I think that actually the relevant distinction here isn't between "games where the fiction influences action resolution" and "games where it doesn't," since as has been stated pretty much all the games we've been looking at (including both 4e and DitV) have the "fiction influences action resolution" in some way. Even if it doesn't appear that it is, you can always make it that way just by defining the mechanic in question to be part of the fiction. (For instance, in the case of hit points, you can say that biology in the game world works by "vital energy" or whatever, and the amount of vital energy a given being has left can be measured*, and "vital energy" is generated or removed by the same things that add or subtract hit points in the game. Now hit points are part of the fiction.)
The relevant distinction instead seems to be between systems in which the relationship between fiction and action resolution is governed by explicit rules and systems in which the relationship is governed by DM judgement. For example, in 4e, there are explicit rules saying how each fictional element (like terrain, positioning, weapons, etc.) affects action resolution. However, consider FATE, a system where your character is mainly based on "Aspects", which can be any word or phrase you want, and you can get bonuses by invoking your aspects if they are relevant to the situation. Fiction still has an impact on action resolution (because how you describe your action affects which Aspects you can invoke) but that impact is based entirely on DM judgement (because the DM is the one that decides if an Aspect is applicable to the situation.)
* There is actually rules support for this idea. There is a magic item called "Goggles of Aura Sight" (I don't remember which book it is from but it is in the CB) that tells you how many hit points a target has. Presumably, then, the goggles must measure hit points in some way.
Now, each system has advantages and disadvantages:
Advantages of Rule-Based Systems over Judgement-Based Systems:
1a. The mechanical descriptions inherent in rule-based systems can be easier to interpret than the "fictional" descriptions inherent in judgement-based systems. For example, "The giant stands nearly thirty feet tall, has bulky but nearly impenetrable armor, and radiates an energy field which seems to repel magic, although he moves relatively slowly" can be immersive, but "The giant has 310 hit points, has an AC of 34 and Reflex of 26, and gets +4 to defenses against attacks of the arcane power source, and has a speed of 4" is a lot more useful for the players in terms of figuring out tactics and understanding what can affect it.
2a. In some cases, having to think about rules can actually increase immersion. For example, I was in a superhero game once using the HERO System where I was a "gadgeteer", and I had the ability to alter gadgets on demand. This was implemented in game by a "variable power pool" where I could create powers up to a certain number of points. So in the middle of the game I would be figuring out what I could do with the amount of points I had and how to use them to get past the situation I was in, taking into account factors like how much endurance the power used and how long I could sustain it. But of course figuring out things like design limitations and power consumption is part of what real engineers do, so having to do that in the game actually added to the immersion, not subtracted from it.
3a. Prioritizing the rules over the fiction can increase consistency, because everyone can agree on what the rules say but people can have different conceptions on what is "fictionally" justifiable. (Of course, it's possible that the rules themselves are unclear, but in that case it's a problem with the rules that can be fixed. With the judgement-based system, the possibility of different judgements is the core of the system.)
4a. In judgement-based systems, it's easier for one player to feel left out if the "judgements" consistently go the wrong way for them - take the example given before of the rogue who can't use his sneak attack because the DM always says that monsters don't have vital spots or aren't distracted enough. Again, it's true this can also happen in rule-based systems (if, say, in 3.5e a rogue is going through a dungeon full of undead) but it can be especially insidious in judgement-based systems because it's easy to do it unintentionally. For example, if it's easier to find "fictional objections" to fighter powers than to wizard powers, then fighters are going to be disadvantaged compared to wizards.
5a. Having rules may help players understand the fiction. For instance, I've never been in a real-life swordfight (well, I've done boffer fighting, but from what I've heard boffer fighting is not a good approximation of a real swordfight), and I've watched maybe 3-5 movies in my life that involved swordfighting (unless Star Wars counts) so I don't have a good sense of what is possible in a swordfight either realistically or "fictionally." And of course there are some elements of the fiction that are made up specifically for D+D (like some of the magic spells) so there is no way to have any external reference frame as to what is "fictionally" possible. So having the rules helps me understand what is "fictionally" possible.
Advantages of Judgement-Based Systems over Rule-Based Systems:
1b. Some of the disadvantages of judgement-based systems listed above may actually be advantages depending on your style of play. For instance, you may want players to only have vague information about their capabilities and threats and have to figure them out by trial and error, or you may want certain players to be useful in fewer situations.
2b. With rule-based systems, only the specific fictional elements listed in the rules matter. With judgement-based systems, any element that you want matters.
3b. Judgement-based systems can encourage players to be more creative, since in order to gain advantages players have to "impress the GM" in some way.
4b. Judgement-based systems make it so players always have to be thinking about the fiction. In contrast, in rule-based systems you can just focus on the rules and still get a good game. (For instance, consider Magic: The Gathering. A lot of effort goes into creating the fiction for the game and designing mechanics that capture the fiction. For example, there's a creature type called "allies" that each gain benefits when you put another ally onto the field, so if you have allies in your deck you want lots of them, which captures the idea of "creatures that are good at working together." And when talking about tne game and building decks players sometimes take this into account - "the next set is going back to the plane of Mirrodin, so it's probably going to have such-and-such kind of creatures." But during a game itself players rarely think about the fiction at all. In fact I am thinking about writing another post specifically focused on Magic as a case study of the relationship between fiction and mechanics in games.)
---
Now, the key question from a design perspective is how to combine the advantages of both styles while minimizing the disadvantages. Some systems (like 4e) deal with this by having a rule-based core, but then having possible add-ons (like page 42) that you can use to go into the judgement-based mode when the desire arises. The problem with this is that you have to have encouragement to go into the judgement-based mode. For instance in 4e, if players' powers do everything they want to do, then players won't need to use page 42.
Another way to do it is to have a judgement-based system as a core, but have clear guidance on how to do the judgements to minimize the problems associated with the subjectivity. Some possibilities:
1. Limit the amount that is riding on one judgement. For instance, consider the example given before of a DM who rules that a party killed themselves by casting a fireball in a wooden building, causing the building to collaps on them. One way to stop this would be to have a rule that limits the damage from one "judgement," so that the players could get hurt but still have a chance. An example of this would be the damage tables on page 42, where even at the "high limited damage" it's not enough to kill a character in one shot.
2. Define the question that you have to judge the answer to in a way that's less controversial. For instance, a question of the form "How likely is Action X to succeed?" or equivalently "How difficult is Action X to perform?" is going to be answered very differently by different players. However, the question "Is Attribute Y useful to perform Action X?" is a relatively simple yes-or-no question, and the variation in answers will be less. An example of this is FATE's system for aspects, where the only relevant question is "Is the aspect useful?" and if so then the rules do the rest of the work (you can spend a fate point to tag it for a +2 bonus).
3. Allow players to ask questions about how the DM would rule ("If I do X, will that count as a good enough description to activate Power Y?") This seems reasonable because the character would presumably have training, experience, and knowledge that the player doesn't, and this is representing that. You might have some sort of roll involved to represent whether the character knows.
4. Keep track of how often each player is getting boosted or hindered by the fiction. An example is in FATE, where you get "fate points" each time an aspect hinders you or provides an obstacle to the team, while you have to spend a fate point to use your aspect. So if one player seems to have a big hoard of unspent fate points, chances are he is being hindered a lot and not getting much chance to use his aspects. Of course this doesn't necessarily mean that "Everyone being equally useful" is an ideal - the player still has to find ways to use his aspects - but it does give you an alert that you might want to look at that player more closely.
The relevant distinction instead seems to be between systems in which the relationship between fiction and action resolution is governed by explicit rules and systems in which the relationship is governed by DM judgement. For example, in 4e, there are explicit rules saying how each fictional element (like terrain, positioning, weapons, etc.) affects action resolution. However, consider FATE, a system where your character is mainly based on "Aspects", which can be any word or phrase you want, and you can get bonuses by invoking your aspects if they are relevant to the situation. Fiction still has an impact on action resolution (because how you describe your action affects which Aspects you can invoke) but that impact is based entirely on DM judgement (because the DM is the one that decides if an Aspect is applicable to the situation.)
* There is actually rules support for this idea. There is a magic item called "Goggles of Aura Sight" (I don't remember which book it is from but it is in the CB) that tells you how many hit points a target has. Presumably, then, the goggles must measure hit points in some way.
Now, each system has advantages and disadvantages:
Advantages of Rule-Based Systems over Judgement-Based Systems:
1a. The mechanical descriptions inherent in rule-based systems can be easier to interpret than the "fictional" descriptions inherent in judgement-based systems. For example, "The giant stands nearly thirty feet tall, has bulky but nearly impenetrable armor, and radiates an energy field which seems to repel magic, although he moves relatively slowly" can be immersive, but "The giant has 310 hit points, has an AC of 34 and Reflex of 26, and gets +4 to defenses against attacks of the arcane power source, and has a speed of 4" is a lot more useful for the players in terms of figuring out tactics and understanding what can affect it.
2a. In some cases, having to think about rules can actually increase immersion. For example, I was in a superhero game once using the HERO System where I was a "gadgeteer", and I had the ability to alter gadgets on demand. This was implemented in game by a "variable power pool" where I could create powers up to a certain number of points. So in the middle of the game I would be figuring out what I could do with the amount of points I had and how to use them to get past the situation I was in, taking into account factors like how much endurance the power used and how long I could sustain it. But of course figuring out things like design limitations and power consumption is part of what real engineers do, so having to do that in the game actually added to the immersion, not subtracted from it.
3a. Prioritizing the rules over the fiction can increase consistency, because everyone can agree on what the rules say but people can have different conceptions on what is "fictionally" justifiable. (Of course, it's possible that the rules themselves are unclear, but in that case it's a problem with the rules that can be fixed. With the judgement-based system, the possibility of different judgements is the core of the system.)
4a. In judgement-based systems, it's easier for one player to feel left out if the "judgements" consistently go the wrong way for them - take the example given before of the rogue who can't use his sneak attack because the DM always says that monsters don't have vital spots or aren't distracted enough. Again, it's true this can also happen in rule-based systems (if, say, in 3.5e a rogue is going through a dungeon full of undead) but it can be especially insidious in judgement-based systems because it's easy to do it unintentionally. For example, if it's easier to find "fictional objections" to fighter powers than to wizard powers, then fighters are going to be disadvantaged compared to wizards.
5a. Having rules may help players understand the fiction. For instance, I've never been in a real-life swordfight (well, I've done boffer fighting, but from what I've heard boffer fighting is not a good approximation of a real swordfight), and I've watched maybe 3-5 movies in my life that involved swordfighting (unless Star Wars counts) so I don't have a good sense of what is possible in a swordfight either realistically or "fictionally." And of course there are some elements of the fiction that are made up specifically for D+D (like some of the magic spells) so there is no way to have any external reference frame as to what is "fictionally" possible. So having the rules helps me understand what is "fictionally" possible.
Advantages of Judgement-Based Systems over Rule-Based Systems:
1b. Some of the disadvantages of judgement-based systems listed above may actually be advantages depending on your style of play. For instance, you may want players to only have vague information about their capabilities and threats and have to figure them out by trial and error, or you may want certain players to be useful in fewer situations.
2b. With rule-based systems, only the specific fictional elements listed in the rules matter. With judgement-based systems, any element that you want matters.
3b. Judgement-based systems can encourage players to be more creative, since in order to gain advantages players have to "impress the GM" in some way.
4b. Judgement-based systems make it so players always have to be thinking about the fiction. In contrast, in rule-based systems you can just focus on the rules and still get a good game. (For instance, consider Magic: The Gathering. A lot of effort goes into creating the fiction for the game and designing mechanics that capture the fiction. For example, there's a creature type called "allies" that each gain benefits when you put another ally onto the field, so if you have allies in your deck you want lots of them, which captures the idea of "creatures that are good at working together." And when talking about tne game and building decks players sometimes take this into account - "the next set is going back to the plane of Mirrodin, so it's probably going to have such-and-such kind of creatures." But during a game itself players rarely think about the fiction at all. In fact I am thinking about writing another post specifically focused on Magic as a case study of the relationship between fiction and mechanics in games.)
---
Now, the key question from a design perspective is how to combine the advantages of both styles while minimizing the disadvantages. Some systems (like 4e) deal with this by having a rule-based core, but then having possible add-ons (like page 42) that you can use to go into the judgement-based mode when the desire arises. The problem with this is that you have to have encouragement to go into the judgement-based mode. For instance in 4e, if players' powers do everything they want to do, then players won't need to use page 42.
Another way to do it is to have a judgement-based system as a core, but have clear guidance on how to do the judgements to minimize the problems associated with the subjectivity. Some possibilities:
1. Limit the amount that is riding on one judgement. For instance, consider the example given before of a DM who rules that a party killed themselves by casting a fireball in a wooden building, causing the building to collaps on them. One way to stop this would be to have a rule that limits the damage from one "judgement," so that the players could get hurt but still have a chance. An example of this would be the damage tables on page 42, where even at the "high limited damage" it's not enough to kill a character in one shot.
2. Define the question that you have to judge the answer to in a way that's less controversial. For instance, a question of the form "How likely is Action X to succeed?" or equivalently "How difficult is Action X to perform?" is going to be answered very differently by different players. However, the question "Is Attribute Y useful to perform Action X?" is a relatively simple yes-or-no question, and the variation in answers will be less. An example of this is FATE's system for aspects, where the only relevant question is "Is the aspect useful?" and if so then the rules do the rest of the work (you can spend a fate point to tag it for a +2 bonus).
3. Allow players to ask questions about how the DM would rule ("If I do X, will that count as a good enough description to activate Power Y?") This seems reasonable because the character would presumably have training, experience, and knowledge that the player doesn't, and this is representing that. You might have some sort of roll involved to represent whether the character knows.
4. Keep track of how often each player is getting boosted or hindered by the fiction. An example is in FATE, where you get "fate points" each time an aspect hinders you or provides an obstacle to the team, while you have to spend a fate point to use your aspect. So if one player seems to have a big hoard of unspent fate points, chances are he is being hindered a lot and not getting much chance to use his aspects. Of course this doesn't necessarily mean that "Everyone being equally useful" is an ideal - the player still has to find ways to use his aspects - but it does give you an alert that you might want to look at that player more closely.