Great WIP once again, Kriole.
Too bad you guys still didn't realize that making the first jump as high as possible in the Gravity Statue Spike room would be a bit faster in the end than trying to land in the spikes as soon as possible. My YouTube demo of this room is apparently not of much use for you... Other than that nice work with many improvements already. Keep it up!
For more news, some of you have eventually heard on IRC that I mentioned a kind of testrun I have finished a while ago, and today, I finally want to present it to you:
Super Metroid - 14% speedrun in 0:27!This run has a pretty long history that would probably interest some of you, so I thought I should share it:
I started to work on it in July 2007 after getting somewhat tired of Redesign. It was intended to actually be a quick testrun using the Ice Beam route, which I always felt like being the faster option than the Speed Booster route moozooh claimed to be faster all the time, so I just wanted to finish it to find out what time I could get with this run. In the very beginning, I didn't bother to micro-optimize Ceres and some of the earlier parts of the run. In fact, I didn't even use some of the tricks from my old any% in it, most noticeably the new OTES climb strategy, Gravity Spike Tech, and Norfair Walljump, because I actually planned to present this run sooner than the fully optimized any% and thus didn't want to spoil the tricks at first. Everyting until the WS lake went very quickly, and I was happy to finally face a new challenge nobody ever passed before in TAS conditions:
The WS-Lake HBJ without collecting the Missile pack!
It only took me ~2 days to fully optimize, and guess what, it only required ~7 sec to get to the other side of the lake (for comparison, moozooh thought it requires ~25 sec, and hero even assumed it to be a 2 min delay at first :-P). After that, I started to take the run more and more seriously, finishing Phantoon in 2 quick rounds while only having one Missile pack, and even spoiling a few strategies in WS such as the attic one. At the Red Brinstar shaft followed the probably last obvious mistake of the run by missing the hero-knockback from the Ripper to enter the next room a half sec faster (a trick that wasn't yet public at that time), but from Kraid's Lair on, I consider the run to be more or less flawless (with the exception of the Norfair Walljump I avoided to spoil intentionally).
I did alot of tests and deep calculations ahead to roughly guess what time I would likely end up with all the run conditions count in (at this point I want to thank hero for his any%v1 0:25 run, which was of great help to me, because it basicly used the same route until the end), and after Ridley, on whom I used PBs and even Ice shields to make a pretty entertaining battle overall, it got more and more clear that I would most likely end up with a low 0:28. I started to regret to have neglected some of the micro optimizations from the early parts of the run, which would be badly needed for the quest to break the 0:28 barrier. Despite of all the previous hard work to get excellent luck in LN, I played with the thought to start over completely, but finally decided to take the risk and still try to go for the 0:27 time. The production pace of the run started to slow down massively, as I from now on took much more effort to optimize the run. Nevertheless I enjoyed the new speedboosterless conditions in Maridia, which although made that part a pain in the ass to optimize, allowed me to develope some really nice and still fast strategies there. The highlight was of course the Draygon fight, to which hero once said the following quote when mentioning the "Pros of the Speed route":
hero of the day wrote:
MUCH MUCH MUCH faster draygon fight. Since your ammo is horrible in a low% run, shinesparks can take out draygon in a very fast 2 rounds.
Well, how about taking out Draygon in a nearly same fast 2-rounder with that horrible ammo nevertheless? :-)
With extreme shot precision and luck manipulation it's still possible, you just have to hit Draygon and his gunk for refills with the same projectile all the time.
Anyway, after exiting Maridia, manipulating all refills from the Metroids without delays, escaping the big Metroid as shown here, the Mother Brain fight and the speedboosterless Zebes escape, I finished the run in early January 2008 with a completion time of 0:28 that is only 2 seconds away of 0:27, just like I feared. It pissed me off, so I decided to just put that run aside. Then, in late June 2008, when the new 6% NBMB TAS came out, Cpadolf presented a new big Metroid escape that was only possible by abusing Samus' hitbox through morphing in the right moments just before the Metroid grabs her, which was something I couldn't even notice without using the memory watcher in first place. This new escape, although requiring a PB which would slow down the MB battle a bit, should, as I calculated, still be faster in the end, and enough to cut off the last 2 seconds missing for the 0:27 to make my goal come true again. So I picked the run in July 2008 once again after a half year, and redid the Big Metroid escape using Cpadolf's technique, while even managing to improve it by a few more frames. The rest of the run was more or less just copied from the previous work, and just because of the new Metroid escape, I finally got a extremely high 0:27, only a few frames away before switching to 0:28!
Because of this fact, I can now say for sure that 0:27 is definitely the absolute limit for a 14% run, no matter what you do. Since this run got absoleted by the new 6% NBMB / glitched any/low% anyway, I won't bother to improve it anymore just to save a few seconds in realtime at most.
To sum the technical details up:
- This run is ~7,6 minutes faster in in-game time than Terimakasih's classic 14%, or even ~8,5 minutes faster in realtime. So I not only reached the fastest possible completion time on the game clock, but also a optimal realtime by skipping all Pause Screens/Menus except when it was unavoidable for the Gravity Jump trick, making the run even more entertaining to watch.
- All Boss strategies except of Kraid are new and unique, developed for this run exclusively, yet they are performed as optimal as possible with only a minimal margin of error (a few frames at most) due to unavoidable luck manipulation.
- Health and ammo management is excellent over the entire run, especially in Lower Norfair and at the Metroids, where it took me alot of work to optimize.
And to clear up the last questions about the Ice VS Speed time and route differences:
No matter whether you take Speed Booster or not, going for early WS with my HBJ technique is a must in any case. Getting Speed Booster early then returning back to WS, like moozooh planned in his run, is much slower due to the long route detours combined with the additional, slow elevator rides. Although having Speed Booster would save a couple of seconds in WS, it would still be far away of reaching the time losses to get it early.
Because of this simple fact, both routes would be equal until entering Norfair, with also almost same detours to get Ice or Speed. So basicly, the true Ice VS Speed comparison only starts at Lower Norfair. Thanks to hero's any%v1 in 0:25 TAS, I could compare the time gains of Speed Booster in every room pretty exactly, until the very end. In total, with all the possible improvements count in, and with the delays of PBing the Metroids, the Speed Booster only saved 70 seconds at most, and this without even counting the recharge station / ship refills you would probably have to use at least once, wasting 10 or more seconds additionally, and therefore resulting in a final gain of just ~1 minute in the end.
The Ice Beam however, already saves ~100 seconds at MB alone, and another ~35 sec at Ridley.
So in the end, the Ice route is (135sec - 60sec) around 75 seconds faster than the Speed route, as unbelievable as it looks, and therefore the absolute limit you could theoretically reach with the Speed route would be a mid till high 0:28 at best or even a low 0:29 at worst, which, to my surprise, moozooh once again failed to realize.
Anyway, I hope I could clear everything up regarding the 14% run now, and I want to appologize to keep it secret for so long. I was just curious to see what some of you would be able to achieve by themself, and I didn't want to kill your attempts with my run too soon. Seeing that nobody works on a 14% run anymore, I thought it's the right time to release it now. Maybe I will eventually upload the smv on my site as well, or just submit it here as an almost 9 minutes improvement to the previous run, but seeing that it's obsoleted by the new 6% run, it probably isn't necessary.
So yeah, enjoy the run!
Uhh I didn't read the entire novel you wrote there Saturn but to burst your bubble me and Kriole got Phantoon in 1 round with 1 missle pack, not completely, but we deemed it possible after some tests.
Furthermore, after me and Cpadolf had done Phantoon like 80 times in our any%, we he also agreed it could be done! =D
Watched the entire run.
Good freakin God Saturn, I deem you the God of manipulation. That run was so insanely awesome words cannot describe it.
All your bombjumping was so amazing.. Just.. wow. >.<
I also now know your sekret, it isn't more or less flawless, there are still many little mistakes myself and Taco found. (by yelling and screaming at each other while watching the run)
Now Spoofer DEMANDS SMV now!!!!!!!!!!!
Trained by Cpadolf. Mission: To Perfect.
Hero says: Yeah bro, I almost went super saiyan once
My SM-RBO
Current Project(s):
1)I don't TAS anymore! Pay me!
fuck youtube is being sooooooo slow on part 1/2 for me -_-! damn i hope you up a smv somewhere
the best thing so far after reading your comment is the bombjump GG dude :D thats cool
Dam guys give Saturn his moment of glory, he did do some really amazing stuff. SMV-wise, yes it would be nice to have, but its his choice, and bitching at him obviously doesn't work. >.<
Trained by Cpadolf. Mission: To Perfect.
Hero says: Yeah bro, I almost went super saiyan once
My SM-RBO
Current Project(s):
1)I don't TAS anymore! Pay me!
Joined: 8/1/2004
Posts: 2687
Location: Seattle, WA
Normally, we wouldn't be so hard on him, but he keeps making statements like this one.
Saturn wrote:
a game completion of only 0:27 minutes, which is the fastest time possible for this kind of run
It's pretty arrogant, and almost all definitive claims he has made along that line have been thrown back in his face with improvements that he did not predict. That, and there is a lot of history of him raining on other peoples' parades, especially when they proved said statements incorrect.
well its true saturn maybe incorrect with that statement.. but still cant we quit that flame? saturn likes those words possible... impossible etc dont take it too seriously
as spoofer said he did nice moves and finally dared to show more than a wip or a part/rip-off, its his first full metroid run we see ;)
edit : what tha heck down there!! xD
Yeah, I read up on all that awesome drama stuff.
Eh, Since he did just barely break the 28 barrier, and I'd estimate a max of about 10 seconds assuming new improvements are found.
Ofc, a better route choice could prove useful as well. =o
He may actually be right on with that statement, ofc, I'm not 100% agreeing, just SEEMS likely.
EDIT: Yo, make sure EVERYBODY reads Evilchen's post above. EVERYBODY MUST READ IT!@
Trained by Cpadolf. Mission: To Perfect.
Hero says: Yeah bro, I almost went super saiyan once
My SM-RBO
Current Project(s):
1)I don't TAS anymore! Pay me!
Great WIP once again, Kriole.
Too bad you guys still didn't realize that making the first jump as high as possible in the Gravity Statue Spike room would be a bit faster in the end than trying to land in the spikes as soon as possible. My YouTube demo of this room is apparently not of much use for you... Other than that nice work with many improvements already. Keep it up!
Don't really get your reasoning. So jumping is faster than running? Think I'm gonna quit TASing =(.
Our last attempt was improved by a frame already.
Cheniways, really nice low%. I think the highlight was the WS lake, I bet you got that idea from your redesign run. It surprised me that you opted not to use most of the bigger timesavers, but still you claimed you struggled for a 27 minute time. This is probably where you pondered if you should restart.
Congratulations on finishing your first TAS of Super Metroid ;]
Joined: 11/14/2005
Posts: 1058
Location: United States
Very nice run saturn. It almost makes me sad to see that the 14% run is no longer a category on this site. I liked how it felt like samus could not be stopped by anything in her way. Each room was handled pretty darn gracefully, in stark contrast to the clunky way in which the old 14% run was played. Drops were collected with such minimal delays, which is a key factor for maintaining the flow of a run. The only thing that bothered me was how many tricks you intentionally left out of this run; it seems kinda selfish and unnecessary imo.
Some points I am curious about:
What did you do differently in the ws attic? Was there some thing special, caused it looked like every other run except for the fact that you killed the hopper dude with missiles.
In the ws basement, you showed off a pretty neat tech. Mockballing into the hole and laying a regular bomb instead of using a PB. Was it faster?
Like namespoofer said, I think phantoon can be killed in a single round.
I noticed a pretty big shortcut in the large acid rising room. Would be beneficial to an any% run. Nice find.
Draygon fight was great. Proving the speed booster to not be needed here at all. Now you gotta find a way to 1 round this guy ;)
I noticed a damage boost in the last metroid room, useful to other runs? I noticed a damage boost in the 4th escape sequence room, beneficial to other runs?
Damn nice run you got there Saturn, happy to see that you are actually capable of finishing a run. The missile lake HBJ was one of the most amazing things I have ever seen done in Super Metroid, pure sweetness.
Thanks everyone.
My 14% run is also now available in a high quality AVI:
http://www.archive.org/download/SuperMetroid_TAS_14p_27_Saturn/SuperMetroid_TAS_14p_27_Saturn.avi
(103 MB)
As for the smv, I will think about releasing it in the near future.
Zurreco:
There is no doubt that 0:27 is the fastest possible time for a 14% run. Getting a 0:26 would require to improve my run by a whole minute, which is just too much to be possible.
Kriole wrote:
Don't really get your reasoning. So jumping is faster than running? Think I'm gonna quit TASing =(.
Just tested it, and it actually doesn't make a difference, because when I jump high on the first jump, I can do the 2nd jump much lower, while you were forced to do the 2nd jump with max height instead. It equals in the end. The reason your attempt was slightly slower is, as you apparently noticed already, because of a slight inaccuracy during the canceled damage boost.
Kriole wrote:
Cheniways, really nice low%. I think the highlight was the WS lake, I bet you got that idea from your redesign run. It surprised me that you opted not to use most of the bigger timesavers, but still you claimed you struggled for a 27 minute time. This is probably where you pondered if you should restart.
Right, my bomb jumping knowledge from Redesign definitely helped me at that lake. And yeah, I obviously didn't know that it would get so close with the timer at first, otherwise I would of course take more effort to optimize the run in the early parts.
hero of the day wrote:
What did you do differently in the ws attic? Was there some thing special, caused it looked like every other run except for the fact that you killed the hopper dude with missiles.
Yeah, it's the Ki-Hunter kill. Wastes a Missile in the end (the Super can be refilled back or exchanged for a big energy drop) but saves a few frames because you can start running sooner to the left.
hero of the day wrote:
In the ws basement, you showed off a pretty neat tech. Mockballing into the hole and laying a regular bomb instead of using a PB. Was it faster?
Nope. It ended up 6 frames slower IIRC. The reason I used it was for experimental sake because my previous tests of this strategy weren't fully optimized. It wasn't much of a loss, so I didn't bother to redo it again.
hero of the day wrote:
Draygon fight was great. Proving the speed booster to not be needed here at all. Now you gotta find a way to 1 round this guy ;)
Impossible, sorry. :-)
hero of the day wrote:
I noticed a damage boost in the last metroid room, useful to other runs? I noticed a damage boost in the 4th escape sequence room, beneficial to other runs?
The Metroid boost was forced, because I had to wait for the last refill. I didn't test, but it might be possible it saves even more time in the next room than it requires in the previous room to wait. Needs to be tested to know for sure. As for the Escape boost, it's at least the fastest option in a speedboosterless low% run. In your any%, you are invincible for a while due to a hit in the previous room after the spark, so you can climb through the pirate on the left wall immediately. Should be around same fast in the end, I guess.
And regarding Phantoon 1-round with 1 Missile Pack, maybe it's indeed possible, but the delay you would require to have all refill drops in the right position before the Missile-Wave would probably end up as long as just doing 2 quick rounds. The 2-rounder allows for some more important refill drops though, which should also be considered.
Any time I read one of Saturn's "this is impossible" sentences, I just mentally add "unless new tricks are found" and it all makes perfect sense. I'm a fan of never saying never, myself.
Anyway, that's besides the point, this is an excellent run and I don't see any reason to fault you for neglecting optimization considering it's not a publishable category. Good work.
Any time I read one of Saturn's "this is impossible" sentences, I just mentally add "unless new tricks are found" and it all makes perfect sense. I'm a fan of never saying never, myself.
Anyway, that's besides the point, this is an excellent run and I don't see any reason to fault you for neglecting optimization considering it's not a publishable category. Good work.
I think I missed something somewhere...since when is 14% no longer a publishable category?
It's all Cpadolf's and Hero's fault we don't have a legit 14% low cata anymore. =(
Trained by Cpadolf. Mission: To Perfect.
Hero says: Yeah bro, I almost went super saiyan once
My SM-RBO
Current Project(s):
1)I don't TAS anymore! Pay me!
Arguably the 14% run is a "non-glitched low%" (or is that 15%?). But from what Saturn was saying about the end of the run (I haven't yet watched the run myself), it sounds like he did use the glitch techs from the 6% run at the end, so I guess that's out.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
