SSX3 - Metro City (NMG) -

First run back from a long smoke break, a lack of recent "warm up" runs made this the most stressful thing I’ve been through this year. I don’t even have to edit the video at all. I’ll come back for actual analysis later but right now it’s just shaking hands and happy tears.

Edit:
Credit to my buddy Micah for replacing my joystick twice with his mad soldering skills.

And to my little brothers who listened to me talk about small movement optimizations and frame counting (that didn’t even find anything) off and on for months.

And to the nice people on the SSX discord, who have answered my questions (even dumb or off the mark ones) in good faith and been very welcoming. Specifically Rithek, who helped me with obs and getting my control overlay going. Despite both things being embarrassingly simple.

Finally notes on the run. Keles glide off the top hill, while super cool, was .033 slower than Ota with no glide. I’ll stop here to note that Otakung is a fucking wizard. Beat Keles glide by 1, literally anything I do w/o AI by 3, and only loses to a good AI boost by 5. Fucking. Wizard. Anyways. While the glide locks in a good time, it’s not necessary. Lost nine frames on Kele in two & three combined from skipping the next glide. I’ve been going 50-50 on the glide if i get the AI boost, gliding saves a good bit more time, but the AI boost saves just enough time to put in the bank alone. This was a "put it in the bank" run. Funny enough, it’s not 2 that is most affected by the glide, it’s 3. Which surprised me a few months back, it’s both faster and much easier if you hit the booth in a glide. 4 was embarrassing and nearly killed the run. While it may look like it, I was never particularly close to crashing. But it bled off enough time as is and could’ve easily bleed more. 5 was unremarkable if a tad slow. 6 took advantage of recent movement optimization. Seventh verse should’ve been same as the sixth, but I veered just a hair too far left to catch the rail clean and ended up a tad slow. Finished off the run as good as the normal strategy can do. I would’ve liked to go for a glide there, but there was no way my nervous/sweaty/shaking hands were pulling that off. Getting the tip around is both easy to nail and to fuck up. Honestly got kinda fortunate at the end, I knew I went off the last rail just after 1.17 (1.17.0667 to be specific) which means I didn’t have much time to spare (exactly .0667 seconds post rail). Normally that would make it glide or bust, because its hard to tell specific frame count from the corner of your eye. But I just had to leave it to the gods. And they smiled down at me.

Final time of 1:25.9. Some pretty obvious places to clean up in 4/5/7 and no more coin flip on the glide in 2. "Best of" pre/post water tower comes out to 1.25.266 while "best of" all 8 is a 1.24.9 (Recently dropped another .233 (dropped further, now faster than the TAS in 1 and 3-7, tied in 2 and .033 off in 8, I think I have a problem)). A little bummed that one of the AI boost + glide attempts didn't get the skip, but that's the cruelty of Metro NMG. 1.25.5 is the eventual goal and is possible rta without an AI boost.

Edit of edit 2: I have not burned out on the category but progress is becoming tricksy. 1.25.5 is not just possible, it’s inevitable. It’s the safe goal. Sub 1.25.00 needs perfect execution and perfect luck on the AI boost and at the water tower. I worked out luck o. Booth 1.


The glide at the bottom of the hill to start 4 both adds speed and is a pain in the ass in multiple ways. Which unfortunately is the easy to implement and relatively free one. I’m still torn on which ones to implement next. the glide up the hill to booth two is both tempting, RTA possible, and possibly the most punishing thing in the run. 


Minor tweaks in the first three section (one in 1 is missing due to the AI boost). Mainly some frame cancelling and better routing (1-2 ish frames earlier on a jump, tilt your board a little better, etc). 4 saw the biggest change. My 1:26 was built upon making 4 easier and more consistent. This 1:25 was built upon me screaming “fuck it”, ripping apart all semblance of consistency, and grinding the TAS route until it was burned into the back of my brain. I didn’t even do it well in this run and it still saved a chunk of time. Very minor changes made five more consistent. Of all the sections I’m probably most proud of 6’s adjustments, turned from a consistent trouble spot to an asset. Seven had some adjustments that only come in to play if you nail the booth cleanly, I did not. 8 was more minor tweaks and some adjustments on the moguls. On the whole added 6 of 8 available frames from the AI boost, and about 4 of 10 available frames from 4’s route change. And the rest was movement optimization.

AI boosting may not be this simple for other tracks, but for Metro just hold L and lean into the AI. I have paragraphs of other notes on it but that’s basically it.