Are there any ‘search patterns’ which are more likely to deliver hits as opposed to just random squares?
A while back I thought about choosing grid locations in diagonals, to create one big ‘X’ across the board (or smaller localized ‘x’) given that I’d be more likely to get a hit eventually. But is there any merit to this? I worry about being too consistent in my tactics, since having a ‘pattern’ would make it easy for the experienced opponent to position ships to make hits extremely rare.
Since the submarine is two units long, you can guarantee finding it if you make a checkboard pattern of strikes. However, once you sink the sub, you can search every third tile in the remaining rows.
The best place to hide your ships is to capitalize on your knowledge of your opponent. If they strike away from the last sunk craft, you can cluster. Or vice versa. If they like to make an X, put your ships near the middle of each edge. Or vice versa.
Finally, if they rely too heavily on the checkboard strategy, place a long ship, such as the battleship 1 space away from an edge (parallel to the edge). Then, hide your submarine along the edge. This way, if they happen upon your battleship, they will probably sink it relatively early, but strike the squares where your sub is hiding last.
If your opponent calls a square that would hit your ship, move it stealthily to a new location. When playing honestly, I’ve found that I had the best luck putting most of the fleet in one corner and a rogue sub off by itself.
The submarine is three units long. The destroyer is two units long.
The cruiser is also three. The battleship is four and the aircraft carrier is five.
You can usually get the part of the carrier to hang off of the grid, so a portion of it can’t be hit. This tatic tends to really piss people off though.
In two-player games of incomplete information such as this, there really isn’t a significantly beter-than-all-others deployment or search strategy. If there was, then the opponent would also know it and use that against you making it a poorer than average strategy.
You pretty much have to go with random looking methods that make it difficult for your opponent to figure out what you are really doing. (Actually being random is fine against a better-than-human opponent. But since humans aren’t that great of players, this gives you a leg up by exploiting their faults.)
I always thought the 2-hole boat was a PT boat.
Sometimes I would put the two and a three unit boat together so if they were sunk the other player might think he sunk my carrier and then when actually hitting my carrier might stop too soon. This worked a few times
This is of course a tatic that was employed by a 8 year-old verses another 8 year-old.
Cheat! :eek:
Though I do not have the ability to run a simulation, intuitively it seems that if you leave three squares between your guesses (i.e. hit every fourth square) you will eliminate options a lot quicker than a checkerboard pattern. First, you can quickly eliminate carrier and battleship locations, while still randomly hoping to hit the smaller ships. Only after you’ve eliminated the carrier and battleships should you shrink the net. This has the advantage of sticking with a checkerboard pattern that you’ll need to adopt if you have to hunt down the minesweeper.
Methodical checkerboards are too slow to develop against a spread-out opponent, and usually take a while to catch a clustered opponent, too. Smaller starting nets (trying to catch all but the minesweeper) also take a bit of time to develop. Random strikes tend to force you to target a lot of adjacent squares (which is a waste), though if you random strike only on checkerboard patterns you’ll be ok, too (or, for that matter, random strike on locations that will eventually lead to the pattern I described above if you think it seems sound as well).
Knowledge of your opponent definitely helps, though. Some people almost steadfastly refuse to put ships along the edge. I do not know why. Also, not many people have good poker faces. 
I used to use the diagonals-separated-by-three method described by erislover, until I took a few art classes. There’s a general principle in 2-d art (painting, photography, etc.) that the main focus of your piece should not be centered. Rather, it should be at one of the four points one-third away from the edges, as this is more aesthetically pleasing. It seems that humans placing battleships have a tendancy to position their fleet aesthetically, too, so something resembling a tic-tac-toe grid (starting with the intersection points) seems to find most of the enemy fleet disproportionately quickly. I usually do try to keep my points consistent with an eventual diagonal or checkerboard strategy, though.
Conversely, most people who try to play randomly will also subconsciously form a pattern similar to the above, so keep your own ships away from those tic-tac-toe points.
Also, once you start hitting ships, your pattern (whatever it is) is going to break down eventually. At that point, you start looking for “how many possibilities can I eliminate with each shot”. At the extreme case, suppose that the board is entirely filled, save for one row of nine squares, and that the only ship left is the big one (5 spots, generally the carrier). There are five possible places the carrier could be in that. If you shoot at one of the ends, then you’ll only hit one of those nine, but if you shoot in the middle, you’re guaranteed a hit. Intersections of open lines are also good targets.
And you should try to sink the big ships first, because you might get the small ones by chance, but if you try for the small ones first, it’ll take forever.
I drafted a pretty detailed post last night, but then decided not to post it. The best you can end up doing in battleship is to approach chance if both players adopt the same strategies. You cannot do better than chance against a good opponent. There are bombing strategies, which we’re discussing, but the flipside is ship placement strategy. Of course they are symmetrical: any failure in ship placement strategy lends itself to a bombing strategy, and any failure in bombing strategy lends itself to a ship placement strategy.
If you notice the length of the ships, you’ll see that there are three odd-numbered ships: two 3-position and one 5-position. If you were to picture the board as a checkerboard, you’ll see that you can bias a “color”. In fact you will always bias a color, but you can have the two 3-ships cancel out their color bias, leaving only the 5-position ship to favor one color by one square.
If you’ll notice, there are five ships. This means you will be forced to create an orientation bias: a majority of ships will point horizonally or vertically.
If your opponent gets a hit on a ship, he will have to make a directional guess as to its orientation. You can minimize orientation bias by arranging your ships so three point in one direction and two point in the other. This will forbid him from adopting an orientation bombing strategy.
As a general rule, don’t try to out-think ship placement. Avoiding “easy” diagonals, always spreading, always clustering, etc, are losing propositions that will allow an opponent to develop a better-than-chance bombing strategy.
As far as bombing goes, play the odds and use deduction to avoid bombing statistically useless tiles. Think about how many n+ length ships remain, and play the odds in guessing orientation when possible. For example, suppose the board looks like this:
0000000000
0000000*5*00
0000000*5*00
0*4444***X**0*5*00
0000000*5*00
0000000*5*00
0000000000
0000000000
0000000000
0000000000
Given this state, two out of the three ships would have to be placed vertically for this hit to be the case.