Forums  /  Zelda series  /  The Legend of Zelda: Link's Awakening DX  /  Horseheads (continued)
  DiskJunky1DiskJunky1

Seems my edits in the other thread got me auto-banned. While I wait for admin to unblock that account...

I've gotten as far as the D6 room with the horseheads and it's admitidly hard to tell when it has/has not bounced off the wall. I'll continue onto the D7 room where it's easier to test but this has taken several hours to get to this point so I'm calling it a night. Based on tests, I'm currently at "it looks like the bounce count works" but I'd like to show something definitive like a video with at least a 10-retry consistency which is very hard to do in D6 given the narrowness of the room and the rnd movements of the heads when thrown. Will post again tomorrow.

 
  DiskJunky1DiskJunky1

Hmm...this has been preying on my mind for a while now. My hypothesis is based on my younger self testing this out in the D7 room many years ago. However, running the numbers on this;

There are 4 possible states.
There is 1 correct state out of the 4 possible.
If truly RNG, then the probability of failure is 3/4 at the first try is 3/4, or 75%.
This implies that the average retry count for a correct solution this is 3 attempts.

My assumption (if correct), states that it takes at most 2 attempts, assuming throw rng is in your favour (exclude that for the moment to simplify the odds).

It's entirely plausible, if not probable, that my hypothesis is solely based on my misinterpretation of whether it bounced or not, e.g., "I must not have counted the bounces correctly that time" to explain the occasional extra attempt.

I have to admit, actually running the probabilities on this, it's not looking good for my hypothesis. Either way, I'll retest tomorrow with D7 if for no other reason than my own peace of mind.

 
  DiskJunky1DiskJunky1

Random. Ah well, worth a shot though. I think the only way to truly figure out if it's 100% would be to look at the code. Even were there a factor in play, given the maths involved, you'd need a ridiculous sample size (1,000's) to separate out signal from the noise. Back to the drawing board we go. Apologies for wasting anyone's time on this.

 
  SpirasterSpiraster

No worries! 🙂
It's always fun looking into these things even if they don't pan out.

DiskJunky1DiskJunky1 likes this. 
  DiscludeDisclude

As Spiraster said, it's always fun to look into this stuff. This game is so strange, it's easy to overlook things, and having people from the outside giving fresh ideas and thoughts on things is always interesting.

Was fun to mess around with testing out frame rules, bounces and frame counts for these horseheads. Maybe one day something will be found to make them less annoying. 😛

DiskJunky1DiskJunky1 likes this. 
  DiskJunky1DiskJunky1

I'm a little embarrassed that I had that stuck in my head all these years to be honest. Still, it was interesting going through the game to get to D6 quickly. Took about 2.5 hours or so. I always 100% the game (non speed running so don't know the tricks - it's the long way round), so I was surprised how fast I got through it. I'm not a speedrunner but can definitely see the appeal. More fun watching though 🙂

BambooShadowBambooShadow and SpirasterSpiraster like this.