As I tried to parse what the poster was actually trying to argue, I realized that it sounded like they believe a good system should remove GM judgement as much as possible, such as in setting target numbers and what not. The game mechanics should, essentially, take inputs, do some processing, and spit out a result. I call this system, "Input things, turn crank, receive bacon." He later clarified that his problem is "Freedom of character concept, mechanical differentiation, rules that work without skilful GM supervision; pick two." Basically, no system in existence meets these criteria for him, and he didn't offer any insight into how such a thing could be designed.
|This is what portions of the discussion made it sound like the problem was|
While I will admit that one of the challenges of being a GM is having to make judgement calls in how to apply the rules, the alternative that follows logically (at least to me) from the poster's observations is not much better. The only way I can see it working to have everything that a GM might need to resolve a situation - the elements going in ("set inputs"), the processing of the information ("turn crank"), and getting the result ("receive bacon") - is to have a large number of dedicated, separate subsystems. Essentially, moving away from unified mechanics (which is what the poster was complaining about as "illusionism"). You could have a number of decision trees to go down with fixed results depending on the inputs, but that isn't much better and certainly not feasible in play. You could also just play a board game.
Now, I can see the appeal in wanting to differentiate between a sword fight and a wrestling match, or shooting a gun versus a bow and arrow at a target. Realistically, they are different things. But from a game design perspective - at least to me - having a unified mechanic serves exactly the purpose that the poster is looking for. It's just that he doesn't want to have to do any thinking about when and where and why to apply it - essentially, all of the things that are considered to be the hallmark of a good GM. The fact that there aren't any systems that supply this magical mechanical meat grinder isn't a failure of those systems - it's a feature, and one that many games coming to the market have learned to leverage.
Overall, I find it an intriguing exercise - setting target numbers and making sure things aren't completely out of whack from the GM's side of the screen is an acquired skill. Consistency is good, and this skill often has to be reset with every new system learned. Taking some of the onus off of the GM, as a result, sounds good on paper. But taking into consideration that the level of GM intervention with regard to interacting with the mechanics is purely personal - and yet another thing that I'm not sure that any system can effectively manage without being entirely too restrictive or cumbersome - I can't really see the idea being anything more than an exercise.