Quoting the published pacifist run description:
Simon is taking another shot at this castle, but feeling horrible for the destruction he caused last time, he is now trying to respect Dracula's property and leave the place untouched (even if his intention is to kill the count). As a minimalist variation of a pacifist goal, this movie does not contain any destruction in the levels: candles and walls are left intact, items stay untouched, and no enemy is harmed — with the exception of the bosses, who are killed mercilessly with the help of the "critical hit" glitch.
With Simon sneaking by most enemies with no subweapons and only his leather whip, this ends up being quite a stealthy run.
Game objectives
- Emulator used: BizHawk 2.2 (but could works on the lastest versions, like the any% run)
- Heavy luck manipulation
- Takes damage to save time
- Abuses programming errors
- Uses death to save time
Same rules as the previous run:
- Like every pacifist run, you must not kill any enemy, except the bosses (in order to advance).
- Don't break the furniture. Candles and walls are left intact.
- Don't pick up items. They can indeed still drop from projectiles for example.
1689 frames ahead of the published run (the extra frame is from NesHawk power-on), thanks to better optimization and manipulation. Also this run fixes 2 rule flaws.
Development of the run
During the work on any% run (w/Morrison), I acquired more knownledge and experience than before, and motivated me to also improve the pacifist run, implementing some new stuff discovered on that work.
For this project, I exported the published run from FCEUX to BizHawk. Initially I tried to resync entirely, but the first room of Stage 13 ruined it, abandoning the resync and starting the work - TAStudio was used.
This run shows the improved turaround (used on any% run), more manipulation, and learns 3 new things: pause manipulation abuse, cancelling damage knockback (used once), and "extra" critical hit!
Stage-by-stage commentary
General notes
- Like the any% run, it's better climbing stairs by pressing up/down + left/right depending of direction. This saves a frame for most of these stairs.
- There are moments where I "unnecessarily" jump, for avoid certain lags.
- When resynching this game, the skeletons would recieve a different pattern, and can be manipulated/resynched with input (except select button), but certain things only can be done with pausing/unpausing game or stage timing.
- Normally it's better finishing a boss with a critical hit, but not did for the last 2 bosses due to obvious reasons.
- It's faster reaching the ground as soon as possible when grabbing an orb, because there's a certain timer before "time left" scoring.
Level 1 | Frames saved | Total |
---|---|---|
Stage 01 | 16 | 16 |
Stage 02 | 6 | 22 |
Stage 03 | 51 | 73 |
- Stage 01: Jumping backwards for ghost manipulation (before the last stairs), allowing Simon to beat a framerule (16 frames)
- Besides better turning around and stair optimization, 1 less lag during the second stair transition for some reason.
- Stage 03: For the Vampire Bat fight, the same any% strategy was used, but due to small hitbox of the leather whip, 4 frames were lost.
Time left: 233 to 234.
Level 2 | Frames saved | Total |
---|---|---|
Stage 04 | 2 | 2 |
Stage 05 | 2 | 4 |
Stage 06 | 162 | 166 |
- No extra lag transition (stage 05 and 06) this time.
- Stage 06: During the work on any% run, I notiched that you can delay the function of these crushers for 16 frames by pausing on "exactly" frames, so he can across without stopping at all!
Note: The pause abuse isn't technically new, because is also present on Metal Gear (NES) since the first published movie, and on the current pacifist run. Both are also Konami games.
- The second bonepillar dodge was improved a bit.
- Starting with Medusa fight, the boss strategies are very new, and being more agressive than before.
- Attacking on the opposite side to also use the ceiling for cancel the knockback when doing a critical attack. A snake was accidentally killed on the previous run - fixed it.
Time left: 328 to 330.
Level 3 | Frames saved | Total |
---|---|---|
Stage 07 | 252 | 252 |
Stage 08 | 1 | 253 |
Stage 09 | 185 | 438 |
- The first room is same as any% run - and 13 frames saved.
- Big change on the next room: Before reaching the second raven, wait a few frames and the raven will descend enough to bypass and skip that difficult effort with three ravens "trying to self-destruct" on Simon. These skeletons are so annoying that I gained a lag frame while crossing for not lose a framerule.
- Stage 08: Extra lag transition added, for some reason.
- On Stage 09, the only optimization for this category is reaching the second skeleton as soon as possible, since he'll toss bones no matter how closer you can try to bypass him. Initially I reached him 2 frames earlier, and the remaining was almost identical as the previous run. After optimizing the mummies, I noticed that would be still possible to reach that skeleton a few frames earlier, at the cost of 1 frame (for raven manipulation) and a lag frame (almost immediately after spawning that skeleton).
- By reaching the skeleton more earlier, this also affects the pattern: He'll move one more time before reaching the bonepillar, so I didn't stop so much. Pause used for high tossing manipulation instead of lower tossing.
- The mummies are difficult to optimize, since you can get a different pattern by entering on different frames. Initially I used only one pause, but when I implemented the new skeleton pattern, the fight was affected so much that I only implemented after finished Dracula optimization. Because of desync pattern, more pauses were used to fix (not exactly) the fight. This resulted on some frames lost and also added more time left for the scoring, losing 2 frames than before.
Time left: 413 to 421 (pausing affects game time too).
Level 4 | Frames saved | Total |
---|---|---|
Stage 10 | 6 | 6 |
Stage 11 | 0 | 0 |
Stage 12 | 79 | 85 |
- Stage 10: Removed lag frames.
- A framerule of stage 11 could be improved by 1 less turnaround, but these hunchbacks and birds ruined the possibility - so, no way (tested several positions).
- Pausing game on skeledragon (used on any% run) is done a bit later, saving some frames - but adding a lag frame (It saves 4 frames to the any% run).
- The Frankestein and Igor battle : Simon decides to only attack Frankestein. Igor was helpful this time, because I used the fireballs for critical hits. And is another critical hit type: If you get hit and not turnaround, the enemy will recieve more damage. Critical hits with leather whip normally does 15 damage, but with extra critical hit, it reaches 18 HP!
- Igor can be manipulated with some input, but not with Frankestein (pausing still works).
Time left: 322 to 323.
Level 5 | Frames saved | Total |
---|---|---|
Stage 13 | 49 | 49 |
Stage 14 | 56 | 105 |
Stage 15 | 189 | 294 |
- Not killing enemies means a laggy room, but thanks for manipulation, and forcing one of these skeletons to leave the screen, most of the lag were removed. One more lag (before damage boost) could be removed, but got some bad luck on the way. What a tough room.
- Frame perfect to beat a framerule on the next room.
- Still on the same room, the new damage boost can't be used since I need HP for the damage boost of the next room (stage 14).
- Stage 14: The first room is same as any% run - improvement.
- No extra lag transition this time.
- Axe knight manipulated to stop earlier, improving a framerule.
- Stage 15: Initially the Death boss was improved, but when I improved the second axe knight, the Death fight wasn't affected (he moved on the same frame before). I used the "timing" opportunity (before entering the fight) for manipulation.
- The second pausing was necessary, because otherwise, he'll not move when doing an "extra" critical attack.
- The fight is bit different than the WIP I've shown before. But both strategies also helped fixing a noticeable flaw: One scythe was destroyed on the previous run.
Time left: 256 to 260.
Level 6 | Frames saved | Total |
---|---|---|
Stage 16 | 1 | 1 |
Stage 17 | 61 | 62 |
Stage 18 | 571 | 633 |
- Improved stairs and entering stage 17 = 1 frame saved, respectively. But a frame is lost during transition (like the any% run).
- Second damage boost improved (more pixels left), different enemy manipulation, "third room" with pause manipulation (waiting 6-7 frames) for different third bird direction, and also jumping for another bird manipulation (otherwise, he'll reach Simon).
Dracula (Stage 18)
Dying is really faster, since restores all HP and the critical hits are definitely useful for phase 1.
Phase 1:
- By entering dracula battle 1 frame later (like the any% run), he can release fireballs earlier. It's possible for get a more earlier (12 frames) reaching a bit closer, but isn't enough for a third whipping attack (critical hit). I also tried jumping backwards and forward (respectively), but when the whip worked, I almost reached the head - so, no way.
- With extra critical hit, dracula is defeated with 3 cycles instead of 4 cycles.
Phase 2
- Hitting ghost dracula without breaking candles is frame perfect, and positioning is also important (so I adjusted him before defeating on the previous cycle).
- When starting cycle 1, he can jump lower or high depending of when you finish phase 1 (single frame difference).
- Whipping on certain frames will make ghost dracula to jump for some reason. This is helpful, since you must wait if you want for attack twice with a single jump - also frame perfect.
- With some farther distance, you can get the fireball pattern, which allows Simon to attack more times in a cycle before jumps.
- For the lastest 2 cycles, making a high jump (pausing is necessary) is better, since I need farther distance for ending input jumping forward while whipping, and not hitting a candle.
Time left: 434 to 443!
Table of improvements
Stage name | Frames saved | Total |
---|---|---|
Main Hall | 73 | 73 |
Torture Lab | 166 | 239 |
Castle Ruins | 438 | 677 |
The Catacombs | 85 | 762 |
The Dungeon | 294 | 1056 |
Castle Keep | 633 | 1689 |
Note: The stage names were retired from vgmuseum's site.
Other comments
Thanks to Grincevent for the published pacifist run and for the RAM addresses (again); And Morrison for the nice experience and knownledge with this game previously.
Well, this is the end of year. Initially I wanted to work another game before finishing 2018, but certain problems appeared, and decided to rest during the remaining days.
That's all folks! Probably the next run will be another NES game.
ThunderAxe31: Judging.
ThunderAxe31: Accepting as improvement over the published movie.
fsvgm777: Processing.