Submission #7805: ktwo, MESHUGGAH & Kirkq's NES Dragon's Lair in 03:50.70

System Nintendo Entertainment System Emulator BizHawk 2.8.0
Game Version JPN Frame Count 13865 (Cycle Count 1238712840)
ROM Filename Dragon's Lair (Japan) Frame Rate 60.09915630721085
Branch Rerecord Count 32217
PowerOn Authors ktwo, MESHUGGAH, Kirkq
Game Dragon's Lair
Submitted by ktwo on 11/12/2022 6:22:16 PM

Submission Comments

Game objectives

  • Emulator used: Bizhawk 2.8
  • Primary objective: speed

Game mechanics

See comments for previous submissions and https://kb.speeddemosarchive.com/Dragon%27s_Lair_(NES) . The latter source currently contains mainly information for the U-version, but many things can be translated to the J-version, despite the significant gameplay differences.

Comments

This is a "side-submission". I originally only intended to work on the U-version of Dragon's Lair, but noticed that some solutions could improve the TAS of the J-version. Since this game is easy and short to TAS, with (mostly) good synching properties, I decided to spend a little time with the J-version as well.
While implementing the improvements that were common to both versions, I also took the opportunity to investigate some of the areas that needed a version-specific approach and came up with a few additional improvements (mainly in the boss fights). The bulk of this TAS is however copied inputs from previous TASes. Comparing Kirkq's (KQ) and MESHUGGAH's (MH) previous efforts, I can see (non-trivial) traces of both in this submission and have therefore included them as co-authors.

Stage by stage comments

Level 0

Based on MH's TAS (which looks like it used some of the inputs from KQ).

Level 1

Based on MH's TAS (which looks like it used some of the inputs from KQ). However, the following improvements were added:
  • Changing the manipulation of the Lizard king (LK) around the first set of gates. This initially lost 18 frames compared to MH's TAS. However, by avoiding bumping into the LK, it could be manipulated again later in the level (if the LK bumps into you, you lose all gold and the LK will not appear again).
  • By avoiding bumping into the LK the first time, two more LK manipulations were possible. The first one saving 39 frames, the second one saving 42 frames.
  • 6 frames saved on the boss fight by starting with a jump (saving wind-up time of the attack animation).
-18 + 39 + 42 + 6 = 69 frames saved in this level.

Level 2

Based on MH's TAS (which looks like it used some of the inputs from KQ), but 21 frames were saved from a faster boss fight (different approach).

Level 3

Based on MH's TAS (which looks like it used some of the inputs from KQ). No improvements in this level.

Level 4

This level is tricky to TAS optimally without external tools. The LK has a 1/816 chance to spawn every game frame (so every 3 actual frames), based on the game's RNG. Other than frame delays, the only way to impact RNG is by pressing R. The "Basic Bot" in Bizhawk was an excellent and invaluable tool for quickly finding suitable solutions. I'm guessing KQ and MH did not use any tools in this level and there were therefore a few delays from unoptimal manipulations in their respective TASes. 21 frames were saved from getting optimal spawns.

Samsara: Claiming for judging. Also replaced the file with a bugfixed version.
Samsara: Solid little set of improvements, here. Accepting as an improvement to the published run.
Quick note about the US version submission: It needs a bit of extra discussion before judgement. Shouldn't take longer than a day or two. This has been taken care of.
Spikestuff: Fun fact MVG showcases footage to the previous TAS (for those that actually watch that channel, since they get brought up time to time in chat). Anyways claiming.

Last Edited by ktwo 15 days ago
Page History Latest diff