Thanks for the info! And yes, I meant Death Incarnate difficulty. If you're curious, for my E2M8 walkthrough I just showed the Aardwolf secret then exited the level.
I've been practicing Episode 1 some more, and today I realized that the game timer isn't always accurate and should not always be used for speed comparisons. Much like enemy aggression and the speed of the Hitler ghosts' fireballs, I think the game clock may be tied to CPU cycles somehow.
I have a timer program that I use for my speedruns, and noticed some discrepancies with roymustang18's episode runs on Youtube. I begin timing once the "Get Psyched!" meter is full, and from there the following occurs:
- Using this timing, I did a run of the first level which took 29 seconds in real time and read 0:28 on the game clock. This makes sense, since the game timer doesn't begin until the level actually begins; I begin timing where I do just for convenience.
- Using the same starting point on roymustang's level 1, he took 35 seconds in real time but got a game time of 0:32!
The game is supposed to run at 6000 CPU cycles, but my suspicion here is that roymustang played at 3000 cycles for all of his runs. 3000 is the default setting on DOSBox, and the enemies are less aggressive in his runs (which has been proven to result from lowering the cycles). I noticed a number of levels where he seemed to play worse than me yet get faster game times, and this is the most likely explanation.
With that in mind, 6000 cycles really should be a rule for any speedrun, as the default DOSBox settings make the game much easier and likely mess up the game clock. Using "real time" and "game time" as separate categories would be interesting, as some secret levels actually reduce the latter but not the former (e.g. episodes 2 & 6, Spear level 19).