May 2015 Time Trial - Marvelynxous

[Ihavenoname248]

(like did Ihavenoname248 miss the quicker way to get the fireball to the SE chip because it was only possible in the latest update?).

Probably- I made a different route for each of the first few versions (and this level is definitely non-trivial to route!) and so when the final version came out, I tested the route I'd made for the version with the broken teleports, it worked, it was somewhat fast, and I was drained from working on the level already. Not going back was "ugh, do I really want to make a route for this again?", especially when I'd already met my goal of scoring a faster time than Time Trial Mixed I.

http://pastebin.com/xaQc4JC0

Here's some thoughts I wrote up after optimizing.

[M11k4]

Probably- I made a different route for each of the first few versions (and this level is definitely non-trivial to route!) and so when the final version came out, I tested the route I'd made for the version with the broken teleports, it worked, it was somewhat fast, and I was drained from working on the level already. Not going back was "ugh, do I really want to make a route for this again?", especially when I'd already met my goal of scoring a faster time than Time Trial Mixed I.



http://pastebin.com/xaQc4JC0


Here's some thoughts I wrote up after optimizing.

Bathing in Battery Acid (MS): 282. Lots of boosting shortcuts and a few backwards overrides. Adjusting my Lynx route only gave 281.




Rivaldi: 95 (!)


Not sure if it was intended, but I found one singular path through the teleports without stepping on the same tile twice, and it was barely enough for 95. To aid in routing, I made a flowchart of each chip and which chips it could take, numbered in reading order. That was enough to note a few paths that had to be in a theoretical 95 route, and connect the dots from there to actually find it. My initial attempt was a 94.


Bathing in Battery Acid: 277


My first effort gave 276. Then I completely reordered things to cut back on backtracking, and it yielded a 1 second improvement.


Age of Ultron: 609


On the first version of the level, I had a busted 689 and was positive I had the best possible route for all 3 levels (watch as BiBA has a 278 route). On the second version, which was still busted with timing tricks, I had a route that would come out to about 645, but couldn't execute it in the hour before I went to sleep. In the morning, the level had been updated again and I had classwork to do for the forseeable future. The level saw another update to prevent further shenanigans, and on the 6th I rerouted the level...and forgot I needed a fireball in trap 3 (the lower of the two in the 3x3 loop) to get the paramecia to chip 5 (numbered counterclockwise from the top).


After rerouting again to account for this, I ran into the issue on the second to last version of the leftmost trap having monsters return to it. This was fixed quickly as it wasn't intended, and I could finally run through my 6th route for the level...and it worked. I scored 608 with a 2 move mistake, and a 2 move suboptimal (1.5?) routing choice. At least with this route, 609 is optimal, and though I feel solidly about this route, there may be better out there.


I'm too drained on routing this level to look further, though.


For reference, here were the notes I used:


BFP (EAST), unload all


Release fire x2, release para


Release fire, release para, release fire x3, take chip 7


Release bug, move fire past para


Release para, fire, bug, para, fire x2, para


Release para, take 2 chips, fire x2, chip, bug


Unload all 3. (PFB)


Release para & bug, fire, para


Release fire x2, release para, take chip 2.


unload both, release fire, release para x2


Take last 2 chips, finish.


As a final note, this route is 1 second faster than J.B.'s route for "Time Trial Mixed I" from a couple months ago.

I had several versions of Rivaldi, and I chose this one because of the single quickest route picking up the chips. I then placed the exit far enough that it would require this route for the bold time. It ended up taking me longer to design the part after the socket (like over two hours) than before it, because I wanted it to seem natural and still only work in Lynx.