SDA logo

Metroid Prime 2: Echoes was released in mid-November 2004. Samus lands on the planet Aether to investigate missing Federation troops and finds a world with a split-personality. In the dark version of the world, she takes damage just from the air and faces an evil black goo called the Ing that can possess all sorts of creatures.

 

Category Note: Entering secret worlds is considered a separate category.

Return to the Game List, the FAQ, or the Home Page.

1:38 by Alastair Campbell, done in 22 segments.

Author's comments:

I'd like to dedicate this run to the late Mark Haigh-Hutchinson.

Several factors contributed to my attempting this playthrough. Firstly, having helped to run Speed Demos Archive for over a year by this point, it was something of an embarrassment to me to be the only administrator never to have had a run listed on the site. Secondly, the training I'd received and the new tricks I'd found in the process of executing my "banned", out-of-bounds abusing low% run in Summer 2006 made me realise that the incumbent 1:43 any% could be beaten without too much difficulty. I also knew that I was the only player likely to attempt to improve this, because nobody else has ever seemed terribly interested in taming this game. Thirdly, a post on the SDA forums claiming that skipping the Grapple Beam requires one of the hardest tricks ever performed in a speed run annoyed me intensely, because this is simply not the case. Finally, the emotional problems with which I was having to deal in the last few months of 2007 left me desperate for something in which I could immerse myself. While running Echoes is probably not the most sensible activity to undertake when one is struggling to maintain one's sanity, it is nevertheless arguably better than sitting around doing nothing.

Originally I wanted to have a crack at the single-segment any% category, because I felt this would be easier to beat than the segmented, and would probably have been more fun. My intent was to try to demonstrate that the above "impossible" Grapple skip trick in Sanctuary could be performed with sufficient consistency to make a single-segment Grapple skip run possible. I practised for a while, but Boost Guardian caused me problems, even with four energy tanks. I also found that landing on the robot in Grand Abyss wasn't as easy as I remembered — it seemed my aptitude for timing this trick had atrophied in the 18 months since my previous run. Realising that a lot of restarts would have been required, I chickened out and decided to have a bash at the any% segmented category instead.

The 1:43 by 'Silent Echo', while a revolutionary and excellent run for its time, had a well documented set of shortcomings. A number of very cheap strategies to improve it had been known for a while — most notably a couple of route changes in Sanctuary and the use of Screw Attack to defeat the final bosses would have saved about two minutes. I felt that by incorporating some smaller time savers and playing more smoothly I ought to be able to grab another minute. My initial goal was 1:40. Assuming 20 segments for the run, that meant I would need to gain 3 seconds per segment on average (a target which quickly proved to be far too conservative). In a game such as this, scientific methods are vital, so I created a spreadsheet for each segment and manually timed the old run, noting the number of frames in between cutscenes and performing a similar process for my own attempts. If possible, I wanted to go faster in every single segment, a desire which I almost, but not quite, fulfilled.

The route is largely the same as the old run, skipping both Dark Suit and Grapple Beam. I open up the gate blocking the Watch Station portal before collecting Screw Attack, which allows me a faster way to return to the light world after the Quadraxis fight, saving 10 seconds or so compared to the previous run. In the same segment, I take the much faster route suggested by XtraX to leave Sanctuary, saving a further 25 seconds; routewise, everything else is as it was before. I considered whether a faster Sky Temple Key route might exist and juggled a few alternatives, but decided eventually that the existing route was probably as good as it was going to get. I also used two more segments than the previous run, saving in Great Temple before the Alpha Splinter fight, and again in Great Temple before heading to collect the final Sky Temple Key in the dark version of Samus's landing site. This save is so cheap that it seemed criminal not to use it every time I passed through the room.

In previous Echoes attempts, runners have made use of the checkpoint after Emperor Ing to achieve a fast Dark Samus 3/4 fight, then edited the videos together to make it look as though these two bosses have been beaten without retrying. I have never liked this practice — it seems inherently deceitful. In this playthrough, I have presented these two boss fights as separate videos. Since players can try Dark Samus 3/4 as many times as they like with no time penalty (although they are not able to save the game to a memory card at this checkpoint), this counts as a segment as far as I am concerned. I have listed this run as being 22 segments.

This run was played on a GameCube, and the European version of the game was used. There are a few differences between the North American and European versions, but they are not significant enough to warrant separate categories. PAL Echoes is a PAL-60 only game — since my DVD recorder is not capable of capturing a PAL-60 signal, I booted the game on a North American GameCube using a Datel Freeloader disc, and captured an NTSC signal instead. This does not affect gameplay in any way.

Echoes is quite buggy. Some of these bugs are decidedly advantageous to a runner, others equally definitely not. Readers unfamiliar with Echoes chicanery may benefit from the following quick overview of some of the more salubrious glitches I exploited on my way to liberating Planet Aether in under 100 minutes. All of these tricks are documented in detail at metroid2002.com, of course.

Before I start on the segment-by-segment stuff, I will just mention a few people. In my 17% run I had a slew of contributors to thank because it was very much me running the game under the direction of others; this time, however, I did nearly everything by myself, so great swathes of fawning gratitude are unnecessary. Thanks, Sparky, Nate and Radix, for taking an interest. I'd especially like to thank Red Scarlet for keeping me company on IRC over Christmas and New Year while I tackled everything from Spider Guardian onwards, and for pushing me to get 1:38 rather than 1:39. As always, Scarlet was full of suggestions, none of which I used — sorry, Scarlet, but I appreciated them anyway !

So, that's that. Saving five minutes over the old run is something I should be happy about, but I can't help but feel slightly disappointed. If I'd had a few months rather than a mere six weeks to work on this, I would definitely have got 1:37. Even now I feel a desire to go back to Grand Abyss (where I still have a save) and improve from there onwards, but I think my time would be better spent on other things.

That 1:37 and probably even 1:36 can be achieved is not in doubt though. Quadraxis, Chykka, Spider Guardian, Emperor Ing and Dark Samus 3/4 can all be improved significantly. A bolder individual would go straight over the fence after Boost Guardian; a more cunning one would manage their energy and ammunition more cleverly. I make many mistakes in segment 13 and throughout the rest of the run which could have been eradicated had I been more patient. There are several speed tricks I didn't use because I'm a coward, and I'm sure there are still some tricks yet to be found.

It is difficult to know what sort of times one should be accepting when the game on which one is playing is so woefully undercontested. I've provided all the information I can possibly think of in these comments — 20,000 words' worth of it — specifically because I want it all written down somewhere for potential future runners to have available. Certain aspects of this game are very frustrating, and it is not as immediately attractive as its beautiful older sister from a speed running perspective, but there is still some way to go before we're looking at times approaching their optima. Simply seeing the mistakes I made in this run should be enough to convince you of that.

Still, this will do for now. Until someone improves it, I hope you enjoy it.

DJ

Single-segment 1:57 by Kim Siafa.

Author's comments:

This is the first single segment any% run for this game. I used the same route as the segmented any% run except that I don't skip the grapple beam, I've to fight the grapple guardian in order to finish the game. The sky temple key route is the same. I had to grab more items though, I grabbed more energy tanks and missiles than in the segmented run because there are too many dangerous bosses in the game, grapple guardian, boost guardian and Quadraxis are a pain in a SS run. I wasn't planning to get the energy tank in the room just below Chykka but I noticed that grabbing that energy tank was as long as going to the save point in venemous pond in order to refill my life. So, I grabbed the energy tank instead. Sometimes during the boss battles, I prefer going much slower because I didn't want to die. Health is more important than speed in some parts of the game.

The run went pretty well until I reached sanctuary fortress where the most difficult tricks are required. Agon wastes and Torvus bog didn't cause me much trouble, I had luck with the 2 underwater dashes in hydrodynamo station and with the ghetto on the laser thingy in Crypt. All the big mistakes I did were in sanctuary fortress. Performing the 2 mid air morphes in Watch station and in hall of comabt mastery is hell in a SS run. Hopefully, I had luck in transit station where I grabbed the early power bomb expansion expansion, this room didn't cause me trouble.

I think that a time of 1:55 or below is more than possible because there are big mistakes in this run. Luck has a big place in the final time. I didn't have very much luck in this run because many dark pirate troopers appeared in upper Torvus and made me lose time. This is a thing which can't be controlled, it's 100% random. Less mistakes and more luck could probably make the final time lower.

Low % entering secret worlds 1:55 by Alastair Campbell, done in 21 segments.

Author's comments:

This was my first true speed run, played and recorded in summer 2006. It is now eligible for submission to SDA following a rule change that legalises runs utilising out-of-bounds glitches, assigning them their own category. Of course, between the completion of the run in 2006 and its submission in 2009, I spent fourteen months administering the very site on which my runs were not legitimate. Bizarrely, in my role as SDA updater, I frequently found myself defending Radix's "in-bounds runs only" policy, and it was not until the site's current administrators softened SDA's stance towards this sort of beyond-the-walls chicanery that my 17% became compatible with the site — or rather, the site became such with my 17%.

I have always thought "Secret World" to be a wholly inappropriate name for what is in reality a hole in the game's scenic collision detection, but these unintentional portals that Retro had left lying around had been allowing players to view rooms from the outside in the Metroid Prime games for some time. Methods to use them to enhance sequence breaking and reduce completion times on the original Metroid Prime had been discussed and developed to some degree, mainly by Andrew Mills and Jonathan Wilson at the now-defunct samus.co.uk. The sequence breakers had left little still to be done, though, and so MP1's pickings were slim. For Prime 2, the developers fixed some bugs and introduced others, which inevitably meant that alternative approaches would need to be found to make serious inroads into the game's completion percentage or time; the secret world dabblers found themselves well poised to take advantage of this opportunity, having a well-established set of techniques to hand already. So, to long-entrenched vocabulary such as "ghetto jump" in the sequence breaking vernacular was added strange new terms like "wallcrawl", "skywalk", "door-warp", "room dump", "aether jump", and so forth.

This run brought together a lot of low percent theories which had existed to that point only as short, inexpertly deinterlaced video clips in diverse locations. I wanted to assemble from them a coherent presentation performed at breakneck speed, with the sort of respect for fractions of seconds that characterised the original Metroid Prime sequence breakers who had inspired me so much, and to shatter the myth that secret world techniques were too slow to be used in speed runs. Inevitably, my playthrough became one of the primary exhibits brought out by both factions in the resulting forum-borne wars over the eligibility of out-of-bounds techniques.

This run was also somewhat experimental in at least one other regard: It was one of the first to feature commentary, as an optional second audio track. Although sternn's Devil May Cry contributions carried audio commentary some time before this run arrived, my videos were important in continuing to popularise speed run monologues, and continuing to prove their workability both compositionally and technologically.

SDA did not officially allow multitrack audio productions in 2006, but this was no problem for me since my run was already ineligible for submission. This freed me from the compatibility problems that plagued the AVI and MP4 containers preferred by SDA at the time, and allowed me to release the run as a set of Matroska DivX files, which you should still be able to find at archive.org. For this re-release, I have of course made the now standard two-audio-track H.264 MPEG-4 — the videos are probably most easily played using VLC. So, if you missed it the first time, you too can now listen to what some have assured me is the quintessential speed run commentary (although to me it still sounds more like some English guy droning insufferably on about the unpredictability of the camera), and get higher quality H.264 visuals while you are at it. (This speed run was digitally remastered ... on my laptop.)

The run was played on a GameCube, and the PAL version of the game was used. Since my DVD recorder was incapable of capturing a PAL-60 signal, I played the game on an NTSC console, and booted it using a Datel Freeloader disc. This does not affect gameplay in any way.

I was already most of the way through this run before I learned of a glitch that I later came to use extensively in my 1:38 any%. Nobody had named it, so I called it the "rollshot". By hitting X and A at at the same instant, the player can unmorph and shoot simultaneously, without having to wait for the unmorph animation to complete. This is useful towards the start of the game for opening doors and activating portals quickly, although its usefulness decreases once the player collects the Boost Ball. This bug could have been used to optimise further the earlier stages of the run, but nobody had told me about it, so unfortunately it does not appear in this contribution.

I do not intend to write detailed segment-by-segment textual comments this time, since the audio commentary provides most of the relevant data. Of course, since the run was completed in 2006, some of this information is now out of date. Therefore, I will provide notes on each of the segments, in which I will correct any anachronisms in the commentary, and clear up a few other issues.

And so, to the segments. Times labelled with [c] denote errors in the commentary. Times without this annotation refer to events in the run.

Single-segment hard mode 2:24 by Kim Siafa.

Author's comments:

This a hard mode any% run completed in one go, without loading and without dying. The route I used in this run isn't the same as the one I used in the normal mode SS any% run. I defeat Amorbis and I grab dark suit in this run. It is possible to complete a hard SS any% run without the dark suit but you need too many energy tanks to survive against the bosses, especially the boost guardian and Quadraxis. 4 energy tanks are on your way before the boost guardian battle, you can grab these e-tanks without losing too much time. The problem is that 4 isn't enough to defeat the boost guardian in hard mode without dark suit. The "real" maximum you can get before fighting the boost guardian is 8 but you would have to grab e-tanks which are very far away from your way : one in temple grounds near windchamber, one in underwater Torvus near catacombs, one in Agon wastes near the temple and the last one in Chykka's room. Getting these 4 e-tanks is very long and they make you do backtracks. You can grab 6 or 7 e-tanks before fighting Quadraxis without really losing time, it's obviously not enough to defeat Quadraxis in hard mode w/o dark suit or at least in a SS run. It's possible to get 10 before fighting him but it would still be difficult to beat time. Between doing countless and endless backtracks to get additionnal energy tanks and grab the dark suit I did my choice : grab the dark suit.

The biggest problem with a SS any% hard mode is the last boss, the emperor ing. DS3/DS4 is a piece of cake in hard mode. The emperor ing was killing me about 85% of the time and it was very frusterating because it is at the end of the run, after playing during 3 hours. He killed about 5 SS runs. I took the first run I killed him succesfully in but it is full of mistakes. The biggest mess up was in transit station where I have to get the early power bomb expansion, I think that I lost 1min 40. There are often other mistakes in the run costing between 5 and 15 seconds. I had luck in this run with the power-ups and the invisible pirates, they didn't appear too often. Though, I didn't have luck with seekers door. I think that I was also lucky to beat the emperor ing and to complete a hard any% in one-go without dying.

100% 2:25 by Paul Evans, done in 24 segments.

Author's comments:

I started this run on December 17, a mere one month after this game came out. At that point, my goal was a time of under three hours, and I doubted whether that was even realistic. I wanted to do my best either way, so i spent a ton of time, probably more than actually working on the segments, going through each part and looking for any tricks that could aid my run. So, not to sound cocky or anything, but no matter how much you think you know about this game, if you want to see some new tricks you should watch this. Those who thought the discoveries have stopped long ago couldn't be more wrong.

As for the actual performance of the run, I would say it is about the same level as my 100% in the first Prime. I'd estimate about three to four weeks were spent actually playing the segments. However, since this game is nearly twice as long, thats less time for each segment. But for the most part, the segments were relatively easy, with some hard tricks usually thrown in. Even if you don't care about the new tricks, and just want to see a fast speed run, this should still entertain.

As for thanks, i'd like to thank the IRC guys and gal for support, and tips here and there. Also, all those who didn't give up on this game, and continued fighting the odds and finding ways to go faster than what was presently known. Finally, nate, who, despite being overrun with work, found the time to capture this run, and Radix, who runs this lovely site which got me interested in speed runs in the first place.

Hard-mode 22% 3:37 by 'kip', done in 30 segments.

Author's comments:

Yes, this is another "who cares" run, a lot like my other hard 22% run in the original Metroid Prime. I put minimal effort into optimizing it and used many more saves than I normally would (some that were very out of the way). I didn't do this out of laziness. I just feel that since the game has just been released, and things are being found every day, I don't think it would be a good idea to put everything I have into a run only for it to become outdated before I'm halfway through it. That's what happened to this run; over maybe 10 minutes worth of time savers were discovered by the time I was done. I'm not annoyed by that, because I knew it would happen, and I was half expecting an item to be skipped before I was done, which would have meant restarting.

As far as I know, 22% is the lowest percentage in the game right now. It's the default percentage, like the original MP's 29% (collecting no expansions but getting all the main items). If a lower percent were possible, this run would have reflected whatever that was instead of being 22%. Here's a list of what counts for the 22%, going by the game's item order, as some things can already be gotten out of order:

missile launcher
violet translator
morph ball bomb
amber translator
space jump boots
dark beam
light beam
dark suit
super missile
emerald translator
boost ball
seeker launcher
gravity boost
grapple beam
dark visor
cobalt translator
spider ball
power bomb
echo visor
screw attack
annihilator beam
light suit

There's also the energy transfer module and 18 temple keys, but those don't add to the percentage for some reason. It's possible to skip the dark suit, but that requires several energy tanks, as there is currently no way to survive against Boost Guardian and Quadraxis with no tanks and an unavoidable life drain of 6 energy per second (if reaching Torvus is even possible in the first place). It's not about just never getting hit, because Dark Aether's life drain will gladly kill you; it doesn't stop at 1 energy or anything like that. So, skipping the dark suit means ending up with a higher percentage, much like skipping the varia suit in the original MP.

If the percent ever drops, I'll probably try to record another run on hard unless someone else wants to instead (but even in that case, I might anyway just for proof). I felt recording this would be better than not just because it doesn't look like 1:04 or something. That said, I'll definitely try harder if/when I do another run.

I died once against Dark Samus at the end, because I forgot about this one attack she has and didn't know how to dodge it yet. The game lets you start back at the escape sequence if you choose to continue from your last save, and it doesn't even count against the final time... so I guess expect to see people killing themselves here until they get a good enough fight.

If anyone's only interested in a certain part, here are most of the bosses: Amorbis (part 5), Boost Guardian (part 8), Chykka (part 13), Dark Samus 2 (part 20), Quadraxis (part 23).

Return to the Game List, the FAQ, or the Home Page.