Thursday, January 12, 2006

Strategy

References: Brian and Tom at Tao of Gaming.

Brian implies that strategy means "plays that give better positions rather than immediate tactical advantage":

Is Caylus Strategic? More so than PR. There are "positional" affects (things that aren't immediately tactical, but that affect play). Larry brought up an example of "If I'm doing the shipping strategy, I'm going to want a harbor, or If I'm doing the factory, it will change my builds." Well, that's certainly true, and you can see "I want this, then this, then this." But I'm not sure it's positional. You could model the same effects as 'evaluation of future values'. If I have the factory, it's better if I have lots of good types. Ditto Harbor, but the Wharf wants specialization. My current position makes those better (or worse), and I don't have to look ahead to see it.


As usual, Brian is of the opinion that Puerto Rico is all tactical, and not strategic. We've argued about this before.

In the comments, Chris Farrel argues that strategy may just mean "what to do when you don't know the right tactics":

I wonder if a game with completely open information and no random elements must be purely tactical by definition. I've talked about it with the folks I've played it with, and nobody thought it was anything more than a highly tactical game.

A game can appear strategic because we just don't understand it. If there are lots of factors I haven't grasped yet, I'll attack them by making a generalized strategic choice ("OK, I'll focus on building castle bricks") because making all the tactical decisions is too burdensome. But that doesn't *generally* make the game itself any less tactical.


Others go on to imply that perhaps strategy is only possible with hidden information. Tom's post is then largely devoted to supporting this.


I don't think there is such a thing as a game where there are no hidden elements. As usual, I find example games to be helpful in this situation.

GAME A: Each player chooses an Ace or a King to play face down. The cards are revealed. If they match, player 1 wins, if they don't, player 2 wins.

Each player has complete control over what he chooses (no dice), and complete information as to the available options available to their opponent (no hand full of random cards). However, the game has a large and very crucial element of hidden information: what your opponent will choose.

GAME B: The same game, but player A reveals his card first, and then player B has to choose what card to play. You no longer have a game. Of course, this is dramatically simplified, but player B now evaluates the board with complete knowledge of what his opponent will do by the end of the game and chooses the best move. What you have is now a puzzle.

So unless we simply want to define strategy out of existence, we will have to come up with something better than "a game with hidden information", since there are no games without.

One thing you have to realize is that even if a game is "understood", it doesn't mean that there can be no strategic choices to make. It all depends on what you mean by "understanding".

If you have "solved chess", then for any given position you know the optimal play to make assuming that there is an optimal play. This assumption is a big one. Consider the following situations:

a) Your opponent can always win from this position if he plays perfectly. However, you know that he has a blind spot when it comes to tactical play. So you make a move that tempts him to make a mistake.

b) There are several different moves that are equally likely to produce a victory. How do you decide between them? For instance, Rock, Paper, Scissors can't be solved except by statistical analysis of all humans. When playing against a single human, it is more important to know THAT human.

c) Even if you have an optimal move, sometimes you may find a sub-optimal move is better against certain opponents, or, dare we say, more rewarding for the game playing experience.

What these factors have in common is the human experience. If you want to talk only about play that does not only mean "the best play given the information that I can calculate", then strategy is the decisions you make that are based on the human experience.

Maybe one day we will solve PR, but a strategic choice is to invest in high cash goods or high shipping goods. These choices are not necessarily "the best" choices, and they are not necessarily "sub-optimal" choices. They are human choices. Maybe your opponent is not good at handling your chosen strategy. Maybe you feel most comfortable handling it, yourself. Maybe it just makes the game more interesting for you or your opponent.

Either way, the real definition of strategy hasn't really changed much. We all can agree that tactics is evaluating your position and making the best of it. Strategy is still, according to me, a guiding philosophy about the game that colors your tactical choices. It may be that from a grand distance that tactics and strategy have no firm clear border, but that doesn't mean that they don't exist as distinct concepts in human terms. We are not computers.

Yehuda

Update: I think I am letting you guys off the hook too easily. Even if you can see the optimal play from the start of the game, you can still define strategy as the overriding series of plays you will be making, and tactics as each of these plays. As Wikipedia puts it:

* The overall goal is to win a war against another country.
* The strategy is to undermine the other nation's ability to wage war by annihilating their military.
* The tactics (told to the combatants) are to do very specific things in a specific place.

In games, a game is strategic (and less tactical) if there are at least two different paths to achieve the same goal and the tactics you choose specifically favor one of these paths. A certain tactic becomes less valuable or more valuable depending on which path you are following. From the same position, two people will choose a different tactic both of which may be close to optimal or equally optimal based on their strategy. It is tactical (and less strategic) if there is only one path, or the paths are sufficiently muddied that from any position you have the same options and make the same choices regardless on any path.

Obviously I feel PR is at least somewhat strategic, but others don't.

Technorati Tags: , , , ,

No comments: