I don’t know how you get to the maxim ““I would like to break the rule that I also think is essential” at the heart of “running a red light”.
I don’t consider traffic lights essential at all (what do you mean by “essential”, essential for what). There are any number of circumstances I would think it is adviseable to run a red light.
Traffic lights themselves are not essential, what is essential is the binding force of the law behind the simple traffic light: although most of us automatically press breaks when approaching a red light, the question that we almost never ask is “why?” – Why should I stop at the red light even if there are no cars going in the other direction? Isn’t the red light there simply to regulate traffic? If the goal of the red light is to help me, to make sure I am safe, to make sure I survive, i.e. if in evaluating my stopping at the red light, I only care about the consequences, then there’s plenty of reasons to run red lights all the time. Consequentialist ethics fails, Kant would say, because it is too dependent on contexts and circumstances, it does not therefore have principles, only contingent guidelines and advices, it’s not scientific then. Traffic laws are formulated in a way that at least attempts to communicate their categorical nature: do not run the red light, period. The question is, for Kant, where does the binding force of such law come from? Mostly from habit, we could say, but habit does not bind neither does fear of punishment or a prospect of happiness – all of these things are contingent, morality must be binding and therefore must be necessary.
When I sit at the red light and I really have to go to the bathroom and my house is right there and there are no cars around, I think about just running the light, my maxim is something like this: “It is ok to break the law in certain circumstances, even though I accept the authority of the law, in this situation I believe that it is ok to break it” – maxim as a principle of volition does not concern itself with actual circumstances as such, at least for Kant, it’s always abstract. Therefore what is then tested (whichever test you like, Kant has three, universalization being one of them) is not so much “if everyone runs the red light, what kind of world would it be?” but a principle of volition which in this case would be a principle that one both affirms the validity of the traffic law and finds it necessary to amend it in certain personal circumstances, i.e. one both thinks that the law is necessary and contingent – Kant thinks that one cannot will such a maxim because it is a contradiction (a point of contention, of course, as it all depends on what one means by “willing a maxim”).
As for poker, since it’s a game, one does not really lie because there’s an expectation that one will attempt to deceive one’s opponents. I think Kantian ethics sleeps soundly at night knowing there are people playing poker. In fact, to suggest that lying in poker is morally evil is to misrepresent Kant as some sort of uptight Christian. Poker deals with a very superficial arrangement where all partying agree to certain rules, one being that I will try to fool others into thinking I have something I might or might not have. I don’t see any problem with poker. Let me stop here, although I intended to do an explanatory post, I think that it’s really something that introductory books are for and I’d rather address specific issues, as I am not really that well-versed in the literature as to pretend to be a kind of introduction machine, even if it helps me a lot to think about these things. In other words, I’d like to assume some sort of familiarity with Kant, and of course my initial interest was not so much in Kantian ethics, but in normativity in general….