A demonstration of the flawed timer display in the bottom right of the screen is shown in the following video:

"https://www.dropbox.com/s/m5g0t327ary97mk/EP%20Timer%20Glitch%20Leaderboard%20comparison.mp4?dl=0"

The top 4 submissions of the leaderboard are lined-up and synced to the same starting point. Over the course of 2m25s one can find spots where the timer easly differs by 0,6s up to a full second.
The error margin can be influenced by run duration or lag (in Multiplayer) or other. That was the experience among many players back in the days who compared times, routes, and captured their runs. I mentioned it briefly when my run got evaluated.

Part of the really close checkup is A ) only recently did I acquire editing software which makes it much easier to show, B ) when checking out the recent ExitPath html-version, out of curiosity, some minor things felt weird.

So upon looking into the Github-history I saw the "Optimize level loading performance". Now I don't know if that impacts Uniplayer or not, and the Video comparison I did here does not show any of that in my opinion.
A possible optimization is surely nice, but also a really dangerous thing to implement, bcz if it is improving the transition between levels in Uniplayer the effect would be amplified by 29 and certainly impact recent runs depending on the version they are played on or not.

If I can help out, let me know.
Thanks and take care!

Thanks for posting this.

I don't fully understand how the in game timer works, but my guess is that the inconsistencies come from the level transition screens, as you can pause and see that right after level 1 the timers are not synced up anymore.

I think it makes more sense to use in game time because if we switched over to real time then slower computers could potentially be penalized. Having the runs be half a second different in real time but still consistent in in game time isn't to big of a problem.