Arathorn's partial report/rant

[Sirian]

Caesar wouldn't trade with me after he declared on me.

Maybe that's part of the explanation. Losing good relations with Caesar would have removed a large chunk of my tech trading.

The AI prefers to attack nearby targets, though occasionally they will go a long way. I don't know if you've noticed in my SGs the amount of times I keep emphasizing "more military". In my game here, I had barracks and four or five veteran units -- plus two workers and two wonders -- before I started to build my first settler. That's extreme only because of the wonders. My second city had a veteran archer and warrior in it and was on a hill, plus I converted JC to my faith with that first free missionary and got a road and some trading and open borders to him very quickly. I never had a window in there where I looked like a juicy target.

By contrast, in games where I have run light on military and stretched my settlers out there, I've been attacked (and honestly, it was correct of the AI to do it). The AI gets itchy over border tensions. If you have border tension penalty, you need a strong military, and even then they may still attack you. It's also about who is the most likely target, so you don't want to end up as low man on somebody's relations totem pole, or you become the "worst enemy", and that can start the dogpile on YOU if you handle it wrongly or catch bad breaks.

I didn't keep exact records of my trades, but I doubt I got more than 5-8 techs total from Mao.

I only got about four techs from Mao. I got about three or four from Money Musa, three or four from Sal, a couple from Alex, and five to eight from Rome. Perhaps it is that I started later, since I did not research Alphabet and did no trading for cheap early techs. I got more total beakers out of tech trades, and got an intensive period where I got three or four techs on one or two turns by trading the same techs to several AIs, thus accelerating myself more than the trades helped them. I went from last to first in tech between 500AD and 1500AD, and I stayed in the tech lead until the very end.

(Contrary to Sulla's notion that the tech leader cannot make use of the Internet, I did by ignoring non-Internet-Prereq techs and HARD, so that even though I was the tech leader, the AIs had researched several techs I skipped).

The system is set up to prevent you from being able to beeline to Alphabet as the Only Right Choice in every damned game you play. You can trade heavily early, but you'd better leverage that in to a winning position because the trade doors will close on you later. Or you can trade later, but you have to be low in score and/or high on relations, or the trade doors will close.

If you're getting five to eight tech trades out of ONE civ and several more out of others, and thinking that's not enough, then we'll just have to disagree on how much tech trading should be going on a game of Civ. There are only about eighty techs on the whole tree, and the first twenty or so are there before tech trading really hits the board (in the normal course of affairs). So out of sixty techs, you traded for a dozen or so and that's not enough?

How many would be enough? Do you believe that tech trading should DOUBLE the actual pace of tech in the game by allowing you to get a tech in trade for each one that you research? I think that would be stupid, to put it bluntly. Once you reach that point, where tech trades are always and readily available, always reliable as the one best path through the game in all situations, then we don't even HAVE a game any more. It's just a puzzle, like Civ3's brokerage maze was. A puzzle with the same easy answers over and over again.

Tech trading in Civ4 is akin to Civ3's singular golden age. You have some control over when to pull the trigger, but the acceleration it offers you will not last the entire game, so it's up to you to make the most of it, one way or another.

There should be some games where the world is peaceful, sure. *Sigh* But why do those always have to be the games I play?

The peaceful thing is real and it's not just you. There's nothing I'm concerned about more. However, I also think you are carrying forward some energy from the first public build, v1.00, and that things have improved since then. Looking at how many players had AI on AI action in this game shows that some progress has occurred.

I would expect another patch at some point, and you need to wipe the slate clean (as best you can) in your mind at that point, and measure from there forward. If the problem persists in an unbalanced way, there's still expansions to try to deal with it. But I must tell you, it's hard work trying to balance something at the macro level like this. You'd be surprised how much goes wrong if the wrong thread is pulled.

JC deciding to use some Arathornian Azteskimos as croutons for his salad seems to have upset the apple cart in your game. There might also be a little bit of bad luck in there, too. After all, the dice do still play some role in who fights whom on which schedules.


- Sirian

[Sirian]

Civ3 diplomatic AI didn't have one set of routines for dealing with the human and another for dealing with another AI...

Maybe your memory or perception of Civ3 differs from mine.


Civ3: AIs don't demand anything of one another. (That hasn't changed).

Civ3: AIs make all diplomatic deals with one another at a discount. (That's gone in Civ4.)

Civ3: AIs hardcoded to sell everything they can, to buy everything they can, and thus if you don't get there first, they LOCK YOU OUT of the old boys' network with their perpetually-renewing gold-per-turn deals. Their cash is locked up in deals with other AIs, and those deals renew between your turns, so that other AIs monopolize the chance to get their business.


Then there's the fact that "fair" can be overrated.

Civ3: AIs hardcoded to accept any request for military aid.

(IF YOU DON'T DO IT TO THEM FIRST, they will always end up ganging up on you in warfare. You should have played Epic Thirty-Six! You'd have a different take on this.)


Civ4 has improved on all of these problems. Greatly improved.

If you look at how Kylearan was instigating wars left and right, yes he was doing it with techs. There is big trouble for the civ who falls out of the tech trading realm, who has no tech to offer anybody. The Civ4 AIs DO have their price for some of these things. Of course, if they are already in a war, they won't want to take on more. If they have buddied up with the target, they won't attack their friend on your say-so.

I know it's not perfect, but it's progress. We can't afford to get stuck on not having everything exactly perfect. That's an elusive ideal anyway, since one man's perfect is another's poison.

Our sample sizes are still too small to render sweeping judgements. I'm always open to looking at evidence. The evidence from this event is pretty favorable on the whole, though. So far, you're the only report who didn't seem to enjoy the game. I guess we'll see what happens in future events.


- Sirian

[Bezhukov]

One thing I'll do when getting Alphabet is to spend a few turns cleaning up the cheap initial techs I missed by researching them myself so I can trade for something good instead of a block of one or two turn techs. I hadn't realized that this preserves trading credits for later.

Once I've gotten a tech lead, or even second place, I'll often trade tech sparingly unless desperate, to preserve diplomatic options, unless I can nab a key tech for expanding tech capacity like currency. The slower the overall tech pace of the game, the more time you have to make up the initial AI advantage. I try to sell tech to the last place AI for some defecit research cash, which over time also changes the last place AI...

[LKendter]

There should be some games where the world is peaceful, sure. *Sigh* But why do those always have to be the games I play? And the world's not entirely peaceful -- the AI will declare on me fairly happily. But I've still seen more AI-decided wars against other AI in my Epic 1 game than all my other games combined.

While there is less # of AI wars in Civ3, I still see a decent number of them. Epic1 saw multiple rounds of AI beating on Ghandi. Washigton ate the Aztecs alive. Several of my other games had wars between AIs.

I think less wars is a good thing, as it eliminates the Civ3 problem of the industrial age arrives and all civs self-destruct fighting each other letting the human player salvage almost any position.

[xenikos]

LKendter -

I'm pretty sure we're not supposed to talk about Adv1 yet....did you mean Epic 1?

[Sirian]

The AI will often red out war declarations on civs you're at peace with. That is, the AI resists being asked to declare war on somebody FOR you, while they are more readily willing to JOIN WITH you. I believe this is more true for some personalities than others.

Sometimes stoking the fires means getting your own hands dirty, too.

And this, too, is more advanced than the Civ3 AI.


- Sirian

[Arathorn]

You may not want to hear this, Arathorn, and this may be a sample size issue, but I think you're falling behind on your openings (that workboat you never built on the Civfanatic game comes to mind).

It’s definitely possible. I didn’t start Adventure 2 with a worker, like so many others did. Was getting 3 early religions worth that? I dunno. I certainly had as much or more land from the landgrab (not necessarily the same land, though) as many others. Too many settlers and not enough military might be true, though. I do think, though, you’re reading a bit much into the first 20 turns of a game where we didn’t even have Sailing to start the game.

I've demanded things of AIs before, and occasionally they've caved in. The -1 hit for a tribute demand is "You made an arrogant demand!" Generally speaking, the AIs will only cave in if they know they'll lose a war against you.

I’ve requested/demanded things of the AI before, too. I just haven’t seen the “occasionally they’ll cave” part. I’m seeing a trend, though, on military. I have enough military that *I* know I can beat the AI, but they probably don’t recognize that. Might be an issue. 3+ units/city is not enough, apparently, especially so when there’s no AI-AI wars to decrease their unit count.

Or you can trade later, but you have to be low in score and/or high on relations, or the trade doors will close.

I was high in relations and dead last in score. One might think that would be enough.

But it still doesn’t address the root of the issue. WHY will the doors close? What gameplay sense does it make? Why would any even semi-intelligent player cease trading with another who’s been a valued trade partner throughout the game? It makes no sense.

then we'll just have to disagree on how much tech trading should be going on a game of Civ

We may well disagree on that. And that’s fundamentally unresolvable. But it’s really outside the scope of (most of my (valid)) complaints.

Putting a brake, of some sort, on tech trading, is perfectly valid. There are just so many ways of doing it that don’t involve counterintuitive, dumb play by an AI that I don’t understand why the “We fear…” thing was ever implemented. It’s such an obviously artificial constraint as to be extremely annoying. Put limits on how many techs you can trade away at a time. Put limits on how often tech trades can be done between two partners. Make tech trading unattractive in some other way (50 gold hit to both civs or something, to pay for “training”). Makes a traded-for tech's benefits (in terms of units/buildings/ensuing techs, etc.) not available for n turns (n=8 sounds fairly reasonable to me, for Normal speed, but it may need tweaking). Dozens of options exist, essentially every single one of which slows down tech trading but doesn’t involve such a ridiculous method of penalizing a good trading partner.

Heck, Mao probably got 12 techs from me. That’s 1/5th of the possible, by your odd math. I count 85 techs and about 8 prereqs to Alphabet. That’s 70+ to trade. The point is NOT the number of trades, though, but how they are limited. The chosen mechanism is about as poor as I could imagine. About the only thing worse would be to have tech trades randomly not work in one direction or the other, in some cases.

However, I also think you are carrying forward some energy from the first public build, v1.00, and that things have improved since then.

Eminently possible. It’s not like I’ve played dozens of games on 1.52 to really know how much that really changed things.

But I must tell you, it's hard work trying to balance something at the macro level like this. You'd be surprised how much goes wrong if the wrong thread is pulled.

Preaching to the choir. I’m not saying AI-AI aggressiveness is an easy fix. I definitely understand some of the issues with that. I’d like there to be more, but…how much more and how do you limit it to reasonable amounts while maintaining the chance of peaceful games and limit the manipulability by the player and…? Tons of issues there. I don’t like the current level much, but I certainly am much more understanding of that than I am some of the diplo issues.

There might also be a little bit of bad luck in there, too.

I had my fair share of that this game, that’s for sure. Reading Snaproll’s report of a longbow surviving multiple attacks from superior units almost made me cry. I couldn’t get an archer (promoted, of course, and fortified) on a hill to survive a single attack from an inferior unit. And bad luck, while it’s statistically guaranteed to happen, is generally no fun to have.

So far, you're the only report who didn't seem to enjoy the game.

Not quite true. Snaproll mentioned extreme frustration and anger. Liquidated didn’t report, as his game was no fun (“That game really depressed me in ways I'll not forget.”). There may be others, as I’ve not read even a significant fraction of the reports yet.

Our sample sizes are still too small to render sweeping judgments.

Bad design is bad design. It’s visible on first viewing. The “Too Advanced” thing is bad design. The AI war rates…well, that’s another kettle of fish, and I agree sample sizes are too small.

Arathorn

[Yosh]

And bad luck, while it’s statistically guaranteed to happen, is generally no fun to have.

I feel your pain, Arathorn. I've talked about this briefly on other boards, and while I think that it's possible I'm paying overly much attention to the unusual results, on the whole, I've noticed that in my games, the "unlikely result" occurs much more often that it "should".

What do I mean by that?

I mean that in a recent game I'm playing, I had 95% chance to pop a Great Prophet, 5% to pop a GA. I popped the GA. Fair enough, let's plug more GPP into GP. Next time comes around, it's 83% for the GP, 17% for the GA. I pop the GA. That's bad luck for sure, and it's an extremely small sample, so there's no real way to truly investigate it.

However, and again, I could just be taking more note of unusual results, I've noticed a LOT (and I mean a LOT) of the situation where I will consistently lose combat where odds are in the 70's-80%. Now, I have no real way of testing this, since I am able to consistently reproduce the problem, but I think I see the gripe about the PRNG.

Now, the Thing to Do, IMHO, would be to find a way to either a) consistently reproduce a situation that displays this error or b) somehow gather enough statistical evidence to support our hypothesis, or some combination of the two. I, for one, kind of enjoy this "bug-hunt", even if it is chasing shadows of our own devising. So if you'd like, I'd be more than happy to work with you on this.

[Sirian]

Putting a brake, of some sort, on tech trading, is perfectly valid. There are just so many ways of doing it that don’t involve counterintuitive, dumb play by an AI...

Tech trading involves "freebies" on the scale of THOUSANDS of units. Civ is a game of details, of working a plot to get ONE extra food or shield per turn, of investing thirty or seventy turns in to cottage maturation, of promotions that add 10% to the Strength value of a military unit, of religion that spreads one city at a time one missionary at a time.

Then along comes tech trading. "Ooh, I've learned Physics. I'm the first to get there. Let's trade it to Civ A for Biology, Civ B for Communism, Civ C for Combustion, and Civs D, E, and F for all their cash, totalling 3000g. I HAVE PICKED UP 20,000 UNITS IN ONE MOVE!"

Um... yay? I don't think so.

All your hard work to pop a Great Person and he can pull 1500 or so units for you, a limited number of times per game. A round of tech trading can pull tens of thousands. There's no strategy to it. There's no challenge to it. There's no THINKING to it. It's a dull and tired mechanism whose day has come and gone.


Because tech trading moves thousands of units while almost everything else in the game moves individual units, it dominates the game. It looms over the entire design, wiping everything else off the board.

Been there, done that. No thank thee.


You can have unlimited tech trading, and that's the whole ball game, or you can have a real strategy game. Take your pick.

If you pick "limits", then you've got no reason to complain. Sorry. A cap is a cap. Limiting tech trading to a quarter of the tree is what it is. You won't make it better by disguising it or making players jump through hoops to utilize it. You won't improve it by forcing a player to dole it out more slowly, so that it trickles in at a constant rate. Better to let the player choose when to engage it. If you are always choosing the maximum amount of early trading and drying up the well, then that's your choice, similar to blowing your golden age under Despotism in Civ3.


I'm reminded of the story about the dog left alone for a week. A couple puts down a week's supply of food for the dog, but comes home after their vacation to find the dog miserable and ravenous. They ask him, "How can you be hungry? We left you a week's supply of food?" He answers, "I have no idea what you're talking about! The only food I found was that really big breakfast on the first morning."


- Sirian

[Bezhukov]

"It’s definitely possible. I didn’t start Adventure 2 with a worker, like so many others did. Was getting 3 early religions worth that? I dunno. I certainly had as much or more land from the landgrab (not necessarily the same land, though) as many others. Too many settlers and not enough military might be true, though. I do think, though, you’re reading a bit much into the first 20 turns of a game where we didn’t even have Sailing to start the game."

Thanks for the smiley. I enjoy debating strategy, and in fact learn myself by doing so, but it can come across the wrong way sometimes (see LK's beating a dead horse icon). Perhaps you can do a similar analysis for opening moves that you've done for combat next? The ability to dedicate food to settler/worker production makes early food subtly but powerfully important. I suspect this may be one area where there is in almost all cases "one right answer".

In my case the capital had four more food from the two camps, with the worker on hand to chop the settler for the second city. I could easily have been as much as 40 or 50 turns ahead on getting the third camp going and third, fourth, and fifth city down (as the capital built settlers/units instead of wonders). With the availability of slavery whips, this would make a big difference.

I'd be interested to see the date at which you got the land from your landgrab, as availability of land is not so much the issue on Highland maps, but quality of land and ability to defend it (as much a question of establishing good relations with other civs before they become surly from overabundance of units, I'm guessing, than one's own military strength. Also, earlier cities have less and easier barbs to fend off).