I personally feel like keyboard inputs for movement are far more precise than controller, though astral flight control feels nicer on controller. Aside from that only minimal aspects like input frequency of my keyboard is most likely higher than controller so for example ledge pull ups might be faster through more inputs per second.

Is there a reason you didn't boost more often in Chapter 4? Definitely wasted time here, I believe. Could've been a near perfect run with only very minor missing optimization and mostly mishaps in Chapter 5 and 6. I honestly have to say, Chapter 6 was far from perfect with pathing and the mishap at entrance of the pilllar room. I could also see some optimization her, like after the purple pillar sacrifice, but I can understand why you took a step back. Then there's a mistake I see many people doing: Not getting a boost from the base platform to the purple pillar at the final sprint. It wastes time climbing the path and resetting evil source, I believe. With a good aimed explosion boost you can even get right up to the purple platform again before then quickly boosting right into the last pillar. That last edge jump was as smooth as the first one. Loved it. Also you should feel free to retime the run properly to the first black frame.

Edit: With runs like this one we have to find some new tech soon or this will become microoptimizing hell. Guess I will have to get my TAS run setup earlier than expected to check the limits, haha.

Yeah, the dev said in an interview that the game engine is construct.

Yeah, I hit that one before too, haha

I personally triggered this bug even from Zyrumis position often and placed myself around the middle part of the platform. Quickly reacting when Patches knocks one back worked overall fine to not loose time.

Overall, I am somewhat shocked over the extended loads for some chapters as the game is only 300MB and I have the game installed on M2 SSDs (tried both of my M2s and even some 2.5" ssds and all had the same load times)

I adressed this topic in my first posted run, which is now obsolete. Loading times for this game are odd going by the size of assets and computation needed. It's roughly equal for all runners at this point, though I didn't analyse the newest WR runs yet. I honestly believe some timing is actually depended on micro-optimizing the run up to this point, since I've both done few runs lately myself that had the wanted sync of platforms while seeing a lot of other runners just missing a lot of micro optimization I usually do. I would need to systematically and quanitiatively review those findings tho before coming to absolute conclusions.

I also did some research into the game engine, which is most likely Construct3, at the least it's Construct, which uses a framerate based physics control.

Lastly, I also did some testing on GPU limited frame caps and came to the conclusion that the difference between 60 and 144 Frames probably isn't too big of a factor, tho anything below 60 will exponentially increase run time and screw with timings. I tried to also investigate how Construct handles frame physics simulation and what the upper cap used usually is, yet couldn't come to conclusive results. Easist solution to this whole discussion is getting in touch with Thomas Lean and getting insights into how the game is designed.

I agree that we simply have to make a majority decision on the topic. I would argue though that the section to use the "glitch" is short and the abuse of the effect of the mechanic is easily avoidable. If somebody accidentally triggers it without abusing it, then the run imo can be valid for glitchless. To again emphasize my point though, any unintended behavior as bugs, which are just unused glitches, do not qualify for the intended playable experience except the Dev would announce that the supposed bug is an intended design. Lastly those bugs could be utilized in Any% instead. Should it be useless there for a plethora of better paths, then well not everything always needs a place in speedrunning. Though at that point different glitch categories might become of interest.

I would personally say nigh anything that isn't the intended way to play should be critically evaluted, discussed and agreed on in the community whether it's a glitch or not. Replenishing Astral flight without touching the ground isn't an intended mechanic.

I mean it's a pretty common thing for indie game engines to run physics based on framerate. The most obvious indicator in the game pointing to wards this is actually the chase scene with patches before Chapter 2, imo. It just doesn't always feel the same going through there.

I am theorized about this too. Will do some testing in the future, especially when I have a more reliable way to precisely measure splits. Also might ask Thomas^2 about it

This would be quite the advancement for Any% tbh

What tool do they recommend?

Toufools Autosplitter seems quite promising, so I will look into that as soon as I find some time for it. Also let's not leave contracting to a 15y/o, no offense.

For anyone interested, running the game on Linux with Wine9.1 and using libTAS we might be able to create both a TAS and a tool to savestate emulate the game. Will give it a shot when I've got some time too.

Nice run, I see some interesting optimization ideas for myself that really makes me want to know how acceleration in the game works.

Aside from that, I think we have to define wether Patches Cutscene skip is a glitch or not. Not using it costs about 11 seconds. I am currently trying to replicate my past experiences of being able to skip Patches cutscene without dying in the hole at the start of the chapter, since I never died there in my first few runs while still being able to use the patches cutscene skip. My current theory is that I triggered some other area close to the hole while playing around with the walls in the elevator.

Tbh, I don't think it's an intended behavior, though I also wonder why we get the skip function after the first fight. I might just have to ask Thomas^2 about it.

Are you pacing each split with manual inputs? I honestly wouldn't want to pace that many. At most just each individual chapter tbh.

Looking forward to any categorys added for the future :) Fan of running all of em

Nice run, definitely some very small optimizations to be achieved for the future :)