This run aims for the fastest completion time in 1p very hard mode. I used the characther yoshi, instead of the common choice , fox. I Tas'ed two test runs before this one and i got to the conclusion that yoshi beats the game in a faster time than fox.
stage 1 : link - Not much seen here i used DJC to get him quickly to the edge and finished with FAIR. Couldn't use less DJC's and i had to manipulate link's DI in the right direction or no direction at all.
stage 2 : yoshi - painfull one, but i got an excelent run, the yoshis fell in favorable position's for the DTILTS which is the only technique that is able to 1 hit Ko them without a star finish on very hard mode.
stage 3 : Fox - re did this from the beggining about 7 times and finally managed to manipulate him to do a left roll which saved me from having to use a grab and saved a few frames. As for the DJCs, beautifull combo and got perfectly timed with the spike in the end
stage 4 : Mario bros. - Maybe the best stage in the whole run, got both the brothers in the same position allowing me to combo them exactly at the same time and saved a lot of frames.
stage 5: Pikachu - had a problem with the start. pikachu would come to the left side of the stage no matter what i did, i had to wait for him to do a dash attack so i could start the DJC. Just after the first NAIR pikachu did a roll to the right which helped a lot and saved 1 DJC, the last FAIR had to be delayed 1 frame in order for pikachu to die quicker saving another 7 frames.
stage 6: DK - Much faster than the usual fox strategy. Beating DK with 2 eggs was very quick, fast falling and delaying the last B allowed DK to die while still inside the egg and that was enough to get the time counter from 4:55 to 4:56.
stage 7 :kirby team - Another perfect stage. Killing the first kirby with a spike made the "mario kirby" come to the left platform and allowing me to manipulate the other kirbys in the same position.
stage 8 : samus - the strategy used was the one i found to be fastest but no matter what i changed in the whole run i had to waste 1 frame before starting running or else she would shoot the energy ball, having to jump early to avoid it, parrying it or anything was slower than wasting 1 frame.
stage 9 : metal mario - same as DK. metal mario dies faster because of his super weight.
stage 10 :polygon team - painfull one, manipulation all the polygons in the same position is very hard. Managed to manipulate a snorlax from a pokeball to kill 4 polygons before they could reach the floor.
final stage - Master hand went just as planned without modifications from my test runs. Just used a diferent manipulation using yoshi's second jump to hit the floor instead of falling normally, both takes the same ammount of frames.
bonus stages: the bonus stages are something that made yoshi an "unfavorable characther" but the improvements in the run make up for the time lost here.
mmbossman: In general, fighting games usually do not make good TASes when speed is the primary goal. The problem usually stems from having to use one or two strategies repeated over and over in order to achieve the fastest time. While this may impress people who are knowledgeable about the game, it tends to be monotonous for the average viewer, which I think the voting for this movie, along with the resulting discussion, has demonstrated.
Most of the fighting games with runs currently published focus on showing off, while still aiming to be moderately fast. I believe that this game also fits that mold; it is preferable to show off much of what the game has to offer while sacrificing some time for the benefit of more entertainment. While I acknowledge the fact that this will be an unpopular decision with some people, I believe that this game works better as a TAS when it focuses on entertainment. Therefore, I am rejecting this submission in favor of this submission.
Nahoc: Added YT module.
Nahoc: Judging...
Nahoc: Rejecting this movie since it is clearly un-optimal by today's standards. An already faster TAS by DennisBalow using Kirby has been posted on YouTube in 2009. A new and faster submission using (most likely Kirby) would probably be accepted for the vault tier. Until then, it is meaningless to publish this as it fails to beat known records.