Allow me to begin with two circumstances of roleplay, the first of which has come up four times in the last two years of my campaign, and the second that never comes up for my players, for good reason.
The first is interrogation and torture. The party is smart enough to realize that killing everything is not always a good idea, that outside a lair they're about to enter it's a good idea to subdue a scout, tie them to a tree and interrogate them for information.
In fiction, this can be a tense, difficult moment. The victim does not want to give in, but the interrogators must have the information. It's a physical/mental struggle of the first order. What will it take to break the victim's will? How does the victim feed information to the captors, and what can the captors afford to believe? There are lies on both sides, tactics, techniques, deceptions ... what will ultimately work? How much can we torture and still retain our self-respect?
In D&D, there's nothing. Why not take out the knife, stab the victim at once, swirl the knife into the victim's intestines, wash our hands in the intestines, eat the victim's kidney while the victim watches while keeping the victim alive with cure serious wound spells, etc., etc. There's no emotion here. And if the victim gives out information, we can know with another spell if its false. And then kill him without any remorse. Whether the victim is cooperative or not. What value is there in letting the goblin live? It's a goblin. Get the info, stab it and let's get going.
The problem is that there's no emotional resistance to overcome. As roleplaying, it's just words. Now, if I could play it so that every time a player gave orders to stab the creature, I could throw an actual bucket of pig blood on the player, that might create the kind of discomfort/nausea inducing reality that actual torture releases. Stab him again, another bucket of pig blood. Some players would get to like it.
Then again, reverse the situation. Interrogate the player and the answer will be (if the information is actually important), "I'll die first, go on, kill me, I'd rather roll up another character." Either that or the player sings at once.
In either direction, it feels stupid in the session for A) the players and DM to be saying to each other, "I hit him again," since there's no connection whatsoever to the actual hits; and B) people sitting around a table pretending to scream in pain. (In a way, making the scream does make some people uncomfortable ... but I play with at least one Dominatrix who likes it when I scream). Everyone feels like an idiot and the whole situation feels like a farce.
Most IMech solutions have as their focus only one thing: does the prisoner blab? Roll the dice and the result will or won't give you the information. But it won't be very interesting as the conflict is resolved with very little emotion. There's no tension.
Let's leave that on the shelf for a moment and talk about circumstance number 2: gameplaying within the game.
The DMG gives a page over to the playing of different dice and card games, as has the occasional book put out by the 3rd Edition universe. I don't know if 4e has, but I'd think it probably does. It always seems like a good idea - to have players gamble with their coins, to place bets with one another and either lose or accumulate income this way.
Problem is, it's all boring.
One doesn't have to get very old before playing cards for matchsticks or meaningless poker chips becomes pretty dull. I still play this way occasionally with people too uncomfortable to play for money, but I don't care if I lose, it doesn't mean anything and my heart rate doesn't climb above normal as the pot gets bigger and bigger. The best game of this variety that I like to play is Rumoli, basically because it's a counting game that requires resource management along with playing poker.
Somehow, the reality that losing hit points will mean that a player will never again be able to run this character in my world is enough to create tension and fear. But losing thousands of gold coins at the casino hasn't got that kind of emotional pull ... characters have most of the gadgets they need already to return to the lair again and at least pick up food money. Food isn't very expensive. A character can live very cheap if they need to (they feel no hunger, lack of companionship, boredom or angst). If they win, its just another big pile of money, which usually just gets plowed into something else that doesn't really matter, like a storage place for their stuff (house), or a bigger storage place (castle).
So, players will either A) not care about gambling, or B) not care about losing. The IMech isn't enough to create the tension of loss, since the loss of money does not carry with it the personal discomfort for characters that we are all familiar with.
The IMech solutions I read through yesterday, including those that were linked, are all ultimately based on a two-sided result: what you want happens, or it doesn't. If it happens, you get the information, if it doesn't, no big deal. Maybe you can get the information some other way. The highly railroaded argument presented by this site, posted by Anthony, argues that you've got to provide three ways to give the player a chance at understanding this information ... as though this won't be cottoned-onto by the players, who will soon be sitting in your world responding to your clues with, "I don't get it either, but the next one will tell us what we need to know. We'll wait for it. No need to think."
Either way, we're still NOT talking about emotional involvement. And this is the key. Combat is more than just rolling dice and getting a resolution. Combat is not about information gathering. And just so I don't forget to say this with emphasis, any IMech that is based on information-gathering will suck. It will suck hard.
Let's look at combat from two sides ... the information-gathering side, and, well, the side that addresses the point I'm moving towards.
If I swing a weapon, I am rolling the dice to see if I hit. This is information gathering. The weapon's affect is calculated against the enemy's resources, and an effect is achieved. This effect is conveyed to me and again, I receive information. I then make an assessment of the situation on the information I've received and decide what I'm going to do.
If combat was not two-sided ... if it was just me rolling the dice to see if I hit, it would be awfully boring. Imagine a shooting game in D&D that had the player standing in front of a target trying to hit that target by rolling dice. Dullsville. But this is what IMech information-gathering is ... one-sided rolls to determine results. The clues do not roll dice against you. And if you miss, all you are is ignorant. No penalty is incurred.
This is the other side I'm getting at. Combat has enemies, who are trying to kill you. There's more at stake than getting information ... you know that if you FAIL to hit, it will allow the other side to try. That could mean your death. That could mean a very lucky roll for your enemy that could smash you for twenty-two points that would be devastating. And that's the chance you take every time you miss killing him. A chance that comes with every die roll. It only takes one to kill you dead.
You understand ... there are consequences. There's more than what you're able to gain, there's what you might lose.
In the gambling problem above, I pointed out that the loss isn't very significant. It's only money. Neither is the gain, for the same reason. Last night, talking about this, I was proposing an ad hoc solution: suppose I laid out the following set of rule:
1) You are able to gamble once per session; you may play for a period not exceeding one hour.
2) You must declare your stake in g.p., that being the number of gold pieces you have on hand at the start of the game. There is no limit on the size of your stake.
3) When your stake is gone, or the hour has passed, you must stop playing for that session.
4) Wins are not assessed until you declare you are stopping.
5) When you declare that you are stopping, for every 10 g.p. you have more than the amount of your stake at the start of the session, you will receive 1 x.p.
6) Gambling houses may have limits on the size of your bets, or ultimately a limit on the size of your total winnings (depending on the size of the city you are in), but beyond this there is no limit on winning.
This should make anyone deeply indoctrinated into the game take pause. There's no specific rule against this sort of thing, but it is generally conceded that players shouldn't get experience just for sitting there gambling. If I start with a stake of 10,000 g.p., and I choose to play in my world's Monte Carlo, I can sit and play roulette at 35:1 odds. If the bet limit is 1,000 g.p. per spin (and it would be, at least), with ten spins I have pretty good odds of hitting for 35,000 g.p. - 2,600 X.P.! Minimum, since I'm assuming there that the last spin is the winner.
Maybe not very much for a character that has 10,000 g.p., since they're at least fifth or sixth level, but still ... a lucky player could win enough to push them from fifth to sixth, if they had a good run.
Scoff if you will, but ask your players if they would play, if you were serious. Don't listen to what their mouths say - watch their eyes as you tell them. Even as the gentle readers right now are shaking they're heads, they're wondering if it's really all that far out there.
Beyond the point that I wouldn't do it, the real problem is that there's no punishment for loss. There needs to be. So let me add this, and really upset the ethical community:
7) For every 100 g.p. you declare as your stake, you must put $1 on the table.
8) If you lose that 100 g.p., the dollar goes in the DM's pocket. Money won back up to the amount of your original stake will be refunded at the end of playing, but the DM is not required to refund the difference between your original stake and your present g.p.
Now, I don't know about you, but I've had players - joking and serious - who have offered to pay me money for experience. It is the dark, seedy side of the game that no one talks about. Business that can only be conducted in the dark hallways between the bathroom and the gaming table. Let me state very clearly that I have scruples, I don't need the money that badly and that, if it isn't clear, I'm not suggesting that any of this be done.
But I am trying to make a point about REAL gains and losses. Experience levels and death are real sides of a coin, much more meaningful than information/no information. An IMech, on some level, must have characteristics that punish the characters for failure.
For months I've been thinking about a post on how the game of D&D is gambling, or rolling dice to get a result. In gambling, you lose. Without loss, there is no gambling.
Let's hope you're still with me at this point, because I still have points to make. Bored yet? Good, we can move forward.
Like the above never to be implemented example, an IMech which is added to the game is going to change aspects of D&D that will be uncomfortable. There is a finite number of existing numbered characteristics acribed to the characters - all of which are held to be sacrosanct - and therefore a limited number of things that can be reasonably changed. Steelcaress makes reference in a comment on the previous post about inventing a new kind of 'hit points,' but I really hate that idea. It lacks creativity, it divorces the whole IMech issue from other aspects of the character and it really reduces the verisimilitude of the familiar stats that have developed meaning over the years in favor of something that feels 'tacked on' and therefore tacky.
In playing the wins and losses of IMech, the stakes should be things that matter. Having some other scale of wins and losses only matters if, when the points are reduced to zero, something happens. Something as permanent and distasteful as death. Something that makes players stop and panic.
Let's return to the torture scenario, and let's say that our prisoner doesn't want to talk. Let's say the success/failure roll indicates that the player WON'T talk. So the player says, "I stab him in the belly with the dagger, and the cleric will stand by if it goes hard with a death's door spell."
(As an aside ... your torturers kill you, actually kill you, just slightly dropping you into the death state, then pull you back from that door to zero h.p. Every ... damn ... day. Pretty nasty, that. Who wouldn't talk?)
In the old system, if the dagger stab doesn't work, we just stab the prisoner again. But suppose you add a wrinkle, and explain it to the player: "If you coldly draw blood against a helpless victim, there's a 1 in 10 chance of losing a point of charisma. Permanently."
Watch how quickly how the prisoner's information seems less important.
But if that seems unfair, or drastic, or problematic for assassins, consider this alternative, as I wrap things up:
Reputation. Assume that anything that happens on account of the party's 'failure' to force information will make its way to the ears of somebody ... if through no other source, than by a failed wisdom-check by the party's dumbest member, who saw the torture (or failed attempt to talk to the king, or failed effort to find the murderer in this town) and just felt the need to talk about it to the local bartender.
What if the players get to be known at those annoying, pestering, supposed do-gooders that can't seem to recognize a goddamned clue if their lives depended on it? What if the players want to be Sherlock Holmes, but they've blown their rolls so often that now everyone knows they're just LeStrade? Are you telling me that won't create tension?
This was, more or less, the destination I was going to reach for with the charisma-related post I was going to write yesterday. I've gotten there by an entirely different means, but the point is still the point. It isn't enough that there's a roll that must be beaten in order to achieve a success. There must also be a roll that must be beaten in order to avoid failure.