YT suggested, I didn’t go looking for it.
It’s funny how the 3 people I watched (I only knew of dwyrin and In sente, I don’t think I knew of Baduk Scholars) made so different reviews, each one focused on entirely different things.
2 Likes
Since I last posted, there have been two feature updates to the game, in addition to a number of patches to resolve issues since launch.
Update 0.3.0 focuses on improving the ease of use of some of the campaign features such as reviewing AI move recommendations, gathering resources, and loading of previous match records. The full details are included in the link below.
Update 0.2.0 focused on improving the difficulty system within the campaign to allow players to fine tune their handicap and AI strength independently for each board size (9x9, 13x13, 19x19), which continue to update independently throughout playing the campaign. The full details are included in the link below.
4 Likes
As of now, the plan is solely a PC release, with potential for Mac later on. There are no plans for a console release, but I won’t say it’ll never happen.
Thanks for the feedback. There are improvements to the campaign that will be coming through Early Access prior to the full release to make the campaign have more of an engaging goal/purpose.
I hope it’s still working well for you. I’ve had some reports of everything working fine under Proton, and others that were having specific problems, so it’s a bit hit or miss depending on the different variables associated with Linux/Proton. There were a few issues related to themes that were patched earlier, though I’m not sure if that was specifically what you were running into.
Thanks for the feedback. I do plan to make the black text a bit darker in the future, as it has been mentioned a few times during feedback. Where are you seeing white letters at in the game?
4 Likes
Sorry, I meant gray letters over white background.
No problem!
Version 0.4.0 was released today which has a number of improvements such as new campaign difficulty presets, better tooltips in menus, better organization and reloading of the main menu, and more. The full details can be found in the link below.
7 Likes
I’ve just tried the game, and the battle map looks very interesting, the music is good, although sometimes the details in the background is a bit distracting, and make my reading difficult and slower (but probably just me used to see a wooden board and not get distracted by the wood texture)
The switch between the battle map and the campaign map is a bit unintuitive, and the territory on the campaign map is not very well defined.
And a few suggestions:
-
maybe we can have a minimap indicating the current campaign map progress (like most turn based strategy games) while in the battle map. With different colors indicating which territory is already “conquered”, or still in contest. I feel it would make the “battles” more connected to the “world”, and can be an intuitive shortcut to switch back and forth between battles and the overall campaign.
-
Maybe we can have some existing board position from some existing unfinished positions, some famous games, as custom battles, or famous battles that players can continue from certain position and see if they can change the results (for better or worse), like historical simulation, but for Go games.
-
Maybe some mini-games like tsumego problems, can be sprinkled in like mini-quest where players can get rewards for resources when they finished them, it would be fun and improve the teaching and promoting.
-
Maybe we can have layers to pick a starting location, like a home base, and put some “buildings” in them that are unique to it, where players can purchase some assistants, or received quests, or can upgrade with “upgrade” battles, where there is no penalty in losing, but some benefit winning (it can act as practice games for new starting players, to have less pressure to start conquest right from the beginning). Maybe even integrated tutorials in them.
-
Maybe we can have more than players vs AIs in a campaign map, but also AI factions that battle each other as well. Multiplayer in the same campaign would be fun, even just for friends to play together, maybe even later on expanded further to online multiplayer, or pair-go like options.
-
Maybe some flavored background “animations” or citizens to show up within the conquest region’s territory, they don’t have to serve any function but would make it feels a lot vibrant and alive. Like towns or small houses, even castles, depending on the territory sized, etc.
Thanks for all the effort in making this game, and hopefully we will get a better and better game.
5 Likes
Thanks for the feedback. I definitely plan to make some improvements to the campaign mode to make it more interesting and strategic before exiting Early Access for the full release. I already have similar plans or variations in mind for your suggestions #2, 4, and 5.
- There are a few different ways to get back to the campaign map, such as clicking the icon in the top right toolbar, clicking the button in the match results panel, or clicking the button in the pause menu. I agree though, they don’t exactly stand out, and can be improved.
- I’m not sure if you’ve run into yet in the campaign, but there are invasions that take place where matches will start from a predetermined board position, usually 4-15 moves into a game. The invasions are a mix of realistic board positions which could have resulted from normal play, as well as unrealistic positions such as black having a number of stones in the center with white having some control over each corner. I’ll consider adding some historical positions into the mix.
- I’ve had similar thoughts about tsumego, however it’s difficult to create a catalog of original tsumego that covers the full range of skill levels and is large enough that players aren’t replaying tsumego.
- Something similar to this is planned to be implemented.
- Something similar to the AI portion of this is planned to be implemented. Multiplayer may be a possibility farther in the future.
- Do you mean on the map itself? Currently, when you click on a game record it will show the end state of the match, with the territories populated with buildings, trees, etc.
2 Likes
Thanks for taking the time to read my long text and reply them :). I want to follow up some of them though
I feel it’s less of a problem of lack of options, but lack of consistency. Multiple ways makes it more confusing than helping. And I feel the second part of my suggestion is more what I was trying to say - the mini-map. It may sound a bit redundant, but the point it to make the battle map feel like part of the larger map, instead of “switching them”, it will be like “zooming” in or out, by just clicking the map, and a minimap is not just a big bottom, but should change according to the current progress of conquest. There is a reason why from strategy games to building games, to even online RPG them all have the same mechanism at play, and most computer game players would recognize it without second thought.
There are tones of categorized tsumego with levels, 围棋级位测试6000题 - 101围棋网
I don’t know if you can read Chinese. There are lots of collections for specific levels with thousands of them 题库 - 101围棋网. The problem might be it would take a long time to implement them. Or maybe can cooperate with some websites if they have a way to automatically link their daily quests and show them inside the game.
Since there are AI options in the game, maybe even AI+player vs AI/player, pair-go could be fun Like having AI generals/advisors to assist you in a very difficult campaign, where players will less skill can still play and enjoy it, maybe need to pay certain upgrade to have a certain period of assistants, etc. Not just a one-time hint.
Yes, on the larger campaign map reflecting some of the building art similar to the battle map, not just different colors dot, so they would feel more connected. And possibly inside the battle maps with some tiny people walk around “the town”, and not just generic builds, but different levels and sizes, maybe even later related to how “powerful” or “upgraded” the overall “empire” is, showing the splendor of the conquest. Visualized progression to encourage players and make them feel connected to their “possessions”.
And I did see invasion start with existing positions, but didn’t think much of it. Maybe we can even have a “player invasion” later on for players to invade AI holdings (instead of start it blank, the maps are started with many AIs holding covering it all). And we could have players to start from a position where the analysis showing a big blunder, and revert back several moves for players to have a do-over invasion.
3 Likes
No problem, I appreciate the feedback.
On the topic of the tsumego, those collections exist, but those are property of that website and I cannot legally just copy them into my game. Perhaps in the future some deal could be worked out with a tsumego repository which has original tsumego content that I could license.
I like the idea of the option of the player and AI being on the same team. This seems like it could be a good option for beginners.
I’ll keep the rest of your feedback in mind as I continue to build upon the game in Early Access.
1 Like
BTW, maybe just one more technical feedback
The new showing played game analyze progress is a nice touch, but I can still enter the conquered region before the analyze is done, and all the analyze options would still be clickable, even though it shows nothing in win rate, scores, et.
Maybe we can show partial analyze data and draw them on the Go according to progress, or like https://www.zbaduk.com/smartreview they have a quick rough review curve, and slowly filling in the details. Not sure how they do it, but if it can be done in the game, would make the experience more fluid and make players want to use the analyze more (I feel most novice players would benefit from some instant feedback when the game is still flesh in their minds, instead of go to a new region to conquer and come back, and might already forgot what happened or don’t even want to review anymore). Not sure how difficult this can be done though…
I’ve heard that chess and go problems cannot be copyrighted, but the solutions can. So sites and apps use a pro, or crowd-source solutions.
A small collection of very easy tsumego for teaching the basics of eyeshape and life and death might be worth it just to get beginners going.
1 Like
It definitely seems to be a legal grey area, requiring assistance from a lawyer. In addition to the solutions, organized collections cannot be copied in full.
There are basic life and death problems sprinkled into the tutorials within the game, but it could definitely be expanded upon.
2 Likes
Just out of curiosity, is it possible to use AIs or programs to generate tsumego problems, or variations of existing tsumego problems?
My thought is that this might be difficult since it counters the way AI thinks I believe, and needs to create “looks promising but actually losing positions”.
1 Like
I know such programs exist for chess. The difference is that chess puzzles tend to be whole board problems.
I’m also pretty sure I read somewhere that somebody wanted to look into this for go. Might even have been on this forum. No idea if anything came out of that.
1 Like
Can AI learns to create them just by looking at a lot of tsumego problems?
Interesting thought. That would not even have to involve solutions and should be easy to build. You are tempting me …
1 Like
A start in that direction is kind of seen in AI Sensei’s Training Mode feature
As far as I know, you can make puzzles out of your mistakes in a game and then do a quiz on them later. I think at the moment it’s a bit more like find the right move, as opposed to solve the puzzle completely and the answer might depend on the level of support you pay for the site. That is, if the AI does more playouts in the game review it could favor different outcomes.
Then a recent version of Katrain has this local positional analysis mode
I wonder if there could be some sort of automated system between the two, where you give the new program an sgf file, it reviews to find the biggest X number of mistakes. Then it uses a local analysis mode to decide if the mistake is a local one - that is, was the best move also local as opposed to a tenuki. If the score loss is big enough, it could be that something got captured, or some endgame went wrong etc and it might be useful as a puzzle. Then save the local variations as a tree to make a puzzle from.
It could be a bunch of work, but it wouldn’t involve training a new AI on generating puzzles at least, that and even if the AI could generate the puzzles, it should also be able to solve them 
Maybe @sanderl has thought about or knows about projects like this before?
Wow, very interesting, although I usually find AI’s suggestions and variations for better moves (especially early mid-game) involved early settle and reduce potential ko threats, instead of just a straight-up kill, or necessary fix (I find AI likes to trade, so it usually doesn’t mind getting killed, and still be able to make up somewhere else). AI tends to be very light in defense and give sente a big value (hence the constant tenuki). Maybe forcing local respond like Katatrain would make it reconsiders, but I wonder if it would just force AIs to create complicated local fight variations instead of a deep trap where only humans would fall for (not to mention low level players would fall for). Setup to fail is not something AIs are good at I imagine. Even when I try to kill a seemly open group, I find AI usually tends to suggest for setting up ko, or even step ko, not to mention, create multiple cut points super complicated interconnected semeai spread to nearby groups, which usually is beyond human reading, and nowhere nearly fun as tsumego problems.
1 Like
Defining what is ‘local’ automatically is very hard, and kind of goes back to why AIs were so bad at go for forever: evaluating a position is hard to describe in an algorithm.
In any case I disagree with AI sensei’s approach in putting that much focus on the AI’s top choice. It’s the only thing it can do, not having a live connection to an engine, but I personally think it destroys much of the openness that makes go so amazing. There usually are many good moves.
However, this is an interesting site by KataGo’s author which does something in the space of auto generated problems: https://neuralnetgoproblems.com
Essentially: Find situations where a 14k would make a certain mistake, but a 10k would play correctly, and present these as 14k problems.
2 Likes