Hi!

After having a chat with @pmcTRILOGY I decided to do some tests about version differences between the NTSC-U/C and the NTSC-J version of the game, as I wanted to find out if there is really a need to get the NTSC-J version as well ;)

The write-up can be found on my GitHub: https://github.com/AmenAnthem/speedrun-notes-brahma-force/blob/master/others/version-differences.md

I also linked it as a game resource.

TLDR: Get the NTSC-J version ;) If you are really masochistic you are ofc free to run the NTSC-U/C version as well, be my guest ;)

Hi everyone!

As I want the mod decisions to be transparent to the community I opened this thread to inform you about my decision made in the case of yeroc5311's run and it's verification, despite it containing the usage of an online emulator.

I looked into this case in detail and here are my findings: First of all, the page is extremely questionable, I do not recommend using it. Despite that it uses emulatorjs.com to emulate games, which uses the open source project EmulatorJS. The payloads to load EmulatorJS in this case requested the core "gb", which according to the documentation will use the Gambatte core. As Gambatte is an allowed emulator according to our rules I did look into if EmulatorJS runs it accuratly. It seems to be doing so (concluded by me after some simple verification steps I did with this video).

Conclusion: I will allow this run to be verified, but please abide by the rules and use standalone versions of one of the allowed emulators (when not using original hardware), my personal recommendation is Gambatte Speedrun. EmulatorJS contains the GPL-3.0 license which requires proper attribution, which is not the case here. As I'm a software engineer I do not condone using anything, which doesn't even properly comply with such a simple license. But obviously this is something where the @yeroc5311 would have needed in depth technical knowledge to even find that out, so I will tolerate it this time.

I hope this clears this up, I will leave this thread open for any questions or discussions regarding this decision.

@yeroc5311: I would love to see more PBs, so please keep submitting runs :) Just please as mentioned above use a standalone version of one of the allowed emulators.

Hi!

I completed my first run of MDK on the PS1 (using the PS2 model SCPH-90001), which I will submit as soon as the upload is finished. My run is far from being optimized, currently above one hour.

There are some differences in the PS1 version which make some parts of the game more difficult or outright not possible in the same way as on the PC version.

• Loading times differ vastly.
• Airvents don't seem to carry Kurt as high as in the PC version.
• Field of vision is extremely limited compared to the PC version, because of the pixelated graphics.
• Some bosses take more hits.
• The shuttle in world 5 is practically impossible to hit in sniper mode when it flies over the platform.
• The nuke save glitch is not possible.
• The item throw glitch seems to be impossible as well (could just be me sucking at it though ;))
• Performance of the game is sometimes terrible which makes some stuff extremely difficult to pull of consistently.
• The snowboarding sections in world 4 are extremely cut down compared to the PC version.

There are some other minor things as well which are different, but don't influence the run too much. Also I was not able to pull off any of the OOB tricks (mainly the world 1 ones) because the jump distance seems to be a bit different, but that's a guess on my site after trying them for an extended amount of time.

To have an incentive to run the PS1 version despite all these differences I'd like to propose to separate PC and PS1 into their own categories.

Have a great day, look forward to feedback :)

Hi!

I just did some 1P Battle runs and had another look at the leaderboards here.

Is there a reason for 1P Battle to use RTA instead of IGT? The score screen at the end shows IGT, which IMO should be the preferred timing method for this category so runners using an emulator are not getting this huge of a disadvantage.

Also 1P Battle and Time Attack both should measure milliseconds as well, because they are quite short, seconds are not enough here IMO. I recommend enabling milliseconds for those categories on the leaderboards.

Are there any news regarding the input in our last thread for the Story sub categories? I'm fine with them staying like they currently are, but if there are any updates for any decision please let us know :)

Have a great day!

Hi! Before doing runs I have a question regarding the short cut option on the PS1: Are we allowed to change this option to on, as it would speed up the transitions between fights? The rules don't state anything about that currently, but as noone is using it I wanted to make sure if changing it is actually prohibited.

Hi!

I'm getting into running this game but there seem to be differences in arcade mode, which the current leaderboards don't account for. I'm using version 2.6 of the game.

First of all a difficulty selection was added and the cpu AI was improved, which seems to make even easy more difficult than the standard difficulty previously was.

Edit : After playing arcade mode with other characters it seems that it depends on the character you choose, but I also tried using Goldar as main and didn't get the same fights as the current WR. Is there any documentation to the arcade mode fights each character gets?

Thanks for any answers in advance!

Hi everyone!

The runner @Nozotakun made the discovery, that using cakes instead of ice creams in Death% can be faster, as the eating animation is faster. It requires solid inputs though, as an additional right input is necessary to choose the cake over the ice cream.

I just did some timing on this, as I was curious, as I didn't know the duration of the animations differ. But they really do! Accourding to my tests the cake eating animation is about 0.4 seconds afster, which is huge in this short run.

I hope this discovery brings some runners back to get some competition in this category again. I know it is a meme run, but one that is really great for getting people into speedrunning! Thanks to @Nozotakun for sharing this discovery with the community and congratulations on the new world record!

Hi!

As there is now a verified run I'm thinking about getting into running this game again. @Cha0smagick thanks for giving me the motifation by doing a run and creating this leaderboard!

I already started routing this game a few months ago, as it was one of my favorite games of my childhood. The last quest stopped me though, because I always hated that one. But with this verified run I'm going to give it another shot!

I need some rule clarifications though:

• I do have the 2011 Attic re-release CD-ROM version of the game, but with a full install on my PC, which removes the loading screens. I also have the Steam version, which still contains loading screens. I didn't see any loading screens in the current verified run, so I assume, that the re-release version with full install is valid, correct?
• The rules state "Reload is OK", does that mean that saving and loading is allowed? The current verified run does it, so I assume that this rule is for that?
• What does the rule "Glitches may appear, max 2 game resets if theres any glitch." mean? Only glitchless is allowed currently, but glitches can occure without provoking them. Does that mean I need to reload the game if a glitch occures? I'm quite confused by this one.
• Could you clarify in the rules when the timer starts and when it ends?
• Would it be ok to run the german version of the game?
• Also it seems you didn't import characters from the previous games, could you clarifiy in the rules, that importing characters is not allowed and using newly created characters which where created prior to the run is allowed?
• Why did you end the timer on the queen kill? Wouldn't it be better to end the timer when the screen turns black when using the amulet?

Also do you have a guide or notes (preferably in english)?

How very you able to remove the loading screens from the steam version?

Thanks again for the run, looking forward to your answers and to trying my luck on running this game!

Hi!

I was thinking about running this game, but as I do not have an original console any more I wanted to ask why emulators are banned. There are extremely accurate emulators like Mednafen that even have accurate load times. Would it be possible to reconsider this decision?

Please provide any feedback about the rules and categories here.

Currently the only emulator allowed is Mednafen as it seems to be the only emulator I tested which was accurate and didn't have any looping sound issues. If there are more emulators which are accurate according to https://emulation.gametechwiki.com/index.php/PlayStation_emulators which don't have issues with looping sound please let me know and it will be added as an allowed emulator. I tested Mednafen, RetroArch with the Beetle and Beetle HW cores (which are forks of Mednafen), BizHawk with the Mednafen core and ePSXe. As ePSXe has troubles keeping the load times consistent across versions and has the sound looping issues it was banned. RetroArch and BizHawk have to be configured properly to run accurately, which isn't a problem, but they also have the looping sound issues, at least with the cores i tested them with, so they are banned as long as the sound loop issues aren't fixed.

The recently found OOB glitch to finish Easy in around 1 minute by JkL87 let me to make this forum post to discuss further optimization of it and my opinion on how the leaderboards should incorporate it as a category.

About the glitch: While performing at the Tiny 10 11 Race a runner known as JkL87 (https://www.twitch.tv/jkl87) performed an OOB glitch in the game, which let him to finish Easy way faster than normally.

Here is his twitch VOD containing the run and a tutorial with some additional theory crafting: The tutorial starts at around 7:53:30.

To summarize the glitch: Platforms which you can drop down from by pressing down and jump have a weird mechanic where the game thinks you're still on the plattform while falling off from the side. Therefore it is possible to press down and jump while falling off from the platform and performing the drop while in mid air. This leads to dropping one tile into the ground where you drop onto. While in the ground you can leave the bound of the level by alternately pressing left and right. The game thinks you ended the level when you left the level's bounds.

Further optimizations: The only real optimization I was able to find so far was on Day 1. Dropping into the spikes as fast as possible while facing right as far to the wall to the left as possible and then using the spikes to perform the glitch seems to be the fastest way. If you're lucky you'll be out of bounds while only taking damage from the spikes once. Taking damage from the spikes while in the floor will throw you back in bounds, so getting into the wall while having i-frames is optimal. This is difficult to achieve, but I'm glad there is something about running with this trick that provides some challenge at least, as it is really easy to perform otherwise. There is a rhythm to alternately pressing left and right, for me personally mashing didn't work well. I'm still testing how to optimize the rhythm.

About the leaderboards: Legs already changed the rules for the normal categories. The OOB glitch is therefore prohibited for the normal categories. Legs also added a new category for Easy, which allows the use of the OOB glitch. As the trick is interesting for all categories my suggestion would be to introduce subcategories, one for no major glitches and one for glitches. My reasoning for not calling the normal category glitchless: 3snow_p7im performed a glitch where the character was off screen to reduce lag, but finishing the level normally otherwise. Also I want to bring up the level skip trick which we discussed in another thread and it's influence on the leaderboards. I would suggest banning it completely, as it seems to be a developer testing feature.

Thanks to JkL87 for his contribution, I'm looking forward to doing runs with this trick!

As starting the timing at the name confirmation isn't really accurate as it is difficult to say when it was clicked when having to do a detailed framecount we are going to change the start of timing to the first frame the tamagotchi is fully visible.