Tips coupcitystate
[currently still waiting for permission from Author of https://medium.com/@wilson97_82892/coup-strategy-part-2-860a36c1ceb8]
Tboard
1. Don’t go crazy early game There is no reason to be couping, stealing, duking, or assassinating early game. Most of the time, the people you couped and assassinated would have eventually been killed by someone else, and by killing them you give people information about your hand, as well as setting up potential revenge coup situations. Also, most players when given a choice between couping someone active vs. someone passive will pick the active person, as they feel the active person is more of a threat than the passive one — not necessarily true of course, but that is a common bias people have. Another way of putting this is that early-game couping/assassinating someone does not bring you any advantage, so there is no reason to do it. Be patient.
2. Figure out the risk tolerance of the people you play with If you can figure out the tendencies of your opponents, you can tailor your strategy to them. For example, if you are certain that someone lies a lot, you are more confident calling their moves, and if you play with someone that hates lying, you shouldn’t call as much. The tricky part is figuring out these tendencies in the first place, but after enough games with someone it usually becomes apparent what they like to do. On a related note, some people are worse at lying than others — they have tells (just like in poker). For example, one of my friends will start drumming his fingers and shaking his leg when he does something risky, so I know to be careful when I see him doing these things.
3. Bluff the ambassador If the people you play with are very tight and don’t like calling early-game, bluffing the ambassador can be very powerful — it lets you see the deck, as well as letting you bluff ambassador later. Once your first bluff goes through the chance of you getting called on your second or third bluff decreases, as people get the impression that if nobody called you the first time, you must be telling the truth (which is a logical fallacy, and one to take advantage of). Be careful, though — if other people are also ambassadoring, they may have enough evidence to call your ambassador bluff, usually by seeing 2/3 ambassadors in the deck. Therefore this strategy works best when you are early in the turn order.
4. Don’t get discouraged when you lose a card Very rarely do games end with the winner having both their cards alive. This means that almost all games come down to a situation where multiple people have one card alive. Thus, if you get called on a bluff or get assassinated early, you are very much still in the game. In Coup, you can’t really get ‘behind’ early, because as mentioned everyone will get taken down to 1 card. A related idea is that you shouldn’t be afraid to bluff early — best case your bluff works, worst case you get taken down to 1 card, which again isn’t a big deal. In other words every game you get one “free” bluff — use it wisely.
5. Don’t use the same strategy all the time If you always do the same thing and someone figures it out, your strategy becomes useless. Therefore, you should switch between various playstyles and risk tolerances to throw people off.
6. Try to get to a hand with at least 1 captain We figured out last time that the captain is by far the best 1v1 card, so when you are ambassadoring early game try to get your hands on a captain. If you can get both captains, even better.
7. Contessas aren’t very useful Again, if you get assassinated and lose a card early, its no big deal. On the other hand, if you end up with your last card being a contessa, you have limited options — you’re basically forced to bluff. So even though a contessa might make you feel safe, in the long run it doesn’t help much. In the next article I’m going to attempt to write a Coup AI. When it works, I’ll have it play against itself to figure out what the best strategies are for situations where logical analysis is too difficult (which is like 99.999% of the game). Stay tuned!