Go for it (just please disclose this and the version in the submissions). Nintendont is a bit of an unknown but this question has come up, it's assumed to be fine, there is some analysis being done but in the worst case I think we'd allow it under a split leaderboard (this may still happen even for Dolphin in the future, it isn't really a big concern right now).

The main issue is the load times varying by significant amounts making it difficult/impossible to mix times together in an ordered ranking. As far as I understand, Nintendont is just a software loader that emulates the disc drive and maybe memory cards? If there's an option to use the regular memory cards that would be preferred, but otherwise I assume it's running the game exactly like the Wii would.

In the long run solutions like this should be embraced because working discs and disc drives will only get rarer over time.

Hello,

Emulator guidelines have been added to the game to clarify the settings and versions for Dolphin. This was in response to some confusion (especially since Dolphin had/has confusing versioning).

The main takeaway is: please update to Dolphin 2412 or later, verify you are using "Emulate disc speed", and when you submit please include the emulator version in your submission. Other settings required are disabling cheats and custom textures.

Please let me know of any questions or concerns.

Thanks,

AKD

PSO Speedrunning has mostly had Discord presence on individual streamer Discord servers and some limited presence on other servers for the last few years, but a new independent server dedicated to PSO speedrunning is now available: https://discord.gg/rhy6AUgU6M

Proposal

Phantasy Star Online is usually timed as Real-Time-Attack (non-stop timer until the game is completed). With the understanding that this category is unreasonable and unhealthy to run without taking breaks, this category uses alternate timing rules: the In-Game Timer displayed on the character-select screen after completing the category and watching the final credits shall be used instead. The game does not display seconds, so times shall be listed with 0 seconds. RTA time should also be tracked and displayed or added to the run comments.

Regular short breaks are permitted at runner's discretion, subject to some restrictions. To take a break, the runner may save and quit the game, or watch the credits.

With the understanding that IGT is the exception for traditional PSO speedrunning, and that this category is specifically distinct from segmented runs (which allow careful planning and repetition until an ideal outcome is achieved), it is expected that the game be completed in a similar amount of time RTA compared to the IGT.

Regular short breaks may not exceed 10 minutes each. After 14 hours of RTA time, one sleep break is allowed up to 8.5 hours long. Another sleep break is allowed at 38 hours if necessary. Regular short break time should not exceed 6.25% of the RTA (except for sleep breaks). For example, a 24 hour RTA run (after subtracting any sleep break time) should have at least 22:30 of In-Game-Time.

Option 1

Use of resets that do not save the game is restricted. Any in-game-time that is discarded this way must be added to the final displayed time.

Option 2

Use of resets that do not save the game is not directly restricted, but any unsaved time counts as an extension of a regular short break.

Option 3

Resetting the game (without a save) is only allowed if no EXP was gained since last save.

Option 4

Instead of using the imperfect In-Game-Time, we will use RTA minus any break time, which does not include gameplay with resets. This is obviously the most tedious to calculate and verify...

Option 5

Despair, just use RTA, and get good at optimizing breaks and beat the game before bed-time.

So I ran FOnewm through Ultimate (theoretically an extreme case because of the longer Hard mode) in part to get a feel of how this works in practice.

My calculation of break time (any time not included in IGT) followed my best understanding of the game:

• Time starts when starting character creation ("Yes" turns green on recreate)
• Time is paused on quit ("Yes" turns green on confirming quit)
• Time is paused on entering Epilogue teleporter (palette highlights corresponding to button press in the teleporter while Epilogue label is displayed)
• Time restarts when selecting "Single Mode"

My IGT was listed as 22:28', which corresponds to my my manually calculated IGT of 22:28:07. RTA was 23:14:38.

| Event     | Time     | Duration | Time Since Last |
|-----------|----------|----------|-----------------|
| Start     | 00:00:00 | 00:00:00 |                 |
| Credits   | 02:10:37 | 00:05:15 | 02:10:37        |
| Save&Quit | 03:51:49 | 00:03:12 | 01:35:56        |
| Save&Quit | 05:29:31 | 00:03:31 | 01:34:29        |
| Save&Quit | 06:23:27 | 00:00:17 | 00:50:24        |
| Save&Quit | 07:17:30 | 00:02:34 | 00:53:46        |
| Save&Quit | 08:44:22 | 00:00:17 | 01:24:17        |
| Save&Quit | 09:07:17 | 00:03:32 | 00:22:37        |
| Credits   | 10:05:20 | 00:05:15 | 00:54:30        |
| Save&Quit | 12:03:29 | 00:04:36 | 01:52:53        |
| Save&Quit | 13:53:04 | 00:02:00 | 01:44:59        |
| Save&Quit | 15:02:31 | 00:03:11 | 01:07:27        |
| Credits   | 17:16:57 | 00:05:15 | 02:11:14        |
| Save&Quit | 17:40:19 | 00:00:16 | 00:18:05        |
| Save&Quit | 19:06:21 | 00:02:52 | 01:25:45        |
| Save&Quit | 20:28:03 | 00:03:10 | 01:18:49        |
| Save&Quit | 20:40:54 | 00:01:11 | 00:09:40        |
| Finish    | 23:14:36 | 00:00:00 | 02:32:30        |

Total Break Time: 46:30

I prepared some food and plenty of water ahead of time, and used breaks for refilling water, heating food, bathroom, etc. This shows that a category like this can be done only taking short breaks, if prepared for ahead of time.

Regrettably, I did not feel like I should take a sleep break. It was sensible to do this from the perspective of focus and the fact that I would have a lot of trouble just falling asleep and having good sleep anyway. Several days later I'm still feeling the effects of that decision. I stand by what I originally said--a sleep break should be permitted.

There is a wrinkle in using displayed IGT, which is how resets can be abused. Though innocently, I carelessly used a reset in my run instead of a Save&Quit, which means I gained some game knowledge but discarded the IGT along with that map in order to play a different one. This technique is allowed in and used occasionally in glitchless RTA runs (it is a calculated risk to try it there), so I didn't think through the consequences of doing so in the moment of the run, 20 hours in.

Perhaps this can be embraced, though reasonable limits should be imposed. I think it would be very bad to allow discarding time excessively. For example, there is a theoretical statistical average point where you should reset the game rather than accrue IGT while farming for God/HP. Allowing that unrestricted would align the category much closer to what a segmented run is. Perhaps such abuse should just be directly penalized.

Next post is my best stab at a recommendation for timing rules in IGT-based categories with a few ideas on how to keep abuse to a minimum. I'm obviously not unbiased here, especially since I accidentally abused IGT (at least what my intention was), so I welcome all feedback.

To make my own preferences clear: IGT is nice, but it has unfortunate side effects. RTA is not great for out-of-game reasons. I don't think this should be a super serious category so I'd lean towards IGT with restrictions/penalties on resetting.

The run is likely in the ballpark of a couple or few dozens of hours of real gameplay. It will vary a lot where the grinding happens for each class and what the grind is specifically, but there are strict HP thresholds to meet to survive Falz and/or not have a bad time in Ultimate mode in general.

The challenge with All Difficulties is that the gameplay is mostly backloaded to the final difficulty. Defeating Hard Falz can be done in under 6 hours as a HUcast. Getting to Ultimate probably can be done in under 6 (not sure how much grind is required specifically), but even if it was longer, it's probably doable in a day even with a break. Since each class has different problems and different solutions to those problems (e.g. FOnewm has a long grind just to get out of either Hard mode or Very Hard mode), so I hesitate about just granting breaks after Falz because it's probably not going to be evenly distributed.

Something very important that I just realized: IGT is NOT paused when sitting on character select. The timer actually starts counting when you pick "Single Mode", and time will be added on to the character you load and next save. Weirdly though, it seems like this "out of game time" can be simply reset by returning to the Single Mode/Multi Mode select. So to minimize IGT, it is advisable to do this "IGT clear". If this is at all problematic, "IGT" could be manually timed instead to some different standard, but that sounds like annoying work especially if the VOD is in multiple chunks. I am not worried much about the fact that IGT doesn't display seconds for the longest category, but if breaks are introduced on things that can be done in a day, manual timing might be better. More reason to try to get one solid recording file and put it on YouTube in pieces afterwards.

If the idea for staging/breaks is to encourage healthy running, then the standard could be set around what we expect people to do on a "daily" basis. One idea: for every 24 hours of RTA, 10 of them could be permitted to be breaks (8 hours sleep, 2 hours for meals / hygiene.) If someone can actually finish a run in 14 hours then the sleep break doesn't make as much sense. If the run takes two days with breaks, only one sleep break makes sense.

Trying some numbers like ".5-1 hours of break every 6-8 hours RTA" and "7-9 hours of break every 24 hours RTA" seem to be awkward. Maybe that's unavoidable and that's the point, to encourage people to pace themselves? 24/14 as a simple ratio also might be kind of high. Maybe this can be written more elegantly around a longer sleep break.

One proposal I'm fairly happy with: grant a free pause while watching credits (teleporter activation to regaining control on the title screen). Any longer category could benefit from this.

I'm reposting this comment because the original thread has disappeared. Since this is relevant to runs I have done and things I want to try eventually I have given some thought to whether longer categories should allow for timer-pauses for breaks. The tl;dr: is probably, because we shouldn't encourage doing unhealthy things.

My current stance is to at the minimum record in-game-time (either the literal time the game reads or just subtract time when the in-game-timer is not running) with the idea that people may save&quit in longer categories to take a break. The limits on breaks I'm not sure of yet--though my gut feel is that an RTA run generally should permit the minimal amount reasonable. I welcome other perspectives.

This was originally responding to a question about the logistics of the category labeled as "All Difficulties Glitchless".

To my knowledge, all leaderboards are real-time-attack by default or at least as they stand.

This category has been sitting in the back of my mind, but even before that, the topic of pausing has been raised for Three Class Relay (6-8 hour gauntlet). That run can be done in one sitting, but with the game offering no downtime longer than 30 seconds (even the intermission credits can be skipped), it's definitely on the borderline of what is a good idea health wise. As far as I know, no one wants to to encourage people to be self-destructive by writing leaderboard rules that reward running an extreme/unhealthy amount continuously.

My approach to that category was to just go as fast as possible and get it done because I'm used to doing one-sitting runs of about that long if I prepare for it. One alternative idea was to prohibit skipping the credits to force a ~5 minute break twice during the run. For what it takes to get through Ultimate Falz, that idea wouldn't be enough.

My personal preference is the least amount of arbitrary rules is the best amount, or "less is more", because there exists some limit on how long of a break is reasonable/necessary, not that there is anything wrong with doing challenges over a longer period. Baten Kaitos 100% just goes with RTA even though it's 14 days minimum, although that idea makes sense for that game with its unique requirements and not necessarily this one.

There are real world analogues to handle this kind of challenge, such as road bicycle races which sometimes take place over the course of days/stages. Perhaps that kind of format makes sense, combined with the force break idea. For example, the run could be RTA timed, but expect people to save&quit at certain points to take a break of at least X minutes/hours, such as after each difficulty and maybe each area in Ultimate.

One approach would be just to also record in-game-time in addition to RTA, which would remove time sitting on character-select. The in-game-time could be used as a primary time with the expectation that RTA doesn't exceed it by some amount, for example if it takes 36 hours IGT (12 hours * 3), perhaps the expectation is that RTA does not exceed 72 hours (3 full days). Or maybe a 3x multiplier is reasonable. This type of system gives plenty of opportunities for healthy breaks while keeping a simple rule to keep it closer to RTA speedrunning. It also scales to shorter categories so it's interesting to me.

Logistically, the simplest solution is to approach the category as segmented instead of RTA, which avoids having to implement arbitrary timing rules, but that may not appeal.

Since there is no competition in the category yet, it's sort of a moot issue that I don't think needs to be resolved to just try the run. I encourage just running what you want to run and how you want to run it, especially if you are pioneering a new category--what makes sense will be more apparent once it is known how long things really take and whether certain ideas worked out or not, or whether other ideas seem really good in hindsight. Nothing has to be set in stone.

Regarding VODs, like any other run, I would strongly recommend making a local recording, to avoid any problems with dropped frames or disconnects from Twitch causing pieces of the run to be missed. With this, there are tools to cut up local recordings even you make one continuous recording.

Besides YT limits, Twitch I believe enforces a 48 hour maximum on one broadcast, something to keep in mind.

One last idea for RTA: without a continuous VOD available and breaks involved, it may be a good idea to additionally capture the current date/time of day in addition to the RTA timer.

(apologies, it looks like the quote function doesn't format well on this forum)

To my knowledge, all leaderboards are real-time-attack by default or at least as they stand.

This category has been sitting in the back of my mind, but even before that, the topic of pausing has been raised for Three Class Relay (6-8 hour gauntlet). That run can be done in one sitting, but with the game offering no downtime longer than 30 seconds (even the intermission credits can be skipped), it's definitely on the borderline of what is a good idea health wise. As far as I know, no one wants to to encourage people to be self-destructive by writing leaderboard rules that reward running an extreme/unhealthy amount continuously.

My approach to that category was to just go as fast as possible and get it done because I'm used to doing one-sitting runs of about that long if I prepare for it. One alternative idea was to prohibit skipping the credits to force a ~5 minute break twice during the run. For what it takes to get through Ultimate Falz, that idea wouldn't be enough.

My personal preference is the least amount of arbitrary rules is the best amount, or "less is more", because there exists some limit on how long of a break is reasonable/necessary, not that there is anything wrong with doing challenges over a longer period. Baten Kaitos 100% just goes with RTA even though it's 14 days minimum, although that idea makes sense for that game with its unique requirements and not necessarily this one.

There are real world analogues to handle this kind of challenge, such as road bicycle races which sometimes take place over the course of days/stages. Perhaps that kind of format makes sense, combined with the force break idea. For example, the run could be RTA timed, but expect people to save&quit at certain points to take a break of at least X minutes/hours, such as after each difficulty and maybe each area in Ultimate.

One approach would be just to also record in-game-time in addition to RTA, which would remove time sitting on character-select. The in-game-time could be used as a primary time with the expectation that RTA doesn't exceed it by some amount, for example if it takes 36 hours IGT (12 hours * 3), perhaps the expectation is that RTA does not exceed 72 hours (3 full days). Or maybe a 3x multiplier is reasonable. This type of system gives plenty of opportunities for healthy breaks while keeping a simple rule to keep it closer to RTA speedrunning. It also scales to shorter categories so it's interesting to me.

Logistically, the simplest solution is to approach the category as segmented instead of RTA, which avoids having to implement arbitrary timing rules, but that may not appeal.

Since there is no competition in the category yet, it's sort of a moot issue that I don't think needs to be resolved to just try the run. I encourage just running what you want to run and how you want to run it, especially if you are pioneering a new category--what makes sense will be more apparent once it is known how long things really take and whether certain ideas worked out or not, or whether other ideas seem really good in hindsight. Nothing has to be set in stone.

Regarding VODs, like any other run, I would strongly recommend making a local recording, to avoid any problems with dropped frames or disconnects from Twitch causing pieces of the run to be missed. With this, there are tools to cut up local recordings even you make one continuous recording.

Besides YT limits, Twitch I believe enforces a 48 hour maximum on one broadcast, something to keep in mind.

One last idea for RTA: without a continuous VOD available and breaks involved, it may be a good idea to additionally capture the current date/time of day in addition to the RTA timer.

Any sort of way of recording the run is accepted as far as I know. Many runners stream to Twitch and many also record locally while doing so (but you can do one or the other). I won't speak for the moderators, but I doubt it'd come up unless there was something fishy about the run (it'd be incredible to post a top time without any apparent prior experience).

Simple answer: No - all categories assume you are starting from a fresh file

"All Difficulties Glitchless" is on the leaderboard for some reason but I wouldn't say there is a great one.

You have found the problem with doing an Ultimate run: it requires an absurd amount of grinding (I have no idea how viable it is in v2, but v3 it's theoretically known that you spend a dozen hours or more just leveling/grinding for HP units). There are also issues with your proposal to use existing files. In my experience in speedruns, when there is the option to start from something other than a new file, there are two options: have a standard file, or anything goes. For PSO, both options seem burdensome or fairly inaccessible.

In the case of the Standard File, what makes sense for what's on the character, and who decides? When it comes to something like this, less is more, but that doesn't exactly work for PSO because you're expected to have a lot of things to run in Ultimate. Also, how difficult is it to make a file up to standards? I suppose you could get to level 200 and ditch all your gear except the starting gear, but you'd have to also be material-less. Seems obnoxious or require extraordinary setup.

In the case of anything goes (e.g. true NG+), then the run becomes "I grinded ___" or "I constructed this file with perfect loot.." which may be interesting, but fairly inaccessible.

Categories with low accessibility/very burdensome requirements don't have much meaning and therefore probably shouldn't be on the leaderboard as a serious category. Contrast this against most speedruns where you just need the game and you're set.

Don't let me tell you what to do, but here are my recommendations:

1. You can play whatever you want to play.
2. Personally, I would not recommend pursuing the category "All Difficulties Glitchless" as defined
3. It sounds like you are interested in running "Ultimate Time Attack" where anything goes for what your character already has, but it would not be appropriate for this leaderboard due to the issues above. "Speedruns" generally put people on equal footing.
4. There is a separate Time Attack community around that handles these types of runs (existing characters) better so you may find better advice seeking them out.
5. Consider another category. Ver2 doesn't get a ton of attention but All Missions Glitchless is a stable*, if long, run (~8 hours). Any% Glitchless is probably ~3.

*Well, remember to save and quit at some point...

All console versions of the game (except Plus and maybe GC 1.1) have item duping. v1 let’s you clip through doors with attack combos. The Dreamcast versions force you to activate the ruins unlock pillars on a second iteration of the first three areas (respawning mobs and locking doors but preserving loot on the ground). That is mitigated in v1 because you just clip through the ruins entrance door anyway so you don’t ever go out of your way for a pillar.

The numeric mechanics in Dreamcast are different, notably RAcast cannot equip Saber at level 1 until v3. This significantly slows things down for RAcast and you’ll observe the best run is a HUcast instead, while RAcast runs are currently opting to stack a second time. Weapon attribute% works somewhat differently and you may be stuck buying weapons with -%.

No quick menu in v1 makes the game more physically demanding for doing things like switching weapons.

Overall Gamecube is the more accessible version, and faster for various reasons. v1 is fun because of the unique fun things it brings and just being a different flavor. v2 trades door clipping for a quick menu and it’s not a good trade for speed so no one does glitched runs for that.

For Glitchless, it’s mostly v3 for the same reasons and also the Forces are a lot weaker and physically slower on Dreamcast.