swap move in Slither

Slither is such a wonderful connection game. I am grateful to LG for having implemented this game.

The only minor unfairness, thus imperfection, rests on the swap move. Since any stone can walk on the board, even a corner inferior swap move can eventually become a powerful stone on the board after a few walks in the beginning of the game.

Could the rule be revised such that only the swapped stone is forever frozen/stationary?

If a similar comment was already made in other existing threads, I apologize.

Enjoy your holiday season, folks!

I agree regarding the problem. The suggested solution is interesting and possibly effective, but it also has the smell of 'quick & dirty' about it, and the casual onlooker stricktly speaking no longer has 'perfect information'. Let's see what Corey thinks about it.

I actually noticed the swap problem earlier and proposed 3 solutions in the previous topic, any thoughts which is best i think for now c) as it do not change rules to much but might be not sufficient.

a) The first stone cant be moved entire game

b) The second player can place two stones on the board in his last move

c) The first player places more than one stone and than second player can swap. For example first player play 4 white stones in the corners and 3 black close to the center. I think second player should not swap in this position.

to be honest I even thought the game might work without the swap rule right after its initial conception, obviously this isn't the case however I think on the 13x13 board the swap is pretty effective because even though you can move your stone out of the corner it costs you a phase in your move. What really imbues the game with balance is the semi-coldness created by the placement restriction and the fewer degrees of connectivity than Hex. So far I'm not seeing an obvious imbalance with the larger size games and white often wins 9x9 games even between higher level players like 5466 and LEESD. At any rate I tried to think of other ways to balance the game, the connect 6 rule doesn't work because making 2 placements might require 2 movements. I don't like the suggested solutions for the following reasons

a) player's have to remember which was the first stone placed

b) this implies the 2nd player must be close to a win.

but actually I like "c" this is the rule from Chain Lightning but I think maybe even just placing 2 stones on the first and second turn would work in tandem with the swap rule for Slither. But as I said the game seems pretty balanced as it is and apparently its in much better shape than Crossway and I suspect chain lightning is more imbalanced than Slither as well because more of the board is filled at once and less of the board needs to be filled in order to win. Could you even imagine a full 19x19 Slither game between experienced players? It would last upwards of 200 moves!

One might have different opinions as to whether that is a desirable property for a game, but I agree that the turn order imbalance seems to decrease rather quickly with increasing board size. It may all come down to voluminous stats at a reasonably high level of play. My impression is that strategy isn't all that well understood yet (certainly not by me), so the scope of the errors on either side of the board may as yet obscure the scope of the problem.

There is no real difference between perfect play and errorless play. After all what should be considered an error? Obviously every game that isn't played with absolute strategic perfection could be considered to have errors but its the errors that make any game fun. Any game might be on the brink of being solved at any moment, even Tanbo the most perfect model of strategic depth (no I'm kidding) and of course I have to fret over this from time to time in the case of Slither but for now I can rest easy knowing it is a deep and complex game. I might try to balance the game out with some ugly rule if black seemed to win a disproportionate number of games however I would probably just increase the board size again before I added any new rules. Let's not even entertain these ideas of rules to "balance" Slither because getting Richard to implement them would be no walk in the park.

My point exactly, it's too early to tell.

"There is no real difference between perfect play and errorless play."

There you're right Sherlock ;-) When I say 'errors' I do indeed refer to something we cannot conceive, but the existence of which has been proved (often referred to as 'Zermelo's Theorem'), namely that every finite two player zero-sum game is completely determined.

There exists a complete tree. We cannot possibly build it, but let's say we could, then we would have the truth of every legal position, including, in Slither's case, the initial empty board.

Slither is drawless, so it means any move in any position is either winning or losing. There may be a non-constructive proof possible for a first player win, like the one John Forbes Nash gave for Hex (but doubtless more complicated).

Such is the nature of our errors: we lack access to the truth (except in cases where we can see or prove it). We blunder blindly, in part, in the sure knowledge if making errors, but not knowing which is which. Errors is what keeps games going. The truth may well kill them. Anyone for Checkers?

Players do not need to remember which the swap stone is, if it is marked slightly differently from other regular stones, such as with a "+" or "*" sign on the top of the stone. :)

Think of it as the Joker in a deck of poker cards.

After many of us have improved our strength in Slither, this swap move will become a big advantage for stronger players, at least on the 9x9 board.

it already is on the 9x9 board. I made a mistake when I okayed the 9x9 board. Richard Wanted to program a 7x7 board and I told him nothing smaller than 9x9 would be interesting and left it up to him so he implemented the 9x9 board. I'm actually surprised how well the game has held up on such a tiny board probably better than Hex on an equivalent sized board. At any rate if 9x9 fails to be a challenge then play 13x13 which will probably keep you busy for the rest of your natural life ; ) . However if 13x13 takes too long then I could ask richard to program an 11x11 board too. Also as far as balancing mechanisms I would opt for the one I just suggested where players place 2 stones on the first 2 turns with a swap option. There will be no marking of stones, Slither was designed for the Goban to be played with Go stones on some square portion of a go board. Isn't anyone here familiar with the expression "if it aint broke don't fix it?".

I think that on sizes 13+ the swap rule probably works well the way it is implemented now. 9x9 will probably become much less interresting when the game is better understood, anyway, so I consider that board mostly an "educational" board.

Oh, and on a side note I'd like to state that I think there IS a difference between "perfect play" and "errorless play". If playin errorless, you just play any winning move if you're in a winning position, and any losing move if you're in a losing position. If playing perfectly, you play the fastest winning move if in a winning position, and othrtwise the move that "makes it hardest for your opponent to play a winning move", often by causing his number of winning moves to be as small as possible.

On the 13x13 board I personally would only consider a swap on the three spots closest (the L-shape) in each corner. I believe that White's best response would then be to play far away (not sure how far), but in the general area that would initiate his initial offensive "line" while simultaneously defending the line created by black's first stone+movement. White playing further away from Black's first stone will tend to split Black's subsequently played stones, thus losing a bit of coordination ... or at least *obvious* coordination. FWIW, that's how I see things so far.

Thank you very much, Zugzwang. Your advice is worth at least 50 thousand Euros. :)

My conjecture is that a corner stone is very powerful as it is in Go and Hex. It can serve for offense or defense, as you keenly pointed out. It is powerful all the time during the game in progress.

Under this conjecture of mine, any player who gets hold of the swap stone is almost like playing Hex with the first stone placed at (3,3) or (3,4), which is a big advantage.

For me, I will close my eyes, and take the swap stone all the time, regardless of its original position, and regardless of the size of the board. :)

I support 11x11 board for Slither and no changes to the exisitng rules. On 9x9 even the corner stone gets rather fast into action. So after playing a couple of 9x9 games my rule is: on 9x9 always swap. Swapping the corner stone is rather extreme for connection games. However even on 9x9 Slither is not trivial. 11x11 should make much better while keeping it reasonably short. On the other hand 13x13 feels big and heavy in comparison. I have a feeling that 11x11 is a very good size for many Go related games (including Go). Too bad it is usually neglected.

Let me see if I can make my points more clearly by itemizing them:

1. On Hex-19, a (3,4) or (3,3) move is a legitimate powerful first move.

2. On Go-19, the same.

3. On Slither-19, a corner swapped stone can easily walk to (3,4) or (3,3) postion for the first 2 moves in the beginning of the game.

4. In the beginning 2 moves of the game, moving central pieces usually is meaningless anyway. So, why not move the corner swapped stone?

5. The power of a stone at the corner remains the same for Hex-19, Go-19, and Slither-19 on any board size. In other words, an angel gives you 25 euros, but gives me only 20 euros. Her giving is unfair. If she gives you 1005 euros, but gives me only 1000 eursos, the unfairness becomes less obvious. However, the advantage of 5 euros remains the same.

6. So, I propose that, if the angel gives you 25 euros, she should give me 24.9 euros, or 25.1 euros, as closely as possible.

7. My proposal is equivalent to marking the swapped stone, which is frozen throughout the game.

Merry Christmas and happy new year to you all. Tim

On Go-19, the same.

Go does not apply here. In Go every stone counts sooner or later because the whole area counts. In connection games a stone counts only if it is part of connection. Hex is also different because of the Y pattern. Because of this it is easy to change the direction of flow. But in Slither it is just horizontal vs. vertical. If your stone is not on the right column / row you will have to move it. If the board is big you will have to move it more on the average. As every move counts it may be too much at some board size with proper counter play. Lets not forget that Hex is deterministically won for first player (or second with the pie rule). Strong opening moves are the ones for which the game is "felt" to be interesting. They do not change the deterministic character of the game. The same goes for Slither. 9x9 is felt to be too small. I say, let's try 11x11 and see what happens. The smaller the better.

The only rule that makes sence is: Place as many stones (of any color) as you want on the first move and the 2nd player decides if he wants to play white or black (e.g. swap or not). This has the additional benefit of random starting positions.

But for example a frozen stone wouldn't work at all, because sometimes to connect you have to move the stone. So sometimes the frozen stone will be more of an obstacle than a helpful thing. You could say you can only move the frozen stone if you place adjacent to it, which makes it only half frozen and you have to keep track. Is it really worth to change the game just because black might have an advantage? In chess white has an advantage, still it's a playable game. So my personal answer to that question is no, the only rule I support is the one above.

In a game with swap rule, the first player should not have too much freedom to create a starting position. Otherwise he could present the opponent a difficult and tricky position he has thoroughly analyzed before.

There's a difference between an extended pie and the 'Marquisian Method'. In an extended pie there's no more trickery involved than in a single pie. In both cases you accept the color that last moved or the color to move. The Marquisian involves the trickery you mention. To show the difference (in simple games), InSight has an extended pie, Swish & Squeeze feature the Marquisian.

Not that I feel Slither urgently needs a balancing rule on say 13x13. Not unless it becomes so succesfull that masters would emerge (and maybe not even then).