Attention in Myth

First I ought to summarize how the game works, in case you're not familiar with it. In the simplest case, there are two or more players. Each player receives a fixed number of points, then trades those points for units of various types. Then the fun begins! The players move their units around in a little virtual world of hills and rivers and such, and make them fight with other players' units. The goal varies, but fighting is always involved.

Now, a player's units don't all have to be in the same place. If they are, then of course the player will watch that place, and keep vis full attention there; if they're not, the player will have to jump back and forth, and divide vis attention in some appropriate way.

The thing that's interesting to me is how the amount of attention interacts with the point value of the units. Let's look at the limiting cases first. Suppose two players are fighting. On the one hand, if both players are paying full attention, the one who has more units will do better, all other things (skill, terrain, etc.) being equal. On the other hand, if one player isn't paying attention, i.e., is distracted, the other will be able to inflict disproportionate damage.

As a result, even if you're far behind in terms of units. it's still often possible to win the game by taking advantage of your opponents' divided attention. Of course there's not much you can do if there's only one other player left, or if one of the players has an overwhelming number of units, but in most other cases you can find something worthwhile to do.

Another result is that if you're playing as a team, with several players sharing one set of units, it's surprisingly useful to have some of the players take relatively small armies. A small army can harass a much larger one by staying engaged but always falling back when pressed; the player controlling the larger army is required to pay full attention or take damage. Even an “army” of a dwarf and two melee units can't be ignored.

Now, here's the point. Because the point value of the units is tangled up with the amount of attention they receive, it ought to be possible to quantify the value of attention! I estimate that full attention is worth about forty points, or about half the value of a full set of units on a small map.

Of course, that whole model is an oversimplification. As a better model, I'd assign three numbers to every kind of unit: its base value when uncontrolled, its full value when microcontrolled, and the amount of attention it takes to microcontrol it. That model accounts for many things, e.g., the fact that it's not very productive to focus all your attention on a single melee unit.

Naturally there are plenty of other factors that one could take into account. A more experienced player effectively has more attention; the full value of a unit depends on the skill of the player with that kind of unit; and so on.

I enjoy thinking about attention just as a component of the game, but there's also something else about it that intrigues me: this attention that we can divide among units, and measure the value of, is the same attention we use in the real world. It doesn't just behave like it, it is it! If I try to play the game and talk to someone at the same time, I'm dividing my attention … some is in the game, animating units, and some is out in the real world. Isn't that strange and interesting?