Detective Go

Go Variants
81

So I thought about this again and found a significant flaw. If at any point of the game, the detectives find an “anchor point” that allows them to deduce where the fugitive stones are positioned, then the detectives would (practically) have complete information for the rest of the game. What I mean with anchor point is a unique revealed shape. For example if the revealed shape shows one two-stone chain and otherwise all single stones, and this unique two-stone chain is revealed, then the detectives would be able to deduce where all the other stones are, pretty much for the rest of the game.

Maybe one way to fix this, is that the fugitive reveals only the shape of their last X moves (for example X = 2). In my opinion this would fit nicely with the mechanic of Scorland Yard, where the fugitive reveals each turn what kind of public transportation they used.

But I imagine that even this is problematic balance-wise. If the detectives at some point figure out what the last move of the fugitive was, then on the following turn there are only two possibilities for the next move (in case X=2).
One last idea to fix this problem, if it is even possible, is that the fugitive reveals the shape of their move relative to one of their stones with minimal distance on the grid.

In any case I personally agree with @NeilAgg that the fugitive is probably at a severe disadvantage given the current ruleset, especially when playing with 4 - 5 detectives, as I believe was intended by @yebellz. In theory this could probably be balanced by playing on bigger boards, but we’d probably need a really big board, like 29x29, where the game may take a long time.
It’s certainly not sure yet whether the fugitive is at a disadvantage or not, but in case we want to play future games with a system that is easier on the fugitive, here are some ideas for a “detective nerf”:

1. Restrict the detectives communication
Cooperation is key for the detective team, so limiting their communication ability would make the game a lot harder for detectives. However I believe this would be detrimental for the fun and experience, especially since this is a forum game.

2. Decrease the information gain
Currently the rule is that entire chains are revealed, i.e.

This seems to give the detectives a lot of information and also limits the fugitives options to expand this chain, for fear of revealing too much to the detectives. Imagine 4 detectives that find a single-stone … they can immediately capture it without fear of collisions.
So the proposed change would be that only single stones are revealed, not the entire chain connected to them.

3. Restrict the detectives “maneuverability”
@NeilAgg already proposed this:

Since this description leaves space for interpretation, and there are many reasonable ways to rule this, here’s what I would suggest:
The detectives submit their moves simultaneously, but the moves are computed in a fixed sequential order, with collisions and captures being resolved separately between the moves.
The fugitive then plays a move and reveals the resulting position as well as information about collisions and captures to the detectives.

2 Likes
82

I think that my original suggestions for handicapping the game by changing the objective were interesting in principle, but somewhat flawed in practice.

  • A fugitive player trying to delay an unavoidable for as long as possible just drags out the game in an uninteresting way, since the endgame just eventually reduces to the detectives filling space as efficiently as possible and a large degree of luck in hunting down the remaining isolated stones.
  • Making a large of a group as possible, when life has already been established, also just prolongs the game.

I much prefer the basic system of just setting the victory condition as kill all the fugitive stones.

I think a strong strategy here (for the fugitive) might not be to try to establish a living group, but instead just place a single stone (and then pass for the rest of the game), that is, assuming that the fugitive cannot simply refuse to place any stones at all.

3 Likes
83

This is interesting and would make things a LOT harder for the detectives.

84

The first game has just finished

What are your thoughts for the next game? Rule suggestions? Parameter adjustments? Was the game balanced?

4 Likes
85

I think the crux of the problem is the number of moves the detectives get for each fugitive move.
At least on 16x16, 3 detectives is too many. I think we should try the game again with 2 detectives.

I think the rules are sufficient, we just need to do some tuning to find the right parameters to play.

86

I wouldn’t want to limit the number of players if there might be more that want to play. I think having >2 detectives helps significantly to counterbalance the hidden moves of the fugitive, and cutting the detectives from 3 to 2 might very drastically swing the balance in the game.

A more gradual way to strike a balance is to increase the board size. Another possibility (if there are a lot players interested) is to have a more than one fugitive working together as a team. A bunch of other ideas have been suggested above as well.

2 Likes
87

This might be interesting:
Have 3 detectives and 2 fugitives. The detectives must collaborate publicly. The fugitives collaborate in secret. The detectives play individually colored stones, which means their stones do not connect to the other detectives’ stones. The fugitives both play identically colored stones so their moves connect with each other.

This might push the pendulum too far towards the fugitives, but we can’t really tell until we play it.

3 Likes
88

What do you think of these possible changes?

  1. Instead of whole chains being revealed after a collision, only single stones get revealed.
  2. The detective moves are resolved sequentially, chains without liberties are captured inbetween. In particular this would mean that the fugitive only needs two eyes to live.
2 Likes
89

I still firmly believe that 3 detectives on a 17x17 board is generous to the fugitive.

I think yebellz had a few unlucky moves and in retrospect some strategy turned out to be not that good. If the D4 stone would not have been revealed (e.g. by not starting the attack on the lower left with D4), for example, then I think it’s very likely that the bottom left would’ve been captured.

Also, there may have been a bit too much spread of stones in the beginning? Maybe the fugitive has a better strategy in just attacking one area and hoping for the best?

3 Likes
90

I’d be happy to play the fugitive against @NeilAgg and two other detectives.

3 Likes
91

Yeah I wouldn’t mind being the detective again, or happy to let someone else jump in so I can kibitz :smiley:

I think 3 vs 1 on the 17x17 could be even ish depending on strategy.

I could be interesting to do 4 vs 1 or 5 vs 1 on a bigger board, 18x18, 19x19 upwards? :slight_smile:

1 Like
92

I guess I should clarify where I’m standing in this debate. With the current rules and one fugitive against 3 detectives on a 17x17 board, I believe that the fugitive is at a disadvantage.

Yes the worst-case scenarios we thought of were scary, but as was pointed out in the kibitz thread, even in those cases we would probably have been able to capture.

2 Likes
93

Just to add my opinion to the mix, with current rules I think the fugitive would be lucky to win 1 game in 100 against competent play. I think yebellz did the best he could, but still was not even close to living.

I think sequential detective moves (as suggested by Martin) would be a good change. I sort of like revealing whole chains though, since I feel the game gets more interesting for the detectives if there is at least some things that can be said for certain.

4 Likes
94

It looks like the four players (including myself) from the first game plus @NeilAgg are interested in playing in the second. Anyone else want to join? @le_4TC, @Samraku, @mark5000, @KAOSkonfused, @pdg137?

So far, @Vsotvep and @martin3141 have expressed interest in playing as the fugitive. With enough players, maybe we could have a team of fugitives? I’ll like to be a detective this time, in order to see the other side.

I like all of the ideas being proposed. I think some choices might depend on the exact size of each team. Eventually, maybe we can even do some polling or bidding, after figuring out the rough size of the player list.

3 Likes
95

I’m in!

2 Likes
96

I think you are right the fugitive has to be to pick one spot and build as quickly as possible.

1 Like
97

I think that would make the game too favorable to the fugitive. It seems like it will be pretty easy for the fugitive to make a group with two eyes.

2 Likes
98

If sequential detective moves give the fugitives a huge boost, I believe that’s a good thing, because it means that more detectives can participate :slightly_smiling_face:

I’m thinking of something along the lines: 2 fugitives, 7 detectives, 19x19

2 Likes
99

That could be interesting, but I was doing some thinking about how board size affects the game. Because it takes a lot more stones to make a living shape going from corner → side → center, there is an inherent bias in the game. The detectives can play very aggressively in the corners at the start of the game knowing they have plenty of time to catch-up in the middle of the board. That means a larger board (Until it gets ridiculously big) has very little benefit for the fugitive(s).

100

Has anyone played Phantom Go? Phantom Go at Sensei's Library
Seems like this game is a variant of that idea.

1 Like