Alexander “Ail” – Pandora: Complete
Pandora: First Contact came out in November of 2013, almost a full year ahead of its terrestrial sci-fi cousin, Civilization: Beyond Earth. As a successor to the the classic 1999 hit Alpha Centauri, Pandora was not all that far off. The problem was the game had issues; most notably a weak A.I. In came Alexander Stumpp (a.k.a. Ail) – a very active community member and lover of what Pandora was trying to do. After the release of its expansion, Eclipse of the Nashira, the developers had mostly moved on, but Ail took over. Since then, he has made many changes that improved the game’s A.I. to the point where the developers QC his work and make them official patches. So, let’s get to know him a little better and find out what he did.
Question: Can you tell us a little about yourself and your background?
Answer: My name is Alexander Stumpp, and I’m occupied as a software developer in a company that, unfortunately, makes something far less fun and interesting than games.
Question: What was your role in creating the mod?
Answer: 99% programmer and a tiny bit of an artist for creating some icons. The game developers were kind enough to grant me access to the source code, so I theoretically have unlimited modding powers!
Question: Why did you chose to mod this particular game?
Answer: I really like how certain aspects of the game work from a design perspective. Having gotten access to the source code allowed me to mod far beyond what usually can be done. In all the other games I modded for before, I eventually reached limits in what I could do. In Pandora, there are no limits! That’s what keeps me motivated to mod this particular game.
Question: How does your mod improve on the base game?
Answer: Next to some minor balance changes and bug and exploit fixes, it is almost exclusively the A.I. There’s not a single aspect that I left untouched.
As for the balance changes: Tech progression has been slowed down. All Eras of the game now have a length similar to the first Era because Eras 2 and 3 felt pretty rushed and you barely got the chance to even build all the stuff you had researched.
The yield of trade and research pacts now scales with the number of players in the game and is quite a bit lower than before. Having the majority of your research and credits come from pacts felt wrong.
All difficulty levels have been nerfed because the A.I. had simply gotten too strong on the higher ones and become mostly unbeatable.
Many changes to the A.I.:
- How the A.I. manages citizens
- How the A.I. explores its surroundings
- How the A.I. uses units in offense and defense
- How the A.I. picks technologies
- How the A.I. designs units
- How the A.I. decides what to build in its cities
- How the A.I. decides what tile improvements to build
- How the A.I. uses operations
- How the A.I. acts in diplomacy
Also the A.I. can now do some things it wasn’t able to before:
- Give a new order to units that have moved but have movement points left
- Use artillery units
- Manage their tax rate
- Slow down their population growth when they run out of space
- Handle the promotion operation
I might have forgotten some things, but I think you get the picture. Overall, the playing strength of the A.I. has been increased dramatically. A lot of new players struggle with the fair difficulty level. The “Very Hard” setting, which I could previously beat despite it having much higher bonuses than now, has become unbeatable for me, as well [Editor’s Note: You can find the latest test builds here].
Question: Where do you feel you could have done more?
Answer: While I think that almost everything could be further improved, A.I. diplomacy is the one area where I feel it’s hardest to be satisfied. For most of the game, when putting enough thought in it, you can eventually translate your own decision-making into algorithms that a machine can understand. But when it comes to decision making on a global scale, it becomes really tough. Also this is where the feeling of the player comes into it.
Most players tend to enjoy in-game wars that are evenly matched. Too easy and it’s boring, too hard and it feels unfair. However, to the A.I., a war where both sides are balanced does not seem efficient or strategically beneficial. Further, players don’t want the other factions to gang up on them – even though this would be an extremely efficient and risk-free way for the A.I. to potentially capture cities.
The ideal behavior for the A.I. would be to only ever declare war when the odds are so far in their favor that they cannot possibly lose. So as long as everyone is about even in strength, the A.I. would never declare war… unless the human player declares war on an A.I. But then the other A.I. would join the war on the side of whoever they think is winning. So the player who wanted a fair fight either ends up in a trivial war against a weak opponent, or gets dogpiled by the other A.I.’s. Another possibility is that due to physical distance, one half of the A.I.’s goes after the player and the other half goes after the other A.I. and both get taken out.
You also have to choose how to position yourself diplomatically. The player can prey on the weak to compete with the strong, or work together with other weak factions to take down stronger enemies. Both are valid strategies and it’s not easy to get the A.I. to choose the ideal one for the situation at hand. Also this part is the hardest to test as the outcome is not immediately visible. It’s quite different from: “They should have built this building instead of that.”
I’m still struggling with finding the best compromise here. It would be ideal if I found a way that feels satisfying for the player but didn’t reduce the overall playing-strength of the A.I. Any ideas?
Question: How did the community receive your mod?
Answer: Unfortunately, the community for Pandora is very small. But then again, it’s also very passionate. One of them even told me just recently he thinks Pandora is the pinnacle of 4X and that my changes play a huge role in that. I’ve also received a lot of positive feedback like: “Your ongoing AI improvements were the reason I got the game in the first place.”
However, when you look at the user reviews on Steam for the game you’ll also find other opinions. There are people claiming the A.I. would cheat on even the lowest levels. So it kinda depends on what a player is looking for in a game. My intention is to challenge players as much as I possibly can with my A.I. So the game is not for people who don’t want to be challenged.
Question: Did your mod achieve the goals you set out to achieve?
Answer: Well, no. But my goal was set in a way that I think is unreachable. The goal was to make the A.I. so good that I would have at best an even shot of winning any given game. But as the A.I. became better, so did I. Medium difficulty is not that challenging for me and I can only lose to it with a very bad starting location. I have to play on “Hard” for the A.I. to be on somewhat even footing with me.
At the beginning, each hour of work I put into the game lead to measurable improvements as there was so much to easily improve. Later on it became much harder to get milestone improvements. Progress came down to small improvements or even changes where it wasn’t clear whether they lead to an improvement at all. Sometimes my changes even made things worse.
It’s really hard to tell what exactly still needs to be improved. When I just watch rival factions play, everything looks fine. It must be small nuances that add up over many turns that eventually get them into a situation where I have a big enough advantage.
Question: Is there another mod that you like to use when you play Pandora? Why?
Answer: Yes, absolutely, I can wholeheartedly recommend using BlackArchon’s SMAX Factions Mod – which adds the factions from Alpha Centauri – to go along with the changes I’ve made to the game. BlackArchon and I even worked together to integrate the mod better with the game. For example the faction “Nautilus Pirates” gets a bonus for food and production on water tiles. The AI did not consider bonuses like that before and would not improve their water tiles accordingly. Now it does consider them. I also had an idea of how to make mechs more cool and useful by giving them Pacific Rim-style waterwalking but I was a little afraid of integrating balance changes like that into the base game, so I asked BlackArchon to integrate it in his mod. It looks gorgeous!
Questions: Is there some other modding project you hope to take on some day?
Answer: I have negotiated a contract with the creators of Dominus Galaxia that’s similar to my contract on Pandora. The game is still in the making, and there are some things unclear about it, but I think it is quite likely that I will try to do something similar to what I’ve done for Pandora for that game.
Question: Is your role with Dominus Galaxia going to be exclusive to A.I. behavior?
Answer: Maybe not exclusively, but for the most part, yes. I will try to influence some parts about the user interface, as I’ve seen some potential for improvement there. However, the game design is pretty much fleshed out and is much to my liking, as it is. I’ll also do what’s normal to do as a beta tester and point out everything I think could be improved or doesn’t do what I think was intended.
Question: How is the Dominus Galaxia A.I. going to be different from the Pandora A.I.?
Answer: Well, it’s a completely different game with completely different requirements for the A.I. So I’d answer that almost everything is going to be different. The similar aspect is likely to be the diplomatic behavior.
Question: Do you have any ideas for the general gameplay that you can share from your experiences with Pandora?
Answer: Well, Pandora suffers from a problem that many 4X games suffer from: The bigger your empire is, the more micro-management you have to do. This, in my experience, can lead to less enjoyable mid- and late-game sections. Dominus Galaxia, by design, tries to limit these issues by giving you fewer but more important decisions to make and by putting things at a more abstract level.
Question: Is there anything you wanted to add that I might have skipped over?
Answer: The most important source of inspiration for me, and most likely any of the other modders, as well, is player feedback. That’s more-or-less what drives me to go further and not stop. People playing the game and just talking about their experience is already a great thing. The more in-depth the coverage is, the better.
I’d like to thank Ail for participating in this Q&A. This further shows what a determined and talented gamer can do with the right mindset and some hard work. Game developers have limited amounts of time and resources to devote to their titles. They can’t always hammer away at their design until it’s right in the eyes of their fans. Ail has shown that all it takes is an ordinary community member with a some talent and a lot of heart. See you next week for part three of the series.