SMALL FIRE SPREADSHEET LINK --> https://docs.google.com/spreadsheets/d/1Br-s42L-yGCJ9_ptST6f2CRNCNlwjyVRTBJ15sRDwkM/edit#gid=931443370
So - Now that Pellsson's newest version of the practice rom (3.1) includes small fire from Title Screen (state "3"), I decided to route some of the rules for the warpless levels with important "rng" (3-4, 5-2, 5-3, 5-4, 6-4) for SMALL FIRE MARIO
Roy has already done this with BIG fire... but interestingly, small fire has totally different "rng" than big fire...
For example, if you enter framerule 1910 (5-4) as small fire mario without having lost powerups, you will encounter "good" bars... if you enter the SAME FRAMERULE as big fire mario (assuming you haven't lost powerups) you'll encounter a more difficult bar pattern.
This happens because there are several variables that determine "rng" (four I think?) and some of them pause during powerup grabs ("framerules" for instance), and some do not, but keep running at normal speed. Because of this, picking up additional powerups or dying effectively scrambles your predictable "randomness". Because small fire picks up THREE powerups and big fire picks up TWO powerups, they have different pseudorandom patterns on the same framerules - so RoyLGamer's Big Fire guide ( https://www.speedrun.com/smb1/guide/fv9u0 ) is incredibly helpful for big fire and worthless for small fire (or small, or if you lose and regain a powerup, or die)
TL/DR small fire framerule rng is different than big fire rng and requires a new guide
Fortunately, that exists now! I have mapped the rules specifically for My (Darpey) and Xeroxfiend's patterns (because those are the two people that I am aware of that are using small fire to go for sub-19:40 runs)... but I'd be happy to map the patterns to any person that's trying for a sub-19:40 or so small fire run.
Assuming you mapped your rules correctly, you could use these in tournaments as well for a slightly faster (and safe) run.
This is a google doc, and it is editable to anyone with link
https://docs.google.com/spreadsheets/d/1Br-s42L-yGCJ9_ptST6f2CRNCNlwjyVRTBJ15sRDwkM/edit?usp=sharing
Shout out to Pellsson for figuring all this madness out... had to do a LOT of testing to see exactly how many frames to "delay" the rng in order to get it to match up to reality.
In case you don't want to open the google doc - here's a "print screen" of the document, to give an example (using Darpey's comparison rules from his 19:17 run - splits link --> https://splits.io/3781
Below is a side-by-side video of the fastest six small fire runs (as of December 14, 2018)
Any% (NTSC) runs below 4:57.000 must now fulfill additional requirements in order to be verified.
- The run's full session must be included in the submission description.
- For emulator runs below 4:57.000, some form of input display must be visible for the duration of the run. A hand-cam or input