While I havenít consciously adopted a "philosophy" of design, these are my observations of what happens.
I prefer fairly simple-to-play games over fiddly, rules-complex, or many-pieces games.
I try to keep the number of pieces, cards, and other Stuff a player controls to under two dozen in multi-player games, and under three dozen in two-player games. Strategic and tactical complexity can be achieved without large numbers of elements to keep track of.
On the other hand, I am not of the "I only want a few alternative choices" school. In other words, a chess-like element is often in my games (lots of different possibilities, though few pieces). I have simplified some of my games to the Euro-style viewpoint that only a few practical moves need to be considered each turn.
I try to reduce the effects of chance, either with lots of die rolls that will tend to "even out" (Britannia) or by using elements such as cards that players have some control over. The more cards are used, the less dice should play a part.
I design games, not simulations; I prefer players to have control of their pieces; perhaps they canít do everything that they want to (thatís good, it makes them make choices), but there is no random element that prevents them from doing something with their pieces, however realistic that may be. In the end, itís a game. Hence, I have little interest in the "card-driven-game" style.
I like the game to represent something. In many Euro-style games, the atmosphere (often wrongly called theme in this case) is "tacked on" (and could be changed considerably), and the players are entirely concerned with pure mechanics and with the other players. I like to be able to understand that when I move something in the game, or do something in the game, itís something like an event that could happen in reality.
An historical game can teach the players something about history. I am not, however, of the "what if" school of varying one factor or one decision to "see what would have happened". My games tend to be at a high (strategic) level where it is practically impossible to "model" the factors that produced history, so it is rarely practical to use the "what if" query.
Despite that, I do design some fairly abstract games, because abstract games are a form of "pure" game. What I rarely do is design an abstract game but pretend it's something else.
People play games for many reasons. I play either (in cooperative games such as D&D) to "succeed in the mission" and keep everyone on my side "whole", or (in competitive games) to win the game. I like to know the rules of a game thoroughly; I much prefer to read a set of rules rather than have someone teach me, probably because I want to thoroughly know whatís going on. I recognize that the rules-reading preference, in particular, is a minority view! Nonetheless, I tend to design games that I like to play.