Managing complexity is usually viewed as an issue exclusively for designers. That’s not unreasonable; players do not, as a rule, have much ability to make a game more or less complicated. It’s possible, however, to envision ways to give them that ability–and there would be concrete benefits to doing so.
As an example of how this might work, consider word count on cards. I think it’s generally agreed that cards with more words on them are harder to parse, tend to involve more elaborate effects, and are more complicated than those with fewer. Designers are aware of this, and try to keep word count down where possible.
Unfortunately, that approach can only go so far. Looking from the designer’s perspective at cards as a group doesn’t address what will be in a certain deck, or which cards will appear in any particular game. Thus, designers have a further challenge: not only do they need to minimize word count overall, they must try to anticipate how the high- and low-word count cards will come up in play. It’s no use saying “90% of cards are very simple” if the remaining 10% are both extremely complex and enormously popular.
What’s more, different players have different tolerances for complexity. Experienced, highly invested players might find it easy to understand even the most difficult cards, while a parent playing with his or her six-year-old could easily wish for the simplest possible version of the game. It’s hard for a designer in the predictive, top-down position to make a game work for all of these players, even if the game is theoretically capable of accommodating them.
Why not, then, approach the problem from the other side? If word count becomes an explicit, player-facing part of the game, it becomes a lot easier to anticipate how much of it there will be. Word count might thus be limited to, say, 200 words per deck (or market, or tableau, or whatever the game has). Just as with any resource, players can decide how to spend it: a few really wordy cards mixed into a generally very simple environment, or a lot of one-step-above basic cards, or any other combination. The average level of complexity will be consistent no matter how players approach the game.
Of course, word count is a blunt instrument for measuring complexity–but there are other tools available. Each card could have the grade level of its text evaluated. (This isn’t as intensive a process as it sounds; Word has this kind of evaluation as a built-in feature.) The limit could then be based on the number of grade levels overall, or on the highest grade level permitted for any one card. Cards could alternatively be categorized based on the number of steps required by their text, or the number of action windows they implicate. An appropriate standard could be chosen for each game.
Handling complexity in this fashion also enables an opt-in approach that helps players customize their experience. Long-time players can push the limit up to allow for new strategies; parents can directly set the bar at the appropriate level for their kids.
Complexity doesn’t, then, need to be the province of designers alone, nor does it need to be unpredictable. There are techniques out there for measuring it objectively, and once measured it can be treated as a resource under player control, with maximums for individual components and/or for a game as a whole. Approaching complexity in this way could benefit designers, who might be able to say more confidently that they know what the play experience will be like, and has particular advantages for the players themselves, who would be able to tune an aspect of the game that is critically important to enjoyment.