Tobold's Blog
Wednesday, April 17, 2024
I lied about Manor Lords

Two posts ago I said: "I won't be getting Manor Lords on April 26th". Today I'd say that this statement is probably not correct. Unless for some reason the 26th is extremely busy for me with something else, I might very well download and play Manor Lords that day. Sorry!

What changed? Well, it is a typical example of customer price sensitivity. They haven't announced a final price for Manor Lords, but it is rumored to be somewhere in the $30 - $40 range. And when I said I wouldn't play it, I meant that I had looked into the game, and decided that at this point in time it isn't worth $30 - $40 to me. So I won't buy it on Steam.

What I had totally missed was that Manor Lords is going to be available on the PC Game Pass from day one. Which I am subscribed to already. So while Manor Lords isn't technically "free", I will be able to download and play it on the 26th at no additional cost to me. That changes the situation. My main worries about Manor Lords are that the game is far from being finished, and I doubt it already has a lot of replayability when considering the large missing elements. I still consider it likely that I'll play Manor Lords once, and then leave it be until the game is finished, which might be years from now. But while playing it once for $30 didn't look like a winning proposition, playing the early access for free seems a lot better. Even if the game is taken out of the PC Game Pass library later, before it is finished, I'll have a much better feeling whether I'd want to buy it later. It's like playing the free Millennia demo at Steam Next Fest, which ultimately convinced me to buy the game, flawed as it still is.

Tuesday, April 16, 2024
Proof of originality in science

Over the last decade or so, media have been full of stories of scientific fraud and plagiarism. And a rather cynical thought crossed my mind: I don't believe that scientific fraud and plagiarism is on the rise; rather we have simply gotten better at detecting it.

I might be in a rather privileged position here: I have written a Ph.D. thesis and a number of scientific papers, but unlike most people, I can prove the originality. That is because I made my Ph.D. is synthetic chemistry; which means that my thesis and a number of my papers contain a description of how I synthesized new molecules, unknown to mankind before. And whenever somebody produces a new molecule, that molecule is officially registered and receives a CAS Registry Number. As these numbers are uniquely linked to the structure of the molecule, and there is a record of who published them first, I have pretty solid proof that I didn't just invent or copy my scientific writing. I can prove that I made that molecule and wrote about it first, and it is possible to verify that my described synthesis method results in that molecule.

In most other scientific fields, the principal product is ideas, not molecules. And while you can give a molecule a unique identification number, you can't do that with an idea. For a very long time that created a sort of stalemate: It was impossible for a scientist to prove that an idea he wrote was original, but it was equally hard for anybody else to prove plagiarism, unless the scientist plagiarized a well-known source. And a number of scientists wrote Ph.D. theses with the certitude that once they got through the process that awarded them their degree, nobody would ever read that thesis again. (I'm pretty sure nobody ever read my thesis since). So they took shortcuts. From a purely human point of view, I understand that: If you studied something in the liberal arts, coming up with an idea which is both brilliant and never thought of on a subject matter that has been studied for a long time isn't easy. It is a lot easier to make a molecule that nobody has done before than to write an interpretation of a work of Shakespeare that nobody has thought of before.

Some of the people who plagiarized then went on and made a career in the public view. They became politicians, or senior administrators for large academic institutions. And then, new technology evolved that allowed computers to "read" and analyze huge quantities of text. If you disliked a certain politician or senior administrator of a large academic institution, you could now do something which would have been impossible before: Take their thesis, which is a public document, and check it paragraph by paragraph against every other document in the public domain. Take their scientific papers, and check every picture in it against every other document in the public domain. That is still a lot of work, and nobody would dream to apply this to everybody alive who has ever published a thesis or scientific paper. But the practice of plagiarism apparently was always common enough that a plagiarism search against a specific target was at least worth a try. Do a Google search for "resigns plagiarism" and you'll get a long list of stories of various high-placed public figures having to resign due to their plagiarism having been detected. It is a pretty good method of political character assassination when it works; most people are ready to believe that somebody who cheated on his thesis must be a dishonest person. Very few of the stories even mention the accuser, or his ulterior motives in having looked for plagiarism.

While I understand the academic pressure of "publish or perish", I have a lot less understanding for the numerous scientists that have been found to fabricate false results, complete with photoshopped "proof". As Newton said, science sees further by standing on the shoulder of giants, by using established knowledge as a basis for further breakthroughs. People publishing fake results, especially in the medical field, risk wasting precious research time of other researchers trying to turn a medical result into a cure. Even worse are the fields of sociology and political science, where authors have been found to publish fake results in defense of their ideology. But that is a whole other can of worms, where a publication of a real result that contradicts current ideology is probably a higher risk for your career than publishing a fake result that people are comfortable with. Again I am not certain that scientific fraud is on the rise, or whether a simultaneous increase in people wanting to hurt other academics combined with better methods of AI detection just results in more stories being in the media today.

Monday, April 15, 2024
Manor Lords hype

Manor Lords is the most wishlisted game on Steam, with over 2 million people. It will be released into early access on April 26th, and this weekend a large number of streamers were allowed to show the current version. Manor Lords is a medieval village / city builder, with an RTS combat part that resembles a small scale Total War. It is the first game of a small indie developer (which started out with a single developer), and the expectations are high, as early versions of the game are extremely pretty. By having some variation of looks for identical buildings, and by allowing for curved roads and irregular shaped plots of land, the resulting villages look very organic. They can be zoomed into very far, and even walked through. It is all very promising.

However, the obvious promise lets people forget the even more obvious limitations. The general idea is that the game plays on a map with several regions, and these regions are populated by AI opponents, who can war each other. But in the early access version, the AI is still missing. You can only have an "external AI" opponent, who lives outside the map, but sends troops to claim regions, so that you can fight him. There are no AI villages being built anywhere yet. The diplomatic system is basically non-existing. And while you can already build nice little villages, there are features missing that would allow you to more easily manage larger villages and towns. You can certainly have fun in early access Manor Lords for some hours creating a pretty village and defending it against bandits and the "external AI". But right now there isn't much replayability.

I won't be getting Manor Lords on April 26th. The parts of the game that are finished are pretty, but offer only gameplay that a lot of other similar games already have. The innovative elements of turning your village into a barony at war with others are hinted at, but not yet implemented. Now of course it is totally possible that 2 million people buy the early access version, that the indie developer with the help of the more experienced publisher manages to use the money well to finish the game, and that in a year or so we get an absolutely brilliant and complete game. But there is absolutely no guarantee for that, and there have been numerous examples of an inflow of cash only leading to mission creep and bad project management. Having actual players means needing to deal with bugs in a timely fashion, and a level of support that can take away from development time. Development time is always longer than players hope, and might take years before a release version, which still might not be 100% of what is promised now. There is even a chance that the hype turns sour after people played the early access version for a while and notice how much stuff there is still missing. This is the quintessential early access game, a crowdfunding opportunity for a game with promise. People who expect more are likely to get disappointed.

Friday, April 12, 2024
What do I owe you? Or you me?

I now have both a YouTube Premium and a Twitch Turbo subscription. In both, I am basically paying to watch ad free, and the revenue from the subscriptions is divided between the platform, and the channels I watch. Thus I rarely feel the need to pay for an added subscription to any channel. Once a month I can hand out my free Prime subscription on Twitch. And Twitch has a strange culture of gift subs, where sometimes I get gifted a subscription. I even have been gifted a sub to a channel I don't remember ever watching. I might have been on that channel for a few seconds due to a raid without remembering it, and then somehow the random algorithm selecting the gift sub subscription beneficiaries landed on me.

The big question here is in an ecosystem of millions of content creators and billions of viewers, who is owing who what? Should the content creators whose channels I watch be happy about the revenue they get from my Premium / Turbo subscriptions? Or should they consider me a freeloader and I owe them a monthly subscription?

As a content creator myself, I never felt that any of my readers owe me anything. I appreciate your readership, and I appreciate comments even more. Having my voice heard is a reward on its own. But I can't help the feeling that my attitude is as outdated as the medium I use to create my content. I remember my surprise when I first heard a streamer saying that he was doing that as a full-time job, that was a concept unknown to bloggers.

Besides entertainment, I also get most of my news from the internet. And I have noticed an increasing trend towards paywalls here. I hate those! I use the internet to have access to a wide variety of different news sources, because you need to hear the same story from different angles to understand the biases. If I gather my news from 20+ different sources, I certainly can't afford a monthly subscription for each of them. But if I subscribe to one, I basically would be stuck to that one source, and wouldn't get any different views. The result is that I simply don't read certain news sources anymore, not willing to pay them a monthly subscription.

We live in a world where even your door bell might ask you for a monthly subscription. While each subscription sounds like a minor sum, if you subscribe to everything out there you can spend a fortune, without actually realizing where your money is going. I am not poor, but I can't afford to subscribe to every streamer I watch, every news source I read, plus paying monthly for internet access, streaming services, and the Game Pass. There are too many sources, and as a consequence I don't spend enough time with each source to justify a subscription.

Thursday, April 11, 2024
Done for now with Millennia

It is A.D. 1909, turn 350, and my colony ship just blasted into space, winning me the game. I played through all ten ages, choosing the standard version of the age each time, and the age of departure as the last victory age. Most of the game was fun, but the age of departure is about collecting 5,000 points for your colony ship with production, which mainly consists of pressing the end turn button. I don't know how long this run took, but overall I have now 88 hours played in Millennia, so I am well over my 1 hour per dollar spent requirement. I think it is time to give the game a rest, and hope for some patches, plus the DLCs announced for the second half of the year.

One lesson that I applied successfully in this run was that you can't play too peacefully. The AI is grabbing land quite aggressively, and as long as he thinks he is stronger than you, the AI will bully you. Let the AI start a war, conquer a few regions from him, offer peace, and you'll have a much nicer neighbor for the rest of the game. I didn't take any warfare national spirits, but I did keep my military research up, and a good number of troops around. That is also very helpful in the otherwise quite annoying age VII, the age of revolution, where neutral rebels suddenly pop up everywhere to steal your cities.

The one feature promised for the next patch that I am looking forward to is the ability to raze a vassal city. Right now you can only raze neutral cities, but if you take a city from the AI, it becomes your vassal, and you can't get rid of the region, even if you could use the space better for your existing regions. I have tried many different runs and different strategies, but a lack of space is a constant. That is really annoying, because the whole city-building aspect of Millennia with its goods chains is somewhat falling flat if you don't have the space to build it.

While I had a lot of fun with Millennia, it ultimately doesn't overcome the problems that other historical 4X games have. The start of the game is fun, but at some point your territory is established, and the whole thing kind of bogs down. Then you can either go for a military victory or peaceful victory, but the end game is always a lot slower and more cumbersome.

Monday, April 08, 2024
Games at release

Once upon a time, this was "Tobold's MMORPG Blog". So I played a lot of MMORPGs. Not only did I play some MMORPGs for thousands of hours, I also played a great number of different MMORPGs to try them out. And among the community of people who played lots of MMORPGs on release, there was a running gag: How bad is release day going to get? Because absolutely every MMORPG release had a bad launch. In some cases it was just login queues and some lag. In other cases the games were completely unplayable in the first week.

Fast forward to 2024, and the discussions I see about the releases of games have very much changed. There is an expectation of perfect launches now. If the latest released game has less than 60 frames per second on older computers, there is a huge outcry, and the game will be review bombed. If the game developers or publishers announce any sort of further monetization, be it DLCs or micro-transactions, the game will be review bombed. If the game isn't perfectly balanced and polished, it will be review bombed.

The modern media culture of the internet and social media, where outrage means clicks, which means ad revenue, is in part to blame for that. And the result is a lot of fake news. If you followed the gaming news, you'll be aware that Capcom's latest release, Dragon's Dogma 2, was review bombed for various heinous crimes, from performance issues, to microtransactions, to bad game design decisions. But if you scratched beyond the surface, you found that some of the outrage-creating headlines were simply not factual. And the Steam reviews follow a now increasingly common trajectory: Mostly negative at launch, when all the outraged people review bomb the game, and then the review score constantly increases over time, when people who actually played the game give their opinion. Dragon's Dogma 2 had 228,585 players concurrently at launch, and the numbers aren't declining particularly fast. That is not a bad launch. For most players the performance of the game isn't game-breaking, the micro-transactions turned out to be completely unnecessary to play the game, and the weirdest bad game design decisions have been fixed in the first major patch.

The danger here is that between the outrage media reporting and the review bombs it becomes very hard to distinguish truly unplayable game from games that are actually rather good, but have a few flaws. It seems as if the whole world has become increasingly unable of differentiate between shades of grey, everything needs to be either black or white.

Saturday, April 06, 2024
An achievement, kind of

I don't generally care about achievements in games. But yesterday I couldn't help but notice that I got a Steam achievement for Millennia, that only 0.1% of players have reached: Age of Old Ones. Which is hard to reach, because you need to play incredibly badly to get there. In the Age of Renaissance you need to get to a negative culture income, which is easiest achieved by integrating a lot of vassal cities. Once you have played 20 turns with negative culture, your next age is locked to Age of Heresy. Of course, like always, you need to stay ahead of the AI opponents to reach the age first, which is why I played on medium difficulty.

In the Age of Heresy you must then deliberately destroy your improvements, so as to *not* fulfill your regions needs. That spawns cultists and cultist leaders. Once you reach 30 cultists in your region, the previously invisible Age of Old Ones locks as a new crisis age. Unusual, because normally you can't have a crisis age after another crisis age.

The Age of Old Ones is also a victory age: When you are the one to trigger it and reach it, all these cultists in your regions and in all the AI nations' regions come under your control. Suddenly you have a huge army, which doesn't need to be paid. So you just need to gather them into stacks, and attack everything. A bit tedious, but funny to do once. Although I am not convinced that this was game development time well spent, as very few people will ever see this. Until you read up on how to reach the Age of Old Ones, you are extremely unlikely to stumble into it by chance.

Thursday, April 04, 2024
The Run

I have played Millennia for over 50 hours now, but I haven't finished a single game yet. The furthest I got was to age 8 out of 10. I didn't lose any of these games, but decided they weren't worth finishing. Basically I am trying to achieve "the run", the perfectly balanced game, that is fun until the end.

One problem here is the conflict between player agency and challenge. If you set the AI to a higher level, the game will be more challenging. Mostly because the AI just cheats more, not that it plays much better, but still, more challenging can be good. However, if you set the AI to something that is challenging to you, chances are that it will be one of the AI nations that reaches the next age first. Which means you are losing player agency, because it isn't you who is selecting whether you'll play a normal age, a crisis age, an alternative history age, or a victory age. In my current run, AI Russia caused an age of plague to appear. That was interesting, because I hadn't played it yet, but definitely messed up my plans for the game.

The same problem appears on the rather basic decision of whether you would like to play a game of peaceful expansion, or a game of military conquest. The diplomatic system of Millennia is rather primitive, and the AI decides whether it wants peace with you based on whether they feel stronger or weaker than you. Thus if you set the AI to strong, they'll constantly decide to declare war on you, and none of the options in the diplomatic system seem to be able to change much about it.

Much of my restarting was done to modify my distance to the enemies. A typical mistake is to start a game with the default settings of continents, medium, 8 players. The continents map always has 2 continents, and about 40% of the map is water between those continents. On a medium map, this much water leaves relatively little dry land if you divide it by 8. So the AI, which is good at aggressive land grabs, will seriously hem you in. Even if you go warfare, and take more territory, you can't currently raze regions conquered from the AI, which means that you are stuck with a network of regions far too close to each other. The medium pangea map would already leave more space for 8 players, due to fewer water tiles. But you could also modify the map size to large, or reduce the player number to 6. In extremis you could do a medium continents map with just 2 players, which would give one continent to each of them, and no interaction until deep water travel is researched mid-game. Of course giving yourself too much space, on a larger map with fewer players, also takes away a lot of interaction with the AI. Thus one needs to experiment a bit which is the best compromise between space and interaction.

To make matters even more complicated, the devs released a "patch zero" yesterday, but the balance changes in that patch only apply to new games. So again I am not certain whether I will finish my current game or start over once again.

Tuesday, April 02, 2024
AI Luddites

I was reading a review of a game on Board Game Quest, when I stumbled upon the following passage: "Beyond current commitments, Board Game Quest is not planning on covering games in the future that use AI art". This was because the game reviewed did contain AI art, and the editor was uncomfortable with the "ethical issues" of that. I think this approach is backwards and far too general.

AI generated images, in my mind, can have two big ethical issues. One is if you ask the AI software to generate an image "in the style of" some known artist. That will produce an image which is obviously mostly based on the works of this one artist, and will be similar to his existing artwork as to be able to compete with it. For example Phil Foglio and his wife produced a lot of very distinctive artwork for Magic the Gathering; it would be unethical for WotC/Hasbro to make new Magic cards using artwork "in the style of Phil Foglio" created by AI, and not pay Phil for it.

The other ethical problem of AI images is deepfakes, like the Taylor Swift deepfake pornography from earlier this year. Besides the aspect that the creator of such an image is trying to deceive others into believing something, which in reality never happened, there is also the aspect of the deepfake using images of a specific person; that person has certain rights regarding his images, which the AI image violates.

But if you ask an AI for a generic image with a generic description, e.g. "draw me a goblin pirate", I don't see the ethical problem. The resulting image is a composite of many different works from many different artists; it isn't fundamentally different from an art student painting an image, inspired by whatever artwork he studied during his courses. If we talk about an art movement like impressionism, that is because the artists in that movement were inspired by each others works. There are obvious similarities between the various paintings of the same art movement. Nobody ever questioned whether that was ethical, or demanded copyright compensation for the artist who inspired his fellows.

Indie game developers, whether for board games or computer games, most of the time aren't great artists themselves. Their interest is in game design. Why would we want to force the game developer to draw that goblin pirate himself, badly, or to pay an artist to draw a goblin pirate, increasing the cost of the game, when instead he can get a perfectly viable goblin pirate for free from an AI? As long as the image isn't copying anyone in particular, but is just generic, I don't see any ethical issues here. I don't think that artists have a right to be employed any more than somebody from all the previous professions that have been made more or less obsolete by new technology. I might as well complain that the board game uses printed components instead of hiring a monk to draw every card by hand.

I believe that AI software should block requests to draw images "in the style of" any specific artist, and not accept names or photos of a specific person to base its AI images on. But companies that want to go for a distinctive art look for their game will always hire human artists. AI is a perfectly viable solution for creating generic images for games.

Sunday, March 31, 2024
A winning strategy for Millennia

Millennia can be played with a wide range of different strategies, which work best when taking into account the geographic conditions of the map you are playing on. But in this post I would like to describe a relatively simple strategy based on military expansion, which should work on most maps. If you find Millennia daunting in its complexity and don't know how to do better than the AI opponents, this is one possible way.

We start our winning strategy by selecting the Homeland: +Culture perk on the setup screen for a new game. That means that at the start of the game, you gain 3 culture each turn instead of 2. As you need 8 culture for the first culture power, it only takes you 3 turns to get there. As first research we select Tribal Elders, and we build a scout cavalry in our first city. We use our two warbands to scout in different directions, because finding Tribal Camps would help, but don't send these units too far off the starting city, in case we need them to defend against barbarians later.

In turn 4 our culture power is ready. We use it for Local Reforms on our city. That increases our culture gain per turn from 3 to 4.5, and our knowledge gain from 2 to 3. When the Tribal Elders research is finished, we research Scouting, afterwards we research Defenses, and finally we move onto the Age of Bronze. After our first production of a scout in the city is finished, we build the Council, which has been unlocked by the Tribal Elders research. We spend our Government points first on Tribal Farming, then on Oral History, and finally on Tribalism (Reformed). We do *not* save points for a settler. Setting reminders with CTRL-click on these Government perks is helpful. Whenever our culture power is ready again, we keep spamming Local Reforms on our only city, which gets even more efficient with the Council building and Oral History to maximize knowledge production. In the early stages this is far better than the Eureka culture power, and it also buffs all other production values of the city.

This way you should be guaranteed to be the first nation to reach the Age of Bronze even on higher AI difficulties, before turn 20. Now you select Discipline as your first Age of Bronze research. And you choose Raiders as your first national spirit. Within the Raiders national spirit, you are going first for Marauders, then for Outlaws. Every perk in the Raiders national spirit gives you 2 more raider band units. So what you are trying to go for now is a sort of snowball effect: By going after barbarians and their camps first, neutral cities second, and AI opponents third, you are trying to collect a lot Warfare XP through combat; which you then spend on Raiders perks, which gives you more raider band units, and thus more combat. While stacks of raider bands can be useful for their speed, their combat efficiency is better if you mix the raiders with archer units and leader units. I would recommend waiting until you researched Discipline to upgrade your warbands to spear units before promoting them to leaders, as that way you get the much better level 2 leaders.

We deliberately neglected building up our city. Of course after the Council building you should build whatever satisfies the needs of your population and grows your city. At some point during the Age of Bronze you might want to interrupt the Local Reform spamming by using Create Town instead. But generally we concentrate on knowledge and culture, and our military.

We do not want to enter the Age of Iron. What we do want is the Age of Blood, which requires killing 6 non-barbarian units. Typically conquering two or three neutral cities does the trick, but we wouldn't shy away from attacking AI opponents either. If we have already researched 3 technologies before triggering the Age of Blood, we catch up on some other research before advancing to the next age. It is best to not accept diplomatic proposals like open borders, because getting from there to war takes longer. Declaring hostilities early can be an advantage.

By turn 40 at the latest we should have reached the Age of Blood. This is a crisis age. There is going to be a lot of warfare between everybody, and there are more barbarians than before. The warfare will also create chaos points (although we vassalize conquests mostly and raze them only if they are really in the way). The chaos points will create more chaos events. Luckily some of the chaos events will also affect all other nations, and being at war with each other, and all the barbarians will force them to fight. The general idea for the Age of Blood is that we are much better prepared than anybody else. We have the spawn raider warfare power to quickly bolster defenses of our cities. And we don't stop taking more cities, neutral or from AI-nations. We can also stop to spam Local Reforms now. Ironically we can afford a peaceful revolution to change our government to Kingdom, which is probably better here, due to all the vassal cities we have.

Once you gain the Victors power of the Raiders national spirit, your armies heal on attacking. That allows you to play even more aggressively. You continue conquering everything in reach during the Age of Blood, and when it ends, you make peace with everybody. Now from the Age of Kings on, you can play with whatever national spirit and playstyle you want. You might be slightly underdeveloped economically, compared to a more peaceful game, but everybody else will be too. And you should have taken control of a rather large chunk of land, possibly even the whole continent on a medium-sized continents map. That allows you to ultimately build much bigger cities, which will be an advantage in the later game. Compared to a peaceful run, in which the AI nations tend to grab land aggressively and don't leave much for you, you should now be in a much better position for the rest of the game.

Saturday, March 30, 2024
Millennia launch

A week ago I predicted: "I would consider it extremely likely that in a week the Steam concurrent user numbers and user ratings of Millennia can only be described as "disappointing"." I wasn't wrong. Peak Steam concurrent user numbers for Millennia were 8k, Steam user reviews are "mixed" at 65% positive. On Metacritic Millennia has a 67% critics rating, and a very similar 7.3 user rating, but frankly, very few people even bothered to leave a rating. This isn't a great launch. Right now where I write this, there are 13 times more people playing Civ6 than Millennia. I don't think Firaxis is worried.

The only thing surprising to me was how many people mentioned the lack of simultaneous multiplayer, a feature that was promised, but is shown as "coming soon" on the start screen. A game of Millennia takes tens of hours, hundreds of turns, and often the interaction with nations that aren't immediate neighbors is minimal. I'm not sure why anyone would want to play that in multiplayer. For a 4X multiplayer game, I would always prefer Solium Infernum, which is a much shorter game with much stronger interaction between players.

Paradox this week announced that they would end further support of Star Trek: Infinite, a slimmed down version of Stellaris that they released just half a year ago. And their other big release half a year ago, Cities Skylines II, also isn't in a good state yet. It is also true that even the successful Paradox games are often somewhat half-baked on release, and only get really good a few patches and DLCs later. All this is reflected in the Steam reviews, with some variation of "yet another unfinished game launch" being frequently mentioned.

Paradox does have good marketing, especially influencer marketing, and Millennia has a great presence on both YouTube and Twitch. That isn't always working out like intended, as Millennia is a complicated game, not terribly well balanced, and things can easily go wrong. So one major streamer ended up losing his main city in a rebellion caused by a chaos event, and did a mild version of a ragequit. That probably wasn't the advertising that Paradox marketing had imagined.

Personally I am still having fun. Probably because since this year I got into a "emergent storytelling" mode with Paradox games, where things going horribly wrong is part of the fun experience. Maybe you recently looked up what the 3-body problem in physics is, due to a popular Netflix series. Millennia is a bit like that, a rather complex system that can go horribly and chaotically wrong. It is easy to overlook something important, and regret that later. Millennia plays a lot less predictably than Civ6. That can be perceived as a lack of player agency, but it is more a case of the player having agency but missing information or being overwhelmed by too much complexity, and ending up doing bad decisions because of that.

A typical example would be you getting the ability of founding a religion, but unless you have played the game before the consequences of that are obscure, and things can go wrong and lead you right into a nasty crisis. In my current game I managed to get faith more or less under control, which gives a huge culture bonus, but then tried to go for a religion victory and failed to achieve that. I would have needed to combine a military "crusade" approach with my peaceful means of spreading my religion to achieve victory. To me that failure was interesting, but I can see how not everybody will like to play that way.

Friday, March 29, 2024
AI verisimilitude

Verisimilitude is defined as the appearance of being true or real. Something doesn't have to actually *be* real, but at least it should do a good job of appearing that way. Arguably verisimilitude is what AI has been making the biggest progress in last year. People are using ChatGPT to get advice on things, which is kind of crazy, especially for things like stock tips. ChatGPT doesn't have a clue about stocks, nor does anyone really know which stocks are going up or down, but ChatGPT does a good job of giving advice that sounds as if it was given by a human who knows about the subject.

I am currently playing Millennia, and I am having a lot of fun with it. But I agree with one of the frequently cited points of criticism, which is the lack of verisimilitude of the AI opponents. Or as other people express it, the AI opponents have "no character", "no personality". That is a combination of two things: First of all the diplomatic game mechanics in Millennia aren't great to start with; but second the AI seems to decide on its diplomatic stance on a turn-by-turn basis, with absolutely no consistency: The AI might reject your demand for peace one turn, and then offer you peace the next turn. You are also very likely to get a bunch of messages at the end of each turn in the latter game where AI opponents change their diplomatic status with each other very frequently, which looks kind of random.

While the lack of verisimilitude in Millennia is true in isolation, it is made worse by Millennia being roughly in the same genre as, and possibly competing with, Civilization 6. And while certainly not perfect, the diplomacy system in Civ 6 is a lot better, and more importantly for this post appears as more real to the player. In Civ 6 you have nicely animated, somewhat cartoonish civilization leaders, which already improves things by putting a face onto the AI you negotiate with. And the Civ 6 leaders have both a public and a hidden agenda (which you can discover), for example Cleopatra likes civilizations with a high food output. That gives her character a certain consistency during a game, and the player an option to influence diplomacy with her in a foreseeable fashion.

Millennia is sharing a problem here with Humankind: Civilization 6 actually has civilizations that remain through the whole game, Millennia and Humankind have factions which change in each era. One could argue that Millennia would be better if instead of calling the AI opponents by the name of civilization, e.g. Germany, they would just be called "Yellow Faction". It would actually be an improvement if at least Germany with the yellow flag would automatically be assigned the color yellow for its borders; right now border colors aren't linked to flag colors, which can make the game a bit confusing, unless you do like me and during setup take care that the colors match. But the fundamental thing is that if you play against the Germans under Frederick Barbarossa in Civ 6, there will be certain consistencies in patterns of behavior with other games against the same leader; while in Millennia playing against the Germans doesn't tell you anything from one game to the next, as they might be peaceful naturalists in one game, and aggressive raiders in the next.

Theoretically future patches and DLCs could fix Millennia, and create a larger set of AI opponents, which then could actually have preferences for certain national spirits and thus playstyles. Leader pictures, at least in 2D, could be added. Diplomatic stances could be made more "sticky", so as to provide more consistency. But honestly, I doubt that will happen. This doesn't seem to be a core concern of the developers. But it sure is a concern of players, and Millennia's success is suffering because of it (obviously not *only* because of this).


  Powered by Blogger   Free Page Rank Tool