United StatesWhen I time, I usually time it on the first frame that I see the boss health value transition to becoming 0 like you say - as a few of the videos seem to use that as the metric, but now that I look through, you're right it does seem to be less than consistent.
I usually try to use consistent visuals where possible as that tends to keep things more clear for timing, but it doesn't much matter to me. There's visuals for the final hit (either unit shake or hit flash are standard on all versions) and the start of the damage numbers (visible and pretty consistent on all versions) too that could work just as well.
Yeah, in my run, I didn't actually know Magician could go that high without literally perfect cards. Resistances are very broken on this version though, only reason why the damage rolls seem to go so high. Considering the work that went into the Human theory TAS to get that possible, it's probably quite literally a 1/25k chance of getting it in a run. Bizarreluck was really lucky to get it, even with a hermit!
Weirdly enough, despite my hand issues, I'm really good at getting frame perfect timing. Don't know if it's just that I can figure out strats like sound that work or what.
For example, in Castlevania Easy Mode I use my peripheral vison to judge the first bat movement and hit the frame perfect bat crit, and I can judge not only what pixel I need to stand on for a first frame Medusa crit (again, a peripherals trick), but use a music cue there to know when to whip. My anti-despawn map object tech that I use to try to kill bosses with in Vortex (SNES) is again a music cue, but due to that game running at like 4 fps, sometimes you just can't get it.
Since this really is a full manipulation, I think I actually needed to hit something like a 10 frame window out of Warren's tarot reading to get all the RNG I need, so I think I figured I really couldn't waste any time selecting anything but the 1st option. There's no good way to tell if I lost frames anywhere. Since everything I do on the in battle map is done to retain some specific in battle RNG so the card does enough damage and the Tigerman does hit, the only concern was getting the timing fast enough and not taking too much time moving units. There's a surprising amount of leeway in movement, timing, and map position to get exact same RNG. Only the Card pattern is actually frame perfect, I believe.
If we could fully crack how "lag" works - and by extension, how pathing works - there's another 4-5 seconds to save, if not more. I took a brief look but what I found in emulator did not seem to match on console, so I scrapped trying things specifically to reduce "lag" in runs.
Someone could also just do exactly what I did only faster or with slightly more optimal routing. I'm fairly sure the TAS movement might be possible for really amazingly handed humans, it would just require routing for it in a way that's SOMEWHAT reproducible. Something I wasn't really able to do unfortunately.
Fair enough - that's actually a good point!
Interesting - I've only ever had runtime errors, so that's good to know. Yeah, I agree, given that knowledge, option 2 seems like a bad idea and would be too beneficial to for folks. So much so that someone might try to exploit it.
I think in this case, it would be relatively easy to discern a real crash as it always crashes due to run time errors that are visible. We could exclude other events that look like crashes and make it a rule to capture the crash runtime error by either capturing desktop or switching to the error message in capture.
Perhaps we could be lenient in areas where there's no benefit - like crashes that happen coming out of a level since it's so common - there wouldn't be a benefit there as it always takes longer to restart the game and you've already passed the screens that take the most time. Plus you haven't seen any RNG yet. But that's the only case I can think of.
I do agree though, all of the runs that are under an hour, I see no need to include in this. There's almost never any crash that occurs in those due to where they fall relative to the amount of cutscenes, and ample reason to believe there might be attempted abuse in shorter runs.
@xSoapBubble recently brought up a good point - how does everyone feel about adjusting times for game crashes? I personally think this is pretty fair for the longer runs given how frequently this game implodes, but I have some times that would be affected by this drastically so I might be a bit biased.
If there were a way to fix it I might be less inclined - but there really isn't. I've tried an extensive number of things to get it running in a more stable fashion, and nothing works. You can take a look in the discord for a history of what's been tried. Might have more folks willing to run the game knowing that a crash won't affect them as much.
As for the timing we could do any of the following:
-
Deduct the time of the crash to the start of the selection of the next level.
-
Deduct the time of the crash to the start of the selection of the next level next level. If the crash happens in a level, deduct he time spent in the incomplete level from start to the restarting of the same level.
We could also switch to in game time, but I'm a little wary of that as it might lead to folks min-maxing the item system without consequence which isn't really in the RTA spirit, and it would also be incredibly rough to do for the runner and the mods on longer runs.
Let me know what you think!
I don't see why not personally. Though I do question if we'll use all the categories. Stuff beyond Any%, Best Ending, Devil and the mini-runs likely won't receive much intentional activity. I'd almost recommend we use the Level leaderboard feature, but it may not be worth it or too awkward.
Anyone have any strong opinions for/against doubling up like this? Or opinions on the board stuff?
Mod Note: time starts at the Excel splash screen being visible, so 0.55 and ends on 30.84 when the Dev room credits are fully visible, so time comes in at 30.29. GG!
The TAS has recently been hardware verified on a TAStm32 to a Toploader and Battle Chess cart!
I'm getting 1:14.316 in avidemux interestingly enough.
No prob! It's a bit surprising to see a run here, and on the Saturn no less! It's more of an honor to review an actual Saturn run, I assure you. =)
Hehe, I'm glad they were helpful! I have a bunch of info in the forums and even a video on how the initial game RNG actually works in all versions of OB:MotBQ that might be helpful as well.
Saturn is one of the versions where it's sort of possible to manipulate your starting frame (i.e. the frame you select the leader's sex) to get a certain pattern of cards, but considering that the loading time on CD based systems can be variable, it's best to just notice the pattern and try to plan accordingly. I also have monitor and initial card pattern mapper bizhawk scripts for most of the versions, but I'm not sure I ever finished the Saturn mapper due to having to retrofit Japanese text into the naming part of the script.
It's probably less of a hassle to just do like you're doing and try to pick out parts of the music that have generally good patterns anyways, that's what I do on PSX and it'll be far more effective on Saturn because the pattern doesn't change when you answer questions!
GG! Looks like you started your timer a little early so my retime puts this at 3:41.80!
For a better comparison please feel free to view my hardware tests on DS3s loading:
Additionally there is a SIGNIFICANT speed difference on 90k PS2s - i.e. 90k might be on par with Vita/PS-TV, which is the fastest in most situations.
This is the nature of the platform unfortunately. We could try to remove loads to make it fair, but the effort might not be worth it considering how few runs there are.
Ah Izzy and Aqua's videos are dead... I'll try to find an alternate source for those.
Ooh neat! I'll have to play with that.
It's occasionally possible to get into the 120s with Chariot on SNES (without changing tactics) though the in battle RNG and the Stage RNG windows are very unfavorable on the seeds I've explored. There's some possible rolls with magician too, but again the windows are way too tight for humans I'd think. We're talking a SINGLE pixel to avoid the witch and keep a good in battle seed in the ones I explored.
There's also two hop manipulations to the boss that are fairly timeloss neutral, but they tend to make things harder to manipulate when you're TASing for good in battle RNG. I had a human theory TAS that's better than what my current time on SNES is, but I found I was actually much less consistent when I did runs with it because it was harder to judge the distances with two hop. I also very much doubted it was feasible for me to do - the windows were too tight for what little time it saved.
Cursor position is hard for me specifically. Even my current fully manipulated time on SNES I consider a fluke, and a lot of the manipulation was designed to cater to me making movement mistakes. I messed up zooming in to my cursor position, was sure I blew the run, and got so flustered I forgot to move the cursor down. I got lucky and moved down in time when my unit was still barely in RNG range. I have a trick from my fighting game days where for the second part of my unit movement I hadouken the cursor in position and I messed that up as well, but my position ended up being close to where it should be.
I think there's about 3-4 pixels on stage x and y where the RNG is the same, in some situations, like terrain changes, it feels like less.
Just having things separated into Console is wonderful! Thanks!
I don't believe anyone but @bizarreluck and I ever ran PS1. Or anything else other than SNES NA.
This is a human theory TAS of Fireseal mode. The RNG to get this is possible for humans but very difficult. This is one of the highest damage roles possible for the Devil card in this situation and it is rare for the Valkyrie to both go first and not miss. That said, it is definitely doable for us fleshes.
There are other ways to achieve this 2 card 1 action route but this is probably more likely for humans to achieve.
With PS1 loads: 133.80 = 2:13.80 With corrected PS2 90k loads: 126.93 = 2:06.93
I'll link a copy of the TAS in resources.
No I'm not interested in doing a non-human version.
Retimed to 7:48 and corrected submission to the right catagorey. GG!
So, it would be possible for a human to do this to match the TAS exactly.
Highly unlikely, but possible. Lot's of frame perfect actions and things that would be rather difficult on a standard NES controller, but since someone got diagonals in a run recently, I figured I'd mention buffering and more efficient start presses can achieve this:
So if you're crazy enough to go for 30+ frame perfects... =P
I think it's likely that someone will get a 1:15 one day on NES 3D.
