This topic is for the purpose of discussing #6821: Xujhan's SNES Breath of Fire II "glitchless, good ending" in 4:44:37.64

om, nom, nom

What was the error? Probably the 34155 frames of blank input at the end. Edit: Yup, it was.

You don't have to resubmit. You can uncancel this and have a Judge replace the submission file with the fixed/correct one.

Okay, thanks very much! Will do that now. Edit: Done. I think everything is as it should be now, but it wouldn't be the first time I missed something that should have been obvious. Thanks again for the help.

By including Bizhawk lag, how does this submission compare?

It's 11345 frames faster than your old 2010 video, but that's not accounting for the lag difference. I don't know off the top of my head how big the emulation difference is, but my video has 57348 lag frames. If you know how many lag frames yours had, we could probably use the difference as a rough estimate.

The other movie has 52,599 lag frames, for a difference of 4,749 frames. So in theory, if you were to play the two movies side by side, Xujhan's movie would finish 6,596 frames before the other one. Maybe someone can double check the math on this. In execution it's a different story, because when I play Xujhan's movie on Bizhawk, at around frame 6279, Ryu enters a forest and the graphics mode switches to something that creates vertical scanlines and lag. This could be a problem that is unique to me. I will try testing on a different machine unless there's a solution I'm not aware of.

No, I get the scanlines as well. Ryu still moves at 16 frames per step though so I think it's purely a visual thing; it doesn't produce any lag frames. It's also only that one screen (which we only enter a few times) so hopefully it's not a serious issue for viewing quality. More weirdly, Bizhawk also doubles the size of window while on that screen. It's only a minor annoyance for fitting to my stream layout, but if someone knows how to get rid of that it'd be nice. Do you have the subtraction backward? My understanding is that the older emulator that janus used in 2010 produced fewer lag frames than a real SNES. So his movie on a modern emulator would be an additional ~5k frames slower. I could well be mistaken though.

480p encode: Link to video

No, I get the scanlines as well. Ryu still moves at 16 frames per step though so I think it's purely a visual thing; it doesn't produce any lag frames. It's also only that one screen (which we only enter a few times) so hopefully it's not a serious issue for viewing quality. More weirdly, Bizhawk also doubles the size of window while on that screen. It's only a minor annoyance for fitting to my stream layout, but if someone knows how to get rid of that it'd be nice. Do you have the subtraction backward? My understanding is that the older emulator that janus used in 2010 produced fewer lag frames than a real SNES. So his movie on a modern emulator would be an additional ~5k frames slower. I could well be mistaken though.

The way I calcultated it is by comparing the exact same moment 9say, when killing Deathevn) and comparing the frame lag at the exact moment

So, once again we have an interesting situation on our hands, here. The currently published run is in Vault, and it uses time-saving glitches (just not weapon glitch as that wasn't discovered at the time, as far as I can tell), which means that glitchless would have to be a separate category, and glitchless isn't a Vaultable category. So what needs to be determined here is whether or not this run is entertaining enough to make it to Moons, and... Given that the published run is sitting comfortably in Vault, and this is a 4 hour and 45 minute TAS (as soon as the file gets updated) to begin with, the deck's a bit stacked against it... But this is a weird case, because at the current moment this is the fastest known TAS of this game, at least that I can find. Nicovideo doesn't return a lot of results other than janus' previous runs, but I could have missed something (searching on Nico directly returns nothing because apparently I'm awful at searches, everything I found was searching through Google), and even if there was something on Nico, I'd imagine a 5ish hour RPG TAS would be significantly faster through the text differences alone, meaning it would be harder to compare actual improvements without watching the runs in full. On the same note of comparisons, it's hard to compare this run and the published run given that this one's purely glitchless, and it's hard to compare this run and the other submission because the other submission is missing techniques used in this run, and comparing this run to the improvement run... Kiiiiiinda requires the improvement run in the first place. So how do we judge the entertainment value of this run in the meantime? As a non-fastest completion category, we can't let it stand on its own. As a 4 hour and 45 minute RPG TAS, I imagine it'd be kind of a difficult watch for a lot of people... And there's also the precedent that foregoing time saves in a Breath of Fire TAS isn't publishable. Yes, I'm aware that's BoF3, which is 3 hours longer than this, and that it's strictly slower than the published BoF3 TAS overall, but in terms of pure comparison I think it's still worth bringing up. Does entertainment even matter right now, though, given that this is currently the fastest completion of the game? Even if an improvement that will be faster is in the works, assuming this run gets judged before the new submission comes in, does the fact that it's the current TAS record make it Vault eligible, or do the known improvements nullify that? I can't immediately think of a precedent for this, a glitchless run obsoleting a glitched Vault run as fastest completion despite the fact that it's known glitches are faster, and honestly I don't even know where to begin to look with that. Obviously, this is a tough ask, but I think this run needs a good amount of audience feedback given the weird situation. I'm planning on watching this run and janus' (EDIT: RECENTLY CANCELLED) submission side by side within the next couple of days to compare them on an entertainment level. I'm sure the categories warrant separate publications, but that would require this run meeting the entertainment level it needs to be published in the first place. ...TASvideos, as always, remains one of the most complicated places on the internet. I love it. Genuinely. No sarcasm. Never change. <3

To clarify, since I assume that most people here aren't super familiar with the game: Janus' 2010 video uses exactly one instance of a minor movement glitch that lets him walk through an object and saves only a few seconds. As far as entertainment considerations go, it might as well be considered a glitchless run. It's also safe to assume that this run will be outdated soon. I'm already partway through a second pass of the glitchless TAS that's going to save a couple minutes, and both janus and I are actively working on any% videos. I've no idea if either of those things have any bearing on the decision - I know next to nothing about the TAS community - but I figured I'd provide as much information as I can.

Believe me, I'm no closer than you to understanding how TASvideos operates, and I helped write some of these rules. At this point, if improvements are in the works, it might be easier to take the same route as janus and cancel this submission in favor of the new one. Having two optimized runs would be a lot easier of a comparison to make regarding whether or not both runs are publishable alongside each other. Are you two working together? Your post makes it sound like you're both separately working on any% runs, which seems a bit weird to me.

Avoiding known significant time-saving techniques for the sake of entertainment isn't eligible for Vault. Speed/entertainment tradeoffs are somewhat ok, if they don't lose significant time. When deciding whether it's a primary goal or just a small trait, we check if that would look different enough to be published as a different branch. Speed/entertainment tradeoffs alone don't result in separate branches. Avoiding time-saving glitches often does.

We're not really working together, no. I started working on this as something fun to do with the BoF2 speedrunning community, which janus isn't part of. I've offered him a few tips for improving his video, but I don't expect we'll collaborate much beyond that. I'm estimating that it'll take me a month or so to finish the improved glitchless TAS. I can take this one down if that's the consensus, but the improvements are all just minor optimizations so they shouldn't affect the entertainment quality much one way or the other.

Ryu enters a forest and the graphics mode switches to something that creates vertical scanlines and lag. This could be a problem that is unique to me. I will try testing on a different machine unless there's a solution I'm not aware of.

This happens due to the snes switching to Hi Rez mode for that area (512x224 instead of 256 x 224)

Hello, I'm a speedrunner of Breath of Fire II Any% and Glitchless. I was also the one that discovered the weapon glitch (both the original herb glitch and then the improved version which janus used in the newer TAS) I have done a lot of work with re-routing the Any% speedrun with the glitched weapons and have speedran I have not watched much of either TAS to be honest. There were a few points of interest that I was curious to see how janus approached them. My guess is that xujhan mentioned these to janus. #1 - In the any% speedrun we glitch about 1.7 mil coins (zenny) into our inventory at the start of the game (in hometown when the first weapon glitch is done) This allows us to avoid the trip to get an uparupa (2:04:30 in janus video) Instead you can pay MC Tusk 900k coins and avoid the trip altogether. There is also a text glitch that we do to leave mc tusk's room and re-enter it, which saves us a trip of going through the circus again. Based on the walking time for the cave, this can be about a 3 minute time save and is something glitchless can't do. -------------------------------------------------- #2 - No need to chase after Patty. This is done in both speedrun categories. At (4:02:30 for janus video and at 4:01:25 in xujhan's video) They are leaving Cotland to go find Patty. The game sends you on a wild goose chase, but in reality the flag is set at this point to talk to Patty in township. Avoiding the extra detours will save janus about 90 seconds. At 3:57:00 in janus also moved the township to help with "finding Patty" but since you can cut that part out, there is no need to move the Township which saves another 60 seconds. In total this will save janus a little over 2 minutes and 30 seconds. -------------------------------------------------- As for the weapon glitches. There are 9 glitched weapons in total. The moondrop is what janus used. all 9 weapons have 64 attack power. The moondrop (item 06) and van ext (item 04) have the unique property to attack twice and have the holy element. Nothing resist holy and spirit enemies take double damage from holy,so it works out wonderfully. In the vanilla game only 1 weapon has twin attack and that is Bow's twinbow. As for the 7 other weapons, again they also have 64 attack power but only attack once and have the fire property. The herb glitch (item 01) was the original discovery. This was a huge discovery because 64 attack power is a LOT for the early game. With the discovery of the other weapons and gaining a weapon that attacked twice made it that much better. For the speedrun, the glitched weapon starts to have diminishing returns because characters start to lack a high enough strength to keep up with the enemies defense. This is even more true for the TAS because your level is much lower thanks to maniping no enemy encounters. Sadly when you ask for what is the entertainment value for glitched weapons vs not having them it is harder to see because random encounters are skipped in the TAS. In the speedrun, Rand's strength gains allow him to become a dps machine in the late game with the glitched weapon, since the developers gave him weak attack power weapons. The glitched weapon has more than all of his weapons accept for one irc. Again, since battles are skipped it's hard to see the value (in great significant's) because the boss battles are handle over so fast in both TAS versions. ----------------------------------- In conclusion. I feel you do have a hard choice ahead of you, but I feel that the entertainment value is the right question to ask. I wanted to lay out some of the time saves that would set janus apart from xujhan, and I also wanted to try to help and educate where those differences are in the two runs. IMO the runs will still look very similar. For where to go from there, I would leave it to the community and what values you guys place first.

Oh. As far as JP goes there has not been too much of a community for the speedrun as far as I know. The current ENG WR holder also has the JP WR. The run is the same, but the text does add up as you guessed. (as of now there is about a 25-28 minute difference between the versions due to text) Also there is a walk anywhere (OOB glitch) that is not real time viable. Even if it was, I found no sequence breaks. It is really cool thou! and you can do a few neat things with it and could result in a very slight time save. This would make the Any% really stick out -----sorry for hijacking glitchless thread but the other forum seems to be directly connected to this at this time.

I'm estimating that it'll take me a month or so to finish the improved glitchless TAS. I can take this one down if that's the consensus, but the improvements are all just minor optimizations so they shouldn't affect the entertainment quality much one way or the other.

If the movie's not going to change much overall, I can set it to delayed until you post the improved version. If it's a major change, it may be better to cancel and resubmit.

From what I saw of the glitched run, it doesn't seem to be a major skip glitch and is more getting a better weapon. So the difference is basically you get to see higher numbers in the glitched run. I don't think that's different enough.

Hey everyone, just swinging by with an update. At the rate it's been going I should have an any% video to submit in about two weeks. Of course I'd be delighted to see my glitchless video get its own category, but I'll fully admit that the differences are mostly only of interest to people who actually run the game (different boss strats, skipping one dungeon, etc). For the casual viewer, the two are mostly interchangeable.

Thanks for the update! Glad to see you're working on the any%, having an optimized run will make it much easier to compare the two categories. Admittedly (and unfortunately), given the lack of interest on this run, the fate of glitchless seems to be rejection at the moment, but I'll wait for the new any% and compare the differences myself before making a decision on what to do with it.

While I'm thinking of it: once the any% video is done, should I update this submission or make a new one? I'd guess the latter, but I figured I should double check.