Yea sorry I'm more used to SDA's system where its a straight accept/reject. Everything else looks good and the trick is near the end of the run so it should be fairly simple to add in.
So I happened to be awake late and noticed this got posted. Sadly, Dash Galaxy has had somewhat of a speedrunning history, and clearly some of that was overlooked. I'd recommend the author check the game page on SDA for the specifics, but a major glitch was overlooked that saves a bunch of time on floor 24. http://speeddemosarchive.com/DashGalaxy.html
Considering the SDA run is several seconds faster than this run (after adjusting for timing differences between TAS and SDA), I feel like this should be rejected for now.
MURPHAGATOR, did you consider making a TAS for this game yourself? This way you're likely to learn even more about the game (which would help you to improve real-time record as well).
I considered it, and actually had worked on it for a while. I found the process really tedious and didn't enjoy it that much, and when I got to the helicopter and realized what I needed to do to truly optimize it, I lost interest. I actually still have to movie files I had made, but considering the approach to fights I used at the time was fundamentally wrong and I basically ignored the off-axis jump, its not particularly well optimized. I definitely continued using the emu tools to learn more about the game when I did attempts. I do appreciate your enthusiasm, but I just don't feel like TASing is for me.
I also don't really have any interest in improving the real-time run because a vast majority of this game in real-time is praying that the enemies co-operate by actually taking hits, and not randomly getting throws. In all honesty, after seeing this TAS, there's only maybe 2-3 things I'd even be able to consider changing real-time, and they'd only save a couple seconds total, tops. The rest of the differences are perfect luck with fights, which gains a couple seconds at a time, and some of the off-angle jumps that are really difficult, save very little time, and missing results in the run being over instantly.
xxNKxx wrote:
MURPHAGATOR wrote:
One side thing I thought of, but is probably irrelevant: a lot of the maps wrap horizontally. It's most notable on 5-2, if you jump off the platform you start the room on, you can see the exit to the room. I very much doubt you could make it over that gap, but I'd never tell a TASer that something like that is impossible.
huh? I don't understand this. Are you mean make it same my last jump in 6-3? If that, it can't, because it's not same 6-3. Can you help me make a example for me? Maybe I'll remake if it's possible for faster
No, it's not like the last jump in 6-3 at all. Basically, at the VERY beginning of 6-2 and 6-3, before going anywhere, if you jump off to the left over the spikes, you can see later parts of those areas. In 6-2 you can see one of the platforms to the right of the conveyors just before the door, but it 6-3 its most striking since you can actually see the door that leads to 6-4. The jumps are WAY too far to make by just jumping normally. I experimented for a while when I first realized that, but I wasn't able to find any way to abuse that. If you could find a way to like, walljump or double jump or something, you could skip most of 6-2 and basically all of 6-3 by going over those spikes. Depending on what exactly gets found, it might be possible to skip parts of 5-2 as well, but I really have no idea if something like this is even possible at all.
Sorry it took me so long to respond to this, I've been out of town for the last several days. After watching the updated run, I'm really happy you decided to watch my run and use it as a reference for some improvements. The movements look a lot cleaner, much more like what a TAS should look like than the last one.
There's a few places where you make minor movements and some frames might be able to be gained, steering around the guy at the end of 3-1 comes to mind, but I don't really know to that much detail. The only other thing I can say is that the helicopter in stage 2 can probably be optimized slightly more. The pattern you get here is just repeating the real time strat, which is still a lot better than the old one you used, but I feel like the helicopter could be manipulated into dropping bonbs even faster by moving in precise ways since it's patterns are based on x/y position to at least some extent. If I had to guess based on my knowledge of the game, the helicopter is where there is most likely any real amount of time to be improved still. At the very least, I know there's a position where it will drop 3 bombs on its first cycle returning right instead of 2, but I couldn't tell you much more than that.
That said, I would probably vote accept instead of reject this time around, this really looks a lot more like what a TAS should look like.
Also, I already knew they would, but the TAS fights make me super jealous. Losing several seconds each fight in real time is painful.
feos wrote:
So what about this suggestion:
MURPHAGATOR wrote:
6-2 is also probably the only place in the game where taking damage might save time on the conveyor section. Without the pause thing it would save time during 3-2 since you wouldn't have to wait for the waterfalls. I took damage on purpose several times during the platforming, but those aren't really relevant to a TAS since they were mostly for consistency.
Yea I wasn't really expecting the damage on the conveyors to save time necessarily, it was just the only place I could think of where it might.
One side thing I thought of, but is probably irrelevant: a lot of the maps wrap horizontally. It's most notable on 5-2, if you jump off the platform you start the room on, you can see the exit to the room. I very much doubt you could make it over that gap, but I'd never tell a TASer that something like that is impossible.
http://speeddemosarchive.com/Rollergames.html is the final run that I did.
I actually ended up watching the rest of this out of curiosity later. The biggest difference between the run that I did and this TAS is that I can't really rely on the jump kick to hit (it has about a 50% chance to hit regular enemies, and I found no way to manipulate it in real time), so I have to stack enemies for the special attacks to kill. I had thought that doing the same would be faster for a TAS, but I never exactly timed it anyways. There's a few jumps that I didn't do, the diagonal jumps on the stage 5-1 bridges coming to mind immediately since those are extremely difficult jumps in real time, and the 6-3 jump into the door since I didn't know about it, although I'm not sure I'd risk that jump in real-time anyways because the diagonal walls are really iffy.
I did also notice when I watched the full thing that this TAS didn't fully utilize the conveyors in 6-2 to gain speed by staying on favorable conveyors longer. 6-2 is also probably the only place in the game where taking damage might save time on the conveyor section. Without the pause thing it would save time during 3-2 since you wouldn't have to wait for the waterfalls. I took damage on purpose several times during the platforming, but those aren't really relevant to a TAS since they were mostly for consistency.
a/ Not any difference. I just like a girl with pink color more than :)
b/ Yes, it have limit. At every sub-stage, I only can use backflip kicks (A+B) in 3 times only. It'll reset when start new sub-stage
Hey, so I was really excited to see this run in submission since I did a real-time run of this game a little back, but I was disappointed to see that a lot of things got missed. I know the racing line got mentioned already, but I can also tell you haven't put time into the game since your first comment is completely wrong. The characters have completely different move speeds, their basic attacks work differently, and their special attacks work differently. I'm not going to put a whole lot of detail into it, but in the end the girl is the fastest anyways, especially in a TAS situation.
The other thing I can say is that I just stopped watching at the helicopter. The helicopter drops a set number of bombs and the way it drops them is based on your position. I never went in depth enough to determine exactly the best way to do it, but what I can say is that my real-time run completes the auto-scroller several seconds faster because of this knowledge alone. Edit: I timed it out, my positioning saves 5 seconds compared to the method in this video, and there could be more time to gain by moving specifically to manipulate these drops, which is the main thing that stopped me from attempting to TAS this in the first place.
I didn't really watch all of this, but just based on what I saw in stage 1 and 2 I would easily say no That said, I can't vote because of the minimum post limit, but I felt my opinion might help others.