1974

I put the blame entirely on Soren Johnson. We need dumber game designers to design games for people like us, the common clay of the new West.

1975

Oh, i really want to see an optimal tile game, please post a screen shot when you’re done.

Honestly, Tom is right don’t sweat it, the AI doesn’t. The battlefield AI is pretty respectably, the empire build AI pretty poor.
You can also buy tiles with Landowner family, and I very seldom bother. The general strategy for border expansion is you put hamlet two squares away from luxury resource, then you develop the resource and put a rural specialist which further pushes your border rinse and repeat.

You also can expand borders a lot by capturing enemy cities :-).

I don’t know he went to a junior university on a farm, that seems fairly plebian to me. (Sharpe will understand the shade).

1977

Let me put it in terms that I think you would understand:

Imagine if someone designed an ingenious boardgame. Now imagine playing it with three morons who are terrible at it. They don’t understand many of the rules and they have no idea how to win. Is it still an amazing boardgame?

-Tom

1979

Maybe I’m thinking of something different but I get events in multiplayer? I haven’t played for a few months so it’s possible it changed. I definitely got the event popups, had to make decisions on them, etc.

1981

How does that affect the design? I thought you would understand the boardgame analogy – a boardgame’s design is independent of the people playing it – but I guess that didn’t work.

My point is that you can look at a design independent of whether it has good AI. If you’re a student of game design, if you enjoy learning clever systems, if you want to experience the work of a designer who questions the assumptions we’ve had for so long in other games, if you appreciate alternative solutions to problems inherent in the genre, Old World is well worth playing. It’s an amazing piece of work.

That has bad AI.

Those aren’t the contradictory statements you seem to think they are.

This is not correct. In fact, I’m not even sure what you’re trying to hypothesize about. What makes you think there are no events in multiplayer?

-Tom

1983

So if you play Justaguy2’s Favorite Boardgame Of All Time with the Three Stooges, it’s no longer a good design. Got it.

What a weird way to evaluate games. And what a complete non-sequitur to what was being discussed in the thread.

-Tom

1985

So the biggest positive to buying tiles for me is not so much to go crazy optimizing. I got a lot of that out of my system in that Babylon game.

Instead, going forward it’s a great safety valve for me. I used to feel constrained that I could “ruin” a city with non-optimal layout but being able to buy tiles if needed fixes that.

1986

A similar conceptual hurdle I had to clear while I was learning the game was thinking that urban improvements could only be built on urban tiles. I didn’t realize that an urban improvement actually converted the tile to an urban tile. So I thought each city site had only four or five spaces for urban improvements. How’s that for constraint?

So, yeah, I understand what you must be feeling now!

-Tom

1987

The way you can acquire tiles is a nice balance to the fixed city sites, I really like it. You’re stuck with the city sites generated on the map but you get away from that “argh, just ONE tile to the west and I would have access to those resources” situations.

1988

Exactly.

1989

There are events in multiplayer. I think the only events that are deleted, are the ones where you have to choose between going to war with country A or B, unless you have ruler with certain skllls. The only problem with multiplayer is it takes a long time. That and Tom Chick, and CraigM ganging up on me and trying a take my 2nd city, the bastards.

I’m really puzzled by your trashing of the game. My impression is you haven’t played it, is that right? It sort of reminds me of people, who trash Teslas but haven’t driven them.

I understand for a lot of smart people, this game has a high learning curve. It wasn’t as hard for me as plenty of other games, despite the non-existent documentation other than tooltips, in early access.

I now understand that plenty of people really want to understand every aspect of the game before the really play it. I’m more of an experimental player, and I’m also perfectly happened to go back and reload10 turns once I’ve figured out a system works.

I think OW, is very much a single player game. I’ll disagree with Tom on the quality of the AI. Is it up to par with a smart human, no? But it provides a challenging opponent for me, and I’m not even at the highest levels. I really like how Soren (and team) has provide the player plenty of tools to customize their experience. For me I turn off the default put me in the center of the map. This gives me some room to expand without having to take out the other Nations early on. I also turned on the play to win option, which means the AI players will declare war as I get close to victory. This makes the end game more interesting for me, but for lots of other players hate the AI ganging up them in the end.

I think the point Tom is trying to make is that is very thoughtfully designed game. It solves many of the major problems with 4x games in a very elegant fashion. Soren has provided some wonderful writeups of thought process behind these tradeoff.

In the same way you can play chess without understanding the concept of pins or forks. You can play OW without understanding adjacency. At some point, I think the designer would want you to read the tooltips or manual, and understand you get bonus, building farms next to each other. But you can win the game by simply building farms if you need food, quarries if you need stone etc, with no regard to where you place them.

Likewise, there are multiple ways to expand your borders, many of which will come naturally. But to play well it helps to understand all the system. My knowledge of chess strategies stops at hold the center, and try to find pins and forks. My friends who are chess masters have read lengthy tomes, that go into all kinds of arcane strategies. I don’t need to know about them to appreciate that chess is an elegantly designed game.

1990

I’ve gotten that event in multiplayer but it’s with regards to AI players. Wouldn’t make sense to fire between just human players, I suppose.

1991

I know you’re a pretty strong critic of the AI and I understand where you are coming from. Out of curiosity, what are some 4X games that you feel have notably superior AI to Old World?

1992

I feel like that’s not a very helpful way to think of a good AI. Because if it’s just a matter of making it challenging, any bad AI will be good if you give it enough of an advantage.

My problem with the AI is that it’s not using certain systems at all – siege engines and World Religions, specifically – and it’s dismayingly suboptimal with other systems, including one of the absolute foundations of the design: city development. None of that has any bearing on whether the game is challenging, because I can just compensate for its shortcomings by giving it advantages elsewhere. But I don’t consider that a good AI. It’s really disheartening to watch it fumble around. It’s really disheartening when I win situations I should have lost.

In the interest of full disclosure, I haven’t played single-player for about three patches, and they all have the vague but tantalizing “AI improvements” listed in the patch notes. I have no idea what these improvements are, but I remain hopeful and it’s entirely possible my observations are outdated.

I like this analogy because I have no idea what pins and forks are! The chess analogy that comes to mind for me – and this is getting back to AI instead of the documentation issue – is playing against someone who doesn’t know how to use the knights and bishops. Sure, you can still play with that person. But it would be a deeply compromised version of chess.

Civ IV, Planetfall, Imperialism, Master of Orion, and GalCiv, off the top of my head.

It’s @justaguy2’s schtick. “Your favorite band sucks” is his usual mode of online conversation. Sometimes offline, too!

-Tom

1994

I think I’ve used a siege engines in two games. I think they take too many orders to be useful except in specific situations. I’ve definitely won games where I only built shrines. Don’t get me wrong the AI development of their land is pretty poor.
However, of the games you listed. I think only GalCiv (I haven’t played it in ages) did a a decent job as a builder. Perhaps Imperialism but again it is been too long, and I suspect we weren’t as savvy as to what the AI really did back then. Civ IV, MOO, Planetfall none of them built up their empires particularly smartly. Actually, let me defer to @SorenJohnson on the Civ IV empire building. I remember capturing Mineral rich worlds in MOO, and found them filled with Research labs, and artifact worlds with lots of mining equipment. I suspect the AI just built what it needed.

What all of those games had is pretty competent battlefield AI that was aggressive and would attack you and could conquer your worlds/cities unless you had decent size military. Unlike say Civ VI, which you could defend your nation with 2 or 3 military units.

If we treat the AI empire as a black box, which produces military units, diplomatic pressure, some type of economic pressure, like grabbing wonders. The bigger the box the more stuff it puts out.

I’d argue what’s important is the quantity of stuff the box puts out. Not necessarily how it does it. After all the religious subsystem is just a means to put out research, civics, money, culture, and happiness. It consumes resources, growth, orders, stone, wood, food etc. There are alternative methods of producing all these outputs with the same inputs, archives, marketplaces, courthouses, theaters etc.
Now my gut tells me that religion is a relatively efficient way of obtaining these outputs. But I have not done an extensive analysis. It’s possible that when we factor in opportunity cost, maybe the AI is being smart by ignoring religion and siege engines. I don’t know. Perhaps, it is devious plot by the developers to make the AI stronger, by having us humans build dumb things :-).

1995

Man, I don’t know I can agree with this, when the AI (especially diplomatic) is one of the things that ultimately makes Civ IV unplayable for me now. Just deeply unsatisfying to interact with.

It’s been a hot minute since I played, but I don’t remember it being particularly adept. Playable, sure, but I don’t know how well it did at things like ship design, force composition, etc. compared to the player. Certainly serviceable, but a certain level of fait accomplis did exist as well.

Now, granted, I haven’t really dug into the game deeply yet. The multiplayer is most of my experience. I intend to dive in more, but have been occupied with other things (like the end of summer and making sure to maximize outdoor adventures before it rains every day for 6 months).

1996

Imperialism II absolutely has an exceptionally competent AI that plays more or less by the rules. You will see their empire start to show cracks as they are denied resources or run out of money, where in many games they’d just cheat up whatever they need.

1997

Boy, I disagree with MOO / MOO2 having good AI. Witness the unending streams of tiny outdated fleets shuttling between systems in 1, and planetary development in 2 that can’t keep any kind of pace with the player as it fails to prioritize the multiplicative improvements that are the absolute core of the econ design.

And the ship design/building in 2 is just woeful.

1998

I believe Tom is defining good AI as AI that competently uses the systems and mechanics of the game, not just the results it achieves. I think the key thing here is that Tom strongly desires a sense of opposition in games, ideally a feeling that the opponent is being aggressive and dynamic in pursuing victory. Tom wants to feel the AI is playing the game not just being a faceless pump of opposing units.

I’m personally ok with a more passive style of challenge but that is a preference.

PS I believe that Tom probably shuddered deep in his gamer soul when he read that black box analogy.

1999

I realize you didn’t mean it this way, but I should have known that this question would turn the thread into people claiming that game x or game y does not, in fact, have good AI. To those people, I would remind them of the question: “What are some 4X games that you feel have notably superior AI to Old World?”

If you’d like to make a case that the AI in Old World is better than game x, I’m all ears.

(As for Master of Orion, I mentioned it because I’ve been playing a ton of Remnants of the Precursors, a MOO remake. But it’s gotten a lot of work, including AI improvements, so I should have said Remnants instead of Master of Orion.)

Yet the AI builds a ton of them, spends orders moving them around the map, and repeatedly fails to use them in combat. This doesn’t stop being a problem because you feel siege weapons aren’t widely useful.

(Which I strongly disagree with, by the way! You might be confusing Assyria’s unique units with what I’m talking about, which are the onagers and ballistas, and their upgraded versions. Those things are beasts in pretty much all situations.)

If all that matters is the end result, why bother with rules at all? Why not just generate the desired quantity of cities and units each turn? Why bother with a rule set that supposedly all the players use?

I agree with the principle that the output matters more than the process, but it’s not a one-size-fits-all solution. Most 4Xs – Old World especially – are rule sets that unite all the players as they vie against each other. That’s the appeal of these games. Otherwise, you might as well just play a solitaire city builder.

So to go back to the earlier chess analogy, when some of the players can’t use their knights and bishops, your solution is to just let them magically drop a knight or bishop at some random place on the table that might capture one of your pieces. At which point, I would argue you’re not playing chess anymore. You’re playing something else entirely, and while I’m sure it would be challenging and even fun for lots of folks, it wouldn’t be chess.

-Tom