly to each other.

...Not in a bad way, like they're trying to get us into double secret probation - just in amusement. The developers have been very responsive to the forum community over time, and some ideas that started here have made their way into the game. I'm really looking forward to the next DA code release, whatever changes is may bring.

With Galactic Showcase you never need to worry about being behind in tech, as you can catch up quickly by trading with each race

Sorry, I meant Galactic Bazaar (always get those two names mixed up) but showcase is really good too

Research score is based on research production not total amount of techs researched.

IF that is the case,,does that mean you can pad the score by keeping the research facilities after the techtreeis all done?

If not,,that would indicate a finite number of points to research,,which in turn would lead to techtrading lowering the total amount.

IF that is the case,,does that mean you can pad the score by keeping the research facilities after the techtreeis all done?

Yes, though the amount the score is "padded" is pretty underwhelming. In my last game I developed a research production in excess of 100,000 RP/week and maintained this for a number of game years. Still, tech was by far the lowest of the four components of my score. I think I would have been far better off replacing my research with production to build ships or even just stock markets for income.

I'm getting to a point where the research tree will be complete.

Should I replace them?

I'm getting to a point where the research tree will be complete. Should I replace them?

Like Mumblefratz said, the numbers are underwhelming. I eliminate all research facilities, usually only keeping the ones built on bonus tiles (just a habit more than anything since they're still "almost" useless to me at that point). Opens up other more favorable options to shape a planet and make it more "efficient".  
And that short people crack Wheeloffire...  

I would not replace existing ones. I usually build research on my planets and those of the first race I defeat. On all the rest they get converted to SMs.

And that short people crack Wheeloffire...

It's a song from the '70s, I think.

It's a song from the '70s, I think

Randy Newman. BTW we're virtual. No short people here. We're all giants. Last time I looked my feet still reached the ground.   

I usually build research on my planets and those of the first race I defeat. On all the rest they get converted to SMs.

In my gigantic galaxy games, that's what I do, too; except I go for 3-4 industrial complexes mixed in with the econ improvements so that I can crank out ships all over the galaxy.

I can't tell if keeping your research facilities going once the tech tree is finished helps your score or not. For my part, it's just too much of a bother to go back and convert all of those buildings, so I just reduced research spending to 0 with the slider.

I've mostly played on medium lately because I just don't get much time to play. On the smaller sizes, of course, you don't get time to worry about finishing the tech tree before you finish off your opponents.

It's a song from the '70s, I think.

Yes, yes it is unfortunately.