Grandmaster difficulty

[whole olive]

Hello everyone, I would like to suggest a new grandmaster difficulty for the NPC's.
I don't get the time to play so much multiplayer, I often just do a few moves on the go.
Below are two out of many games I have with the NPC's on master. I definitely find it somewhat challenging, and still need to really think about what I'm doing. Though on a whole personally I'd rate it medium difficulty and would love to see some real challenge.

[whole olive]

Another example Tonight. Anyone else finding them too easy? Or am i just playing really well? XD

[whole olive]

Ok so tonight, i entertained the idea that i could possibly be making it too easy for myself by only having two opponents. This was the result, they did even worse 😆 I am 100% certain they are on master difficulty.

[glass stag]

keep playing AI is adjusted the higher your level, the better they get

[ebon terrace]

Hm yeah, I would've thought it'll likely be more difficult if you have 3 opponents... I'd suggest playing some more 4-player games. Maybe you were lucky they couldn't get in your way too much that game 🤔 but I'm sure it is related to skill as well 🧐

[whole olive]

leveling is very slow, but thats good to know!

[whole olive]

yeah i played another one forgot to screenshot it, was a bit more challenging but i got half way through the game it suddenly snowballed and i wiped the floor with them haha XD

[whole olive]

What do you think about, in regards to a grandmaster difficulty anyways? Possible? Something you guys would consider? 🙂

[inner totem]

Would definitely be nice, especially with how unstable multiplayer can be right now, could just enjoy some more challenging singleplayer games meanwhile. Although multiplayer is still more fun for me personally.

That said, I think it’s very hard to replicate real human unpredictability with AI in this game. There are a few things I notice the AI doesn’t really do, which good players consistently take into account:

-taking into consideration which resource is actually valuable to a player in that specific moment, not just generally strong resource fields, for example recognizing that someone has very limited wheat access right before a barbarian attack and forcing them into a bad position where they might lose a city

-tracking opponents’ hands and adapting decisions based on that, like knowing what they can or cannot build soon and playing around it

-pushing trades based on scarcity, for example if ore is very rare on the board, good players will often force 2:1 or even 3:1 trades, especially if it enables something critical like saving a city or a rare brick to building a settlement.

-understanding risk situations, like when someone is close to 7 cards and might accept a worse trade just to avoid getting robbed

-reacting to the overall game state and indirectly teaming up, for example favoring trades with a rival player so he can get a metropolis, so that the leading player doesn’t get it wins the game

The AI seems quite weak at reading and adapting to these kinds of dynamic situations. It could theoretically be very strong at tracking exact hands, but it doesn’t really translate that into smart, human-like decisions, I feel.

So while a higher difficulty would definitely be welcome, I’m not sure it will ever feel like a real challenge in the same way good human players do. A lot of what makes Catan competitive is the social and psychological layer, especially early game, which is hard to simulate.

If you want a real challenge, I’d still recommend climbing the multiplayer ladder, games can get very intense and close with strong players. But yeah, that comes with the downside of time investment and the current instability issues.

also for cities and knights, i also think it's not too great at reading which win conditions will be open and less contested to others, and planning ahead accordingly, which is crucial at a certain level, I'd say

[whole olive]

From my experience they do that on master difficulty with knowing players needing food and can't get. Like I've had many times where i know they have plenty but they won't trade when the barbarian is coming up. Then soon as it passes they will trade the food for something, though that could just be coincidence 🙂

I agree with the rest of what you've said there 🙂

Yeah I'd like to play more multiplayer but so often end up with people who aren't focusing on the game, so you're waiting almost the whole turn timer just for them to roll, over and over again and that's just frustrating.

[ebon terrace]

Sadly, creating another AI level is not on the list.

[inner totem]

yea I get that, understandable.

Out of curiosity, I did some testing yesterday with a game with masters AI.

First, Pink already had an activated knight out and a 6 wheat field, while I did not have a knight yet, needed an ore, and she was willing to trade 1:1 wheat for ore, which is a bad trade. She could easily push a 2:1 or 3:1 trade. Besides, I do not know what she even needed the wheat for at that time.

Also, she was first in turn order. She had ore and sheep on her starting city tile and the opportunity to block and cut me off from a good expansion spot, aswell as taking my 8 commodity knight spot. Even if she did not want to commit to it and instead stick with an ore-wheat-sheep strategy, she still would have had the better knight position there to defend commodities. And if she never planned to expand there, why did she place the road downward instead of up-left for an ore expansion and ore protection, which she later committed to, in the first place? Bad play.

[whole olive]

Out of curiosity what is on the list? Or is it already published somewhere i can read? 🙂

[whole olive]

One of the fastest games I've had, that was fun haha 😆