I’m in a pitched battle with my own incompetence to design an n-in-a-row game I enjoy, even though I don’t enjoy most n-in-a-row games. I’m doing it because trying to make lemonade from lemons is a way to improve at game design. The battlefield is littered with rinds, yet I fight on.
The challenge has forced me to understand the problems of n-in-a-row games and to think creatively about avoiding them. My best success so far is Morro, but I’m not sure about it. Now I have a new attempt to explore. First I’ll present the rules and then the why’s and wherefore’s.
Breach is for 2 players and is played with Go stones on an initially-empty square grid that looks like this:
The column on the right is called the scoring track. One black stone and one white stone are set aside to be scoring markers for the scoring track.
I don’t know what the best size is for the board, but 10×10 seems a good place to start.
A Row is an orthogonal or diagonal straight line of same-colored stones. The Score of a row is the number of stones it contains.
- White begins by placing 1 stone on any empty space.
- Then, starting with Black, the players take turns. On your turn you must place 2 stones on any 2 empty spaces.
- If you complete a row with a score of at least 2, and that score is higher than any previous score by either player, you must move your scoring marker to that score on the scoring track.
- After you move your scoring marker, your opponent gets an additional option on her next turn (and only on that turn): instead of placing 2 stones, she may choose to place 1 stone and then replace any one of your stones on the board with one of her own.
- If the longest row is broken up due to a stone replacement, the scoring marker is not moved back – scoring markers never move backwards.
- The game ends when the board is full and the player with the highest score on the scoring track at that time wins (in practice the winner will be obvious well before the board is full, and the trailing player should resign at that point).
Where did it come from?
All n-in-a-row games (that I know of) suffer from some mixture of 4 problems:
- not enough strategy (tactics dominate)
- the pylon problem
I only recently added the pylon problem to this list (the term is from fellow game designer Corey Clarke). It refers to the tendency of stones on the board to clump together and stop mattering before the game is over. They become dead pylons. I’d like to design an n-in-a-row game where more stones “live”.
For the record, my favorite n-in-a-row game at the moment is Pente, because its capture rule partially addresses all 4 problems. I think that rule’s brilliant now that I understand all of its effects. However,
- Though Pente is more strategic than most n-in-a-row games, it’s still too tactical.
- Though it’s more balanced than some other n-in-a-row games, there’s still a first-mover advantage.
- Though draws are rare, they’re still possible.
- It isn’t a “pure” n-in-a-row game because you can also win by making 5 captures (a condition which, one suspects, was added to reduce draws).
- The rules are more complex than I like.
Maybe there’s still room for improvement?
The lack of strategy is the most difficult issue to fix so I start with that. Morro creates strategy through negative feedback. Negative feedback here refers to a penalty for taking the lead or getting closer to the win condition. In Morro, when you take the lead, your opponent gets a stone advantage.
It’s the opposite of “the rich get richer” dynamic. With negative feedback “the rich get poorer”: sometimes taking the “lead” is a bad idea. As a result the players have to take the long view and play toward the endgame, rather than focus just on taking the lead. Voila. Strategy.
The great difficulty is providing just the right amount of negative feedback. If it’s too strong it can wipe out tactics and if it’s too weak it can fail to create strategy.
Morro’s drawless and pretty balanced, but maybe:
- it’s too strategic
- it’s too opaque
- it suffers some from the pylon problem.
The first two problems arise because players place an increasing number of stones per turn as the game progresses. The result is a tsunami of stones in which both tactics and clarity are lost. The pylon problem exists because it always exists in n-in-a-row games unless you design it away and I didn’t – I wasn’t paying much attention to it back then.
So back to the drawing board. The tsunami of stones can’t be calmed without also dumping the feedback mechanism, so I decided to try other negative feedback mechanisms. I lit upon one which also addresses the pylon problem: when you take the lead (i.e. when you build a row longer than any built up to that point in the game), your opponent can (optionally) replace one of your stones with her own, in lieu of placing one of her stones on her next turn.
Example: lets say you’re playing on an 8×8 board and your opponent has a row of three and a row of four on the same column, and they’re separated by one of your stones. In a normal n-in-a-row game the whole column would be dead. But here you can’t take the lead without allowing your opponent to remove that one stone of yours and to create a row of 8, which is an automatic win for him. By this kind of effect, the stones on one part of the board are important to what goes on in other parts.
This illustrates a key point: your opponent’s longest row is actually the sum of his longest two rows which are separated by one of your stones, unless you’re in the lead and can remain that way for the rest of the game. That’s the key insight around which to start evaluating possible moves and build your strategy.
Time will tell how good this game is. It shows promise but I’ll reserve my opinion until I’ve played it more, per general policy.
There are two other feedback rules that I tried and (tentatively) discarded. The first, which I felt was too weak, entailed choosing optionally to flip an enemy stone instead of taking your normal turn, after your opponent takes the lead. The second, which I felt was too strong, was to flip an enemy stone in addition to taking your normal turn. If further experience shows that I chose wrong, I’ll revisit these. It’s easy to get designs like this wrong, as I’ve discovered time and again.