A new game, a new member. Realm by Apemantus.

[greyorm]

Well do YOU need to know your own 'stats' in order to act in real life?

Yes, I do in fact. If I have no idea how strong I am, I cannot -- in any reasonable fashion -- determine what my course of action should be in a variety of given situations. Yes, I cannot attach a number to my "Strength attribute", but I know the generalities. "I am strong enough to do X" or "I am not strong enough to do Y."

I KNOW whether or not I can pick up a sword and swing it around for a bit without becoming fatigued (in my case, I know I can't because I'm not strong enough. I have had friends, however, who could whip a claymore around without problem), because I know how strong I am.

The same goes for skills. Even very young people know what sorts of skills they have and how good they are at them, in general. Say I design a character whose background includes "wizard's apprentice" and I describe in that his voracious appetite for the written word. Later on, I find you have given him either an average or limited skill at reading and writing, or for various knowledge skills.

As you can see, unless you are creating completely blank slates (ie: babies), this is not going to work. What sorts of skills YOU decide the characters have is going to conflict with the concepts developed by the players regarding who their characters are and what they can do.

Now, I'm going to play pop-psychologist for a minute: It seems to me you are really fighting against the above truth because you want to do your game the way you want to do it, and you feel you need the support of the argument you have given. Showing the argument is flawed removes that support for you, so you reject the argument. But you're rejecting the response being given to your idea for the wrong reasons: you WANT your statement to be true because it allows you to do a thing in a certain way, regardless of whether or not your statement is in fact true.

That's a bad way to design a game (and more, but I desist), because if you insist on using a false argument to support your design, people will reject your design. Now, you can still design your game using the method you have described above, you just can't use the current argument to support your decision and expect anyone to accept it as a valid design choice.

[Certified]

Maybe there is a balance between the two perspectives. If the game begins with player and GM building the character's background story then the GM might be able to say, You're Strong and tough, but not so quick, yeah, you did get into a lot of fights in the schoolyard, some you won, but not many, you preferred to say in and study, learning all your school had to offer on subjects X, Y and Z. Now the player knows, he's not a wimp but he is more studied than an outdoors type. I've used the tactic of no character sheets to build a beginning for games, it's a good way to set the mood and tone, but can become overwhelming as a GM quickly.

It's a trick that can be done in any system, but simpler works better since it's less work behind the scenes, at least from my experience. My personal opinion on the subject is it's a good way to start things but once play begins in full better to hand back the character sheets. However, I've noticed Henry mentioned adding in two new support roles for handling the number crunching, which sounds like a really good idea for achieving the goal of blind gaming. I would recommend a "Styles of Play" section, detailing the Hidden Character Sheet Method as well as Tradition lay out and maybe a mixed middle. Thank you for reading though this post. 

[knicknevin]

And the combat system needs to be complex, even out of the sight of the PC's, because I do NOT want decisions to be arbitrary--Arbitrary decisions are always the most mundane, of course.

OK, thats a bit of a jump there from 'not complex' to 'arbitary'... I've played in plenty of games with simple rules that do not come to arbitary conclusions, e.g. Savage Worlds, which has a very basic system of target numbers & rolls, but still manages to create a variety of useful outcomes; and I've played in games with evry complex rules that the GM did'nt tell em about and I just found it incredibly frustrating, for example:
Me: "I do X"
GM: "You can't do X"
Me: "But I've done X before!"
GM: "Ah, but you'd done Y first that time"
Me: "Well, can I do Y then?"
GM: "No, you don't have enough Z points"
Me: (Quietly slumping in my chair and entering a state of apathy where I let the Gm just do whatever they want until I can get out of here)

Finally, "Arbitrary decisions are always the most mundane, of course"? Thats a VERY sweeping statement and, to be honest, its decisions which stem from complex and mechanisic rule sets that I have generally found to be mundane; simply winging it on a handful of rules has always encouraged my creativity and that of the people I've played with and you have stated yourself

My argument is that not having direct access to probabilities will introduce a new kind of drama in character actions, and that uncertainty in general makes things more prone to cause fantasy in the mind, to make play more profound and exciting.

OK, so why not have 'uncertainty' for the GM too? Why pin them down with a very exact & complex rules set, instead of letting their imagination flourish like the players?

I still like the idea here, but I'm not sure you'r two stated goals are achievable in the same game: hidden statistics seems more suited to a mystery/investigation/espionage-style game, whereas having strategic depth & detail requires that the players have a strong idea of what they can accomplish: do you really want to come up with 20 great combat maneuvers and then only have 3 of them ever get used, because the players have found that they work and simply never think to try most of the others?

[Sven Seeland]

Hm, this may be a bit general and rambling but I hope it helps nevertheless...

First I can suggest some more good reads:
Universalis for a game that creates an incredible amount of uncertainty with simple rules and without any dice or GMs at all.
Dogs in the Vineyard for a combat mechanic that creates highly exciting and thrilling combat (or rather "conflicts") with a minimum amount of rules.

I'm not saying you should do what those games did but these games have changed the way I think about RPGs and they might just tear down some assumptions for you.

Let me just inject some statements. I don't claim that these are absolute thruths, though they are insights that have helped me greatly. For you they are mainly meant to provoke some thoughts:
1) Complex rules are not better rules. Games can be tactical, exciting and interesting with simple rules, too.
2) Complex rules are not bad just because they are complex. They are especially neat if mastering the rules is one of the challenges in the game ("Who can use the rules to their advantage the best?")
3) You don't need to be totally immersed to create a good story (or a game that focuses on story)
4) You can trust your players. With the "right" distribution of "power", arbitrary choices can lead to very, very fun play (see Universalis, for example).

Maybe you should familiarize yourself with what we call "naratavist" games around here. I'm not sure that's the style of play that you want but you sound a bit like it. Two I've already mentioned (Dogs and Universalis). Burning Wheel falls into that category as well, as far as I know and I believe The Riddle of Steel could be interesting for you as well, though I can't say for sure since haven't played or read it.

Again, I'm not suggesting you should do anything any particular way, I'm just trying to show some options and "free your mind" so to speak. It did wonders for me when I came here.

Best wishes,
Sven

[knicknevin]

Y'know, it's one of those things that just pops into your head after you've already said your piece, but thinking on it while walking the dog, I've now revised my opinion and think that maybe your two goals are compatible, with one proviso: the GM and the players collaborate in creating the tactical rules.

What if, instead of working out all the rules you want beforehand, you just create the very, very basics, the sort of things you learn in childhood, e.g. punching, kicking, using a stick & throwing a rock; then, players can have hidden stats, but still have an approximate idea about what they can & can't do. When anything not covered by these simple rules is attempted for the first time, e.g. trying to disarm an opponent, grapple them, aim for a weak spot, etc then the players and the GM enter into a dialogue about what the player expects the result would be and how the GM thinks that fits in with what has already been written. The player's success or failure at the task could even form part of the rule, e.g. "Well, you failed at that, so it must be very hard" or "That was very successful, so that type of attack does extra damage". Once the rule is agreed, it gets written into the rules that this grou are using; another group could have different rules variats. In this way, players can have very little knowledge about combat, reflecting the uncertainty of their characters (or is that vice versa?) and knowledge becomes a reward for creativity & experience.
Also, this would level the playing field; if the player's don't know the combat rules but the GM does, how can he give the players a fair challenge? Won't all the NPCs appear to be more capable than the PCs in combat, if the GM is using the full scope of the combat system? If he tries to play more balanced NPCs, he has to more or less feign ignorance of the combat system: tricky at best.
Anyway, sorry for posting twice before you've had a chance to reply, but I just felt that this was more of a constructive post than my previous one and I hope it helps. :-)