volveré y seré millones
Indeed, quality demos are the result of dedication, aside from engine capabilities, but as far as comparisons go it's all hard to measure, and the effect varies with the circumstances. A one in a million chance may be missed by not seeing something during one instant, with low resolution, for example. Such a setback can mean the difference between achieving a hard-to-obtain record or failing altogether in the long run. A subjective and vague "I didn't notice much difference" doesn't address these tiny but potentially telling differences.
But if you don't have good moving and aiming skills, reaction and stamina, these tools won't help you to record serious demos.
I play public CTF on ZDaemon with low res and a vanilla compatible key setup and I can tell I'm kind of "shortsighted" in respect to many other players, as they become clearly "pixelated and blurry" with a reasonable distance. This doesn't mean one can't achieve much with such settings; I believe Chewy, one of the best CTF players in the community, used 320x200. Key arrangements make a bit more of a difference, and being able to have all the important weapons close to one's fingers can be the difference between life and death.
There's also the fact that it's not possible to know what the player is using unless its all written down on the demo text file or somewhere. There is already some variance in this respect in vanilla, and the habit of writing key and mouse setups was not unknown in the vanilla era (such as in Compet-n profiles and occasional text files) but it gets more extreme with engines that keep adding possibilities.
After all, PrBoom can do whatever Doom can do. I use it in a Doom compatible way in respect to key setup and resolution. That is, a vanilla demo is within the range of what is expected when using PrBoom. Vanilla behavior, down to player input, is an option in PrBoom. This isn't generally valid the other way around, as PrBoom does things vanilla can't.
For example, kimo said something like "I thought h214 1:02 was beyond keyboard time"... it's still not clear how possible it is without key setup arrangements not normally available in vanilla, at least. That's in comparing keyboard-only setups, because the mouse is a bonus regardless of other features or settings, of course.
What I like about the idea of Compet-n is that it allows a pretty objective and strict "base" for determining validity, going back to the beginning, much like in sports when the weight, balance and make-up of rackets, shoes, balls or other implements are established, which is all more or less beyond debate and interpretation. These change in sports too, with time, but people can bring up the differences, and question historical comparisons. So you can do the same around PrBoom-plus, but then the standard becomes centered around it. It's slightly different and while the slightly (I've called it mild-TAS versus actual TAS, from a vanilla perspective) can be emphasized, it should still be valid as long as there are people interested in vanilla-specific playing or want to have a perspective in comparison to when vanilla was used more regularly.
None of this should demerit the value of well-worked demos. It's just something to consider for more objective comparative purposes. "It's not a valid vanilla demo" doesn't mean it isn't great.
Aside from engine differences, we shouldn't also forget the cumulative understanding of technical details that allow us to improve times or make the right movement choices, both from playing over the years and the open examination of demo data, demo route precedents and the source code. This part of skill is a community achievement and cannot be expected from most demos from years back.
To really compete with Anders Johnsen's demos in a "truly fair" way, one needs to be propelled back in time and to forget the newer tools and the things we learned in 10 years or more.
PS: I've split this off because the posts regarding differences are many and long.