Rage Against the Machine
For our pre-game game yesterday, my soon-to-leave-for-Zambia buddy and I took a second swing at the latest COIN game from GMT, Pendragon.
Last time, we divided the four factions between us in the obvious way: Britons (Dux and Civitates) for him, barbarians (Scotti and Saxons) for me. Fairly straightforward but, you know, first read/play is always tough. We made it about five cards into the game before we had to put it away.
This time we decided to give the NPC AI a go! Isn’t it silly how we call algorithms “AI” any more? Such scifi. I’ve had mixed luck sussing out GMT’s flowcharts in the past, mostly because they’re so frigging conditional in order to have the algorithm play as hard as it can. It plays pretty damned hard in Pendragon.
So I played the Civitates (Romanized civilian rulers, the folks basically learning how to be the aristocracy in a couple hundred years, dedicated to civilian rule and yearning to break away from Rome), he played the Dux (Romanized military rulers, still dedicated to Roman rule, total hardasses stomping around the island but inexorably losing everything). The AI ran both barbarian factions.
Really, the flowchart is mostly common-sensical if you know what the best plays already are. Like, you could convincingly play “against” yourself and it’d look about the same, right? The barbarians prioritize pulling their plunder back home to bolster their renown, or taking advantage of their best event cards if they have no plunder, or starting shit if they’re in a good position to do so, or raid, and so on down the list of all their options. The tiny conditionals that help you decide where they’ll hit, that stuff is fussy to get exactly right. But I had a better time, for sure, playing against the algorithm than when we combined forces.
This time we played about 3 hours and made it to the first Epoch (there are only two Epochs in the “short” intro scenario). That was really our goal, because in all COIN games the big “stop now and do a bunch of other stuff” rounds are mostly where the games are won or lost. It’s also why I mostly suck at winning COIN games even while understanding perfectly well how to play COIN games.
The Epoch round in Pendragon is pretty cool! You start out paying for all the foederati you probably stupidly hired up, and dealing with the fallout when you can’t or won’t pay them (spoiler alert: they wreck your shit), checking to see if the political landscape has changed (Roman to autonomous, military to civilian, or total collapse into fragmentation), then lots of maintenance dealies and kind of a map reset — Roman cav runs home, barbarian warbands return to their villages, stuff like that. It’s time consuming and involved enough that I feel like this is probably where a masterful player can snake a win without less-masterful players even knowing what’s going on.
Playing through their algorithm is also super-educational about what “good” play looks like. I had taken a swing at the barbarians last time and sorta-kinda felt out what they looked like, and the NPC algorithm both confirmed that and reinforced stuff I wasn’t sure of. Specifically, the value of settling down on the island.
I feel like at this point we could prrrrobably get through a two-players-two-NPCs short game in 5ish hours. Onboarding two other live humans, though? It’ll be a long day.
Pictured below: asshole Saxons putting down roots on the little spit of land that’ll someday become Kent. Canterbury I guess? It’s super fun to try and map the deep-history names to the modern.