That being said, I’m forgiving when a Civilization or a Total War produces an AI that can do the basics but which still has problems. Coding AI is incredibly difficult, and I’m not expecting it to reproduce anything more than what a player with a few hours experience in the game would be able to do, with the rest of the difficulty provided by behind-the-scenes stat bonuses. The reason I’ve been so annoyed at Rome 2 and Civ 6 and Stellaris is because their AIs couldn’t even do that, which is the point at which you need to seriously rethink you basic game structure.

I wonder if boardgames are our saviour in times of developers embracing what the tech allows them to do. First, there’s a problem of devs being able to model whole WW2 with a meticulous rivet-counting and keeping tabs on how much toilet paper does 12th squad of 3rd rifle battalion of 14th infantry division of Irish Army has. Even worse players think that they want this.

But it’s an obvious problem that boardgame limitations help to solve, the one that you made me think of is AI. Boardgames are balanced and tested, they have some rare updated editions with some fixes to the rules but those are rarely complete. Meanwhile most of those rivet-counting strategy games fall apart once you learn the rules; many modern strategy games with wonderful rules fail to have an AI that is able to do anything against a player who knows the rules. Not mastered the rules, not understood everything, just learned enough to not do stupid mistakes. Many games like Paradox ones evade this problem by having varying starts and randomness in many ways (e.g. you can’t predict how the game will go in Europa Universalis 4 because there are hundreds of independent actors and no sane prediction is possible even if each faction is more or less predictable). But many of currently developed turn-based strategy games – Civilization 6, Endless Space 2, Stellaris – may have elegant complex mechanics that they’re constantly tinkering with. I can understand devs who see little point in trying to make a good AI because the better is your AI’s understanding of the rules the dumber it looks the dumber it looks when a new mechanic is added. Or, in case of Stellaris, removed. Soren Johnson, of Civ4 fame, says that it’s the biggest problem in AI development. He wrote the AI in parallel with game system and he basically programmed most of the mechanical stuff alone which is a rare and I doubt it would even work for later Civilization games with their more complex mechanics. It’s still possible when you confine yourself to making a boardgame port. It may not have 500 hours of gameplay but it won’t leave a poor taste in your mouth when you realize that AI is dumb and the only things that stopped you from realizing it were obscure hidden mechanics and poor UI.

It is a good game; I’ve not delved into buying any of my board games in digital format as I’m not sure I would really enjoy them on my own – board games are definitely a social activity for me. This might at least be a way for me to play more games though. It sounds like you’ve played quite a few digital incarnations of board games; any to particularly avoid or which were good beyond this?