I now have a PhD in floating boards after this quest
Despite the thumbnail, I did enjoy this record. I just wanted to recreate my favourite board racing poster-child. However, I went through quite an ordeal on Monday which inspired my cinematographic decisions today.
Every single action and jump here is made with purpose and has been concluded as the best action after several dozen iterations of runs that I compared frame-by-frame to see what was gaining me speed. Having to adjust from the generous hitboxes of the score symbols in field races to the incredibly small point cubes here was a big task in routing. I probably put in over 3000 retries here, and I can't count how many times I kept recording times under 1:57.100, never breaking into 1:56, even recording a perfectly even 1:57.000 just to tilt me.
I've only managed runs under 1:57 twice. This is the second one, but the first one has a sad story. The last clip in the intro is an example of me finishing during a lag spike, which added almost 0.6 seconds to my run (notice how far I make it into the goal before the trial actually ends versus every other try); that run being the first ever legitimate 1:56 I pulled off, which was ruined due to latency. It's possible that run is still faster than this one, but it's dead as far as the game is concerned. I nearly quit this record on the spot after that bullshit, but luckily I still had a day left to recuperate.
This is a board race with no random elements, except connection quality, which can influence my time due to lag upon touching the goal. This on average raises my clear time by .2 seconds as opposed to the true clear time measured via counting frames. So really, this run should be about a 1:56.7XX, but that's just the penalty I take for playing on Global.
If this game had better lag compensation, then this race would be perfect, as it features no wandering enemies, other players, or jagged terrain with random rocks and divots in the ground to jam you up.
I also had to realise that spamming jump, as everyone and I have normally done, is not always the best option, and is just a way to hasten the onset of carpal tunnel. On a flat surface, jumping does basically nothing. On a hill, jumping will only help if you boost while in the air, as a boost immediately sets your lateral velocity to its maximum. This is important because as soon as you touch a hill, your lateral velocity gets reduced proportional to the angle of the hill, up or down. Therefore you want to be on flat ground in a boosted state, or jumping just before every boost ring on every hill, or boosting manually after jumping if a ring's not available.
This whole race forced me to re-discover mechanics about the floating board and correct the misconceptions I had created over the last year of racing in the open world environments. Hopefully I can go back to them with this knowledge and improve my times.
As for the first two seconds, the fastest way to reach the floating board that I found was to precast Triumphant Shift and dash forward with it, then use Assault Buster, then use Spinning Gust and step out of it nearly instantly, so that I touch the board terminal as my step ends. As my step ends, I hold down E so that I get on a board as soon as possible, and boost as soon as possible. I put an unusable skill on my middle mouse button so that I don't accidentally use it when I try to immediately boost after getting on the board. This one combo of moves made a hell of a difference to simply just using Spinning Gust spam, nearly half a second's worth of time gain.
I'm honestly not sure how much lower the time could get here. I feel like I was approaching total optimisation, but there may be a thing I missed about the floating board physics that means this route isn't the best. But after 3000+ retries, I was mentally done after securing my 1:56 goal.
0:00 Intro
0:50 The run
Информация по комментариям в разработке