TASVideos

Tool-assisted game movies
When human skills are just not enough

Submission #1940: Inzult and Rising Tempest's NES Zelda II - The Adventure of Link in 47:57.05

Console: Nintendo Entertainment System
Game name: Zelda II - The Adventure of Link
Game version: USA
ROM filename: Zelda II - The Adventure of Link (U) [!].nes
Branch:
Emulator: (unknown)
Movie length: 47:57.05
FrameCount: 172623
Re-record count: 38442
Author's real name: Phil C.M. and Jack William Wedge II
Author's nickname: Inzult and Rising Tempest
Submitter: Rising Tempest
Submitted at: 2008-04-17 04:46:05
Text last edited at: 2008-05-17 15:31:20
Text last edited by: Bisqwit
Download: Download (12101 bytes)
Status: published
Click to view the actual publication
Submission instructions
Discuss this submission (also rating / voting)
List all submissions by this submitter
List pages on this site that refer to this submission
View submission text history
Back to the submission list
Author's comments and explanations:
Zelda II, improved by 1077 frames (17.95 seconds).

Uses FCEU 0.98.16

Goal nonsense

This run came about by Inz's incredibly silly idea of levelling magic over attack in the first half of the game. While this seems to be a laughable idea on paper, it's in practice around ten seconds faster overall up to the point where the levels are equal to the attack route, though the actual gameplay sequences are slightly slower.

...that doesn't make sense you say? Then, allow me to elaborate my unenlightened friends. At the end of every dungeon in this game, there is a crystal you place, which refills your life/magic at a rate of 32 frames per container. Having higher magic levels means cheaper spells, hence less magic to refill. This was paid careful attention to throughout the run, a couple magic refills were gotten without need of casting more spells to save time refilling on crystals. Palace three for example: Link's magic meter is full at the crystal allowing gameplay to resume much quicker.

In addition, you also get a free levelup on the crystals. In the first three palaces, this level is used towards the new levelling strategy, but as magic levels are cheaper than attack some more time is saved from waiting less for the experience to tick. Level 8 magic is flat out needed for the fastest route through the final palace, and anything below attack level six later on in the game, bosses simply take too many hits to kill quickly. As opposed to the last run, attack level 7 was declined in this run saving 148 frames from not waiting through the level-up fanfare. As a result, 16 frames were lost on Barba, and 32 frames on Thunderbird, as well as an additional 27 frames from an extra level cancel, the result is a net gain of 73 frames however.

With regards to the actual bosses, time is lost on the first three but not nearly enough to negate the gains from faster magic/experience fills from the Crystals. To those unfamiliar with earlier runs, Link can attack _extremely_ quickly with frame by frame precision in this game, so an extra 10 (horsehead), 12 (helmethead), or even 21 (blue rider) hits doesn't take long to get in. Oddly enough, the fights are more fun to watch as the bosses have time to breathe before they are killed.

As for the overworld, all random encounters gotten into are over a forced battle square. In all instances, the random encounter is quicker than the forced battle.

For time saved outside of the levelling strategy itself;

In the Death Moutain caves, less experience is needed for 2 level ups, so less large enemies were killed, saving time as Link loses momentum to kill these guys without the downstab. In the Hammercave itself, the red Dairas were manipulated into not throwing axes as Link runs away, saving a frame each from not having to slow down mid-air to dodge them. The red Daira at the bottom of the elevator was manipulated into a closer position, and killed quicker than before through the use of a momentum-unfriendly strategy, but as it's required to wait for the experience from him to start ticking before exiting the screen (or he respawns on the way back and throws an undodgeable, unmanipulateable axe) this doesn't end up mattering.

After palace three, we return to the old strategy of restarting (UP and A on controller 2) after the crystal. Upon realising that we can spawn red ironknuckles from the statues at the entrances to palaces three, four, and six, there is no need for the extra enemy in the cave on the way to the dock for the six count (every six kills of a large or small enemy drops either a point bag or a magic refill). We are 1279 frames ahead of the published run at this point, though 745 of those are artificial as 4 level ups are moved to later in the run (see paragraph below).

Palace four uses the same route. Here we spam sword levels, the result is after the restart at the boots we have the same levels as in the published run (6 attack, 7 magic).

The route through palace five was changed drastically, and saved 180 frames to the spider on the overworld. Instead of getting the flute then killing the boss, it is done in reverse with a restart at the flute. The boss here was killed 13 frames quicker than the old run by manipulating him into walking towards Link sooner, and carefully controlling Link's height to be just above the mace, and just inside the detection for downstabbing. The sword level was cancelled, and 1000 points were acquired returning to the flute for a magic refill/level, the last level-up of the run. At the spider, we are 738 frames ahead of the published run.

In addition we take two hits in palace five. Both are as a fairy and do not slow Link down (8 and 9 frames gained respectedly). As more magic is needed to refill at the crystal than life, these hits are taken safely without losing time. The one other place this would save time (6 frames in palace 6) the skull would drain Link's MAGIC, costing more time at the crystal refill.

Palace six is mostly the same as before, though the blue rider was killed 3 frames faster (odd since we have a lower attack level in this run) and Barba 16 frames slower. Declining the attack level on the bridge again saves 148 frames on the published run, minus 29 as this is still an extra level cancelled overall. As said earlier, the extra attack power for the last two bosses that matter doesn't save nearly that much time.

The overworld to the Great Palace went MUCH better in this run. We wound up being 989 frames ahead at the entrance to the Great Palace as a result, a gain of 116 frames over the old run in this area.

Inside the Palace itself, Inz discovered a way to save 6 frames each on the underhang elevators, by bouncing off the opposing wall with the sword. The room where reflect is cast to block fireballs was improved 27 frames by queueing the spell up at the wizard who teaches you Thunder (during the forced pause, rather than a voluntary one in this room). The room where we jump over a huge wall with the dragonhead enemies was improved 72 frames, by stunning one of the heads it was possible to manipulate his position without losing any forward momentum at all.

Thunderbird took an expected hit from the lower attack level, 32 frames lost here. Point of interest, he is killed without using the Jump spell in this run. The keygrab is slower by 27 frames- exactly the same time it would take to pause and cast Jump, so this method was used for more "wow" factor.

And finally, due to some random douchebaggery on Shadow Link's behalf, the input had to end six frames later in the fight than in the published run (though the time between each hit is completely optimized for his 56 frame invulnerability.)

Otherwise, there are some insignificant gains and losses here and there throughout the run thanks to luck manipulation being easier or harder. Also Inzult is a slow worker (so he claims anyway.)

something about audio commentary to come, closing statements blah blah


NesVideoAgent: Hi! I am a robot. I took a few screenshots of this movie and placed them here. Oh! I also corrected the ROM name.
adelikat: Accepting for publication.
Bisqwit: Processing for publication.

Similar submissions (by title and categories where applicable):