Slashdot Mirror
Metroid II, Prime Get New Speed Run Records
PrinceBrightstar writes "Both the Metroid Prime pure speed run and the Metroid II 100% completion speed record have been shattered by Zoidi and Brightstar (myself) respectively. The Metroid Prime (GameCube) pure speed record is now 1hr 17mins, and the Metroid II (GameBoy) 100% is now set at 1hr 12mins, with further decreases planned - these records were recorded into video form and no emulators were used." We've previously covered Metroid Prime 'speed runs', which are an extension of classic Quake speed runs.
It seems that the run was broken into segments which were based on save points.
Did the person actually run through the game in the time he said he did? Or did he just repeat each segment for the best time possible and then add all the shortest segments up for the time stated?
-J
100% is just getting every item... pretty simple :-) Since the weapons respawn and are in different places, we just decided that you have to get each one once.
As for a complete map, check GameFAQs, I just submitted once recently.
Speed Demos Archive - Lots of speed runs!
Recorded with emulators using thousands of save states and slow motion too.
Everything at SDA is on the real game.
Speed Demos Archive - Lots of speed runs!
Considering I personally have beaten Super Metroid in less than 1 hour with roughly 60% completion, its not hard to imagine some hardcore gamer out there getting 100%, with an extra half hour. (Unless you do a bare minimum run through, you will get at LEAST 30% completion rate. As it stands the lowest is 15% so 30% isn't that much more.)
Not only that, there are reports and entire FAQs dedicated to bare MINIMUM runs (theres even a report for a 1% run through in Metroid Fusion!) Naturally this were made on the emulator so they wouldn't count, but the fact that they CAN BE DONE remains.
The one by the japanese player who did the 11 minute run admitted that they were faked by going frame by frame through an emulator.
Why would a video recording be more reliable as evidence? Really, what would prevent a cheater from recording on an emulator and then creating a video from that (through his TV or whatever)?
Your chances of catching a cheater are in fact greater when dealing with emulator input recordings, since you can analyze the exact input and, judging from reaction times and various oddities, to some degree determine the likelihood of slowmotion being used. Then there's visual analysis, of course. A skilled player can often tell when there's something superhuman going on. I know that a few cheaters on the Doom speedrunning scene have been caught this way.
In the end, the safest way is knowing the person. And trust me, there's no reason the guy who set these records would have cheated. He's made tons of recordings, each containing mistakes and occasionally improving his time slightly. I doubt a cheater would go through that.
On the other hand, if I saw a flawless run...