I wonder if people are envisioning things as much more complicated in this whole gametype scenario than they really need to?
What if we think of this entire thought experiment like this...
We are playing D&D. Same way we always play it. Everyone runs their characters the same way... everyone makes choices and decisions and queries with the same logical consistency as they always do (in other words, no player ever bothers asking "Can I jump to the moon?" because they know full well what the answer
would be, which is 'no', because they know what the internal logic of the game world is). Any time a player
does make a choice for which there is a chance of failure-- a time when we would ask for dice to be rolled and check to be made-- the roll
just happens to roll a success.
Every single time.
Every time there is a chance for failure, there just isn't. We don't know why this happens... and yes, this could result in a huge amount of weird coincidences or unlikely happenstances... but how does the D&D really change? Or even DOES it really change?
After all... whenever we as DMs ask for checks to be made, most of the time we are expecting that indeed there is a possibility of the check succeeding. Sometimes easily, sometimes a very small chance. And sure, we will occasionally see that rare occasion where a players asks to do something
so unlikely that we DMs will just set the DC for the action so high it is basically is an undeclared 'No' (IE setting the DC to 40 knowing full-well no one can actually roll it, so you are ostensibly telling the player 'No' without actually saying it out loud)... but more often than not the players-- when playing within the logical confines of the ruleset and physics of the campaign world they are in-- will only make choices that
could have a chance to succeed.
Knowing this... knowing that the game is played in such a way that-- odds-be-damned-- every time a player rolls to see if the choice they made did in fact work,
it does... does that really change the game all that much? Because that is what
@Reynard's thought experiment truly is representational of-- a D&D game played just like any other except every time someone rolled the dice, they won the roll. And in this particular scenario of Reynard's... we just skip ahead past the successful die roll to the DM describing the results of what would have been the successful check.
To me... the biggest stumbling block I can see for some people to really wrap their head around this idea are those DMs who do in fact play with players who are prone to making choices that ARE logically impossible to succeed with in their particular game world. Where the need to say 'No' is important because the players just can't help but "try and introduce a ray gun into Dark Sun" and other such world-breaking actions. And for those DMs, I absolutely see their need to be able to say 'No', because the players aren't willing to police themselves. But for other DMs-- ones with players who are self-policing of their choices and actions to things they think could actually be possible? Just skipping past the "successful die roll" straight to the resolution I don't think would really make the game all that different personally.