Mario 64 Glitch Reproduction earns you $1000

[Arwin]

This caught my attention:
http://www.eurogamer.net/articles/2...ner-posts-usd1k-bounty-for-reproducing-glitch

Thinking about this as a programmer, I would think that the glitch happens when the two ceilings above have just the tiniest of room between them, so that a different ceiling is hit-detected instead. It would probably be useful to look if it is possible that if there is room between the two detected ceilings, you could draw a line to the hangable ceiling that Mario spawns above, and then, considering the low precision of calculations on that system, if it is possible that at such a large distance the spawning could happen too high, or if there is a maximum height.

Sounds like it could be insanely hard to reproduce though, if the inbetween seam is so small, which is probably why he hasn't managed it himself already.

Then I thought - there are some much smarter and more experienced games programmers here, and now I'm really curious how this would work and if the above could be right, or if there's something else going on here ...

 

[Shifty Geezer]

The test for the ceiling above would be something like a ray cast upwards. I expect the other upwards glitch is a result of the above-ceiling test happening before the move forwards, and then the vertical repositioning happening after that test. That is

1) Test if ceiling above is hangable
2) Move forwards
3) Set to hangable height

Alternatively, a ray cast with forwards momentum could hit the wrong surface and would be more susceptible to poor spacial (floating point on N64?) resolution.