let there just be new thread #52

[eggyaytso]

my new time spreadsheet https://docs.google.com/spreadsheets/d/18bPh1NatNamQAHRtp05zXBVkjjo5_yTtPyPZ8uR7mYg/edit#gid=548872536 fmc singles https://docs.google.com/spreadsheets/d/1miFV_yV1f1Ik1VZY-Aa-HDfwT0fuRSLON_5eAh34eoo/edit#gid=2074135236

[eggyaytso]

i see 42 here?

[jrj5423]

yes, an alt accout of 42 which was registered in 2020

[eggyaytso]

is there still 2 jze on lb?

[jrj5423]

yes sum of time - ranking 38th just ignore it

[jrj5423]

jze2 has no big puzzle scores

[eggyaytso]

that's not good for people <38th place

[jrj5423]

yes, i absolutely agree with you

[jrj5423]

where can i find slidysim fmc lb

[eggyaytso]

in slidysim app or for wrs: http://slidysim2.000webhostapp.com/leaderboard/records_moves.html

[eggyaytso]

@jrj5423 also my spreadsheet for 6-10 singles https://docs.google.com/spreadsheets/d/1miFV_yV1f1Ik1VZY-Aa-HDfwT0fuRSLON_5eAh34eoo/edit#gid=2074135236

and here is my rankings for slidysim for some categories https://docs.google.com/spreadsheets/d/18bPh1NatNamQAHRtp05zXBVkjjo5_yTtPyPZ8uR7mYg/edit#gid=1445879275

[eggyaytso]

now i'm in combined top 10 for 5x5 single:

1) jrj 74
1) hamburger 74
3) ZYCI 75 (??? who is he)
4) cslimit 76
5) vovker 81 (obviously cheated)
6) jze 82
7) ben 83
8) daanbe 84
9) 42... 86
10) dphdmn 87

[jrj5423]

ok thanks

[eggyaytso]

i should get 85 to make it 81 82 83 84 85

[jrj5423]

my 74 (2.658)

0 3 7 8 10
1 11 6 17 19
22 2 18 24 9
5 21 16 4 20
23 14 15 13 12
U L L U R R U L L U R R D L D R D L U L L U R D D R D L U L D R R U U L D D R U L L U L D D R U U U U L D D R U U R D L U R R D L D R U L D L U L U

hamburger 74 (3.706)

2 9 4 10 5
1 12 7 18 23
6 3 17 8 0
11 16 24 19 15
21 20 13 14 22
R3 D D R U U U L D D L U L D D R U U R U L L D L D R U R D L U U U L D R U R D L U R R D L D L U U R R D L D L L U U R D D R U L U R D D L L U U

[eggyaytso]

@jrj5423 can you explain your solving method for big puzzles? what do you use that much time for? i can't even think what else i can see, and if i spend more time, i just lose flow of the other elements, idk if i can explain that, maybe you are doing something specific? do you measuring every move or is it still more like a strategy, planing sequneces etc? do you have a guide somewhere in the app to read? (maybe in chinise)?

[eggyaytso]

basically, what i'm trying to do, is mostly just solving two fringe layers, and while doing that trying to do moves that affects other elements the better way, sometimes solve obvious cases, but i just can't see what will be with them after i solve first fringe layers, it's too chaotic for me, on 8x8-10x10 is extremely hard i just have nothing to do but just trying for lucky cases and slowly learning the places of different numbers and color ordering, for example, solving 6x6 part in 9x9 is much harder because of the weird numbers and less-gradient colors..

[jrj5423]

I was trying to evaluate the current and potential movement in each direction, trying to "greedily" use fewer steps that would increase the manhattan distance. For example, I would try to move as many pieces which have upper target positions as possible up when the space moves down, and move as many pieces which have lefter target positions as possible left when the space moves right, and so on. Because of this Manhattan minimization idea, I will give priority to solving the entire outer circle instead of just layer.

[jrj5423]

from what I know, cslimit's skill is similar to mine

[eggyaytso]

that's sounds obvious

[eggyaytso]

can you tell more

[eggyaytso]

it's extremely impressive when you just finish the later and already have almost-solved everything else

[jrj5423]

for example, the solving priorities of 15, 42, 49 and 85 (in 10x10) should be relatively close

[eggyaytso]

solving from right and bottom is very hard

[jrj5423]

Considering that the space ends up going back to the bottom right, maybe 49 and 85 are a little less important than 15 and 42 -- but only a little.

[eggyaytso]

major problem is that i don't know the exact wanted-position of each of 99 elements all the time, so i can't intuitively find the good pieces on-fly

[jrj5423]

yes, it is different from the method of we are more familiar with in speedsolving, so I'll spend more time thinking of the target position of every pieces and trying to make my decision as rationally as possible

[eggyaytso]

but that's insane even if you calcualte like 30-40 pieces every move, you can't controll the difference of these position after few moves, or maybe just a little, of only few pieces you are currently moving.. i just don't understand what to think no matter how much time i have

[jrj5423]

I made a color scheme last year that directly reflected this idea. But it doesn't work as well as imagined, maybe because it's too extreme, or maybe because I'm not enough of it

[jrj5423]

if you're willing to try it, it's probably not as hard as you might think now

[eggyaytso]

that's a nice idea but too chaotic

[jrj5423]

indeed

[eggyaytso]

@jrj5423

[eggyaytso]

i should actually try it probably

[eggyaytso]

but would be better with diffrent shades for bottom-right

[eggyaytso]

( impossible to make in slidysim)

[eggyaytso]

naaah that's too hard, i should do same in LM but with more colors, well, not this month

[jrj5423]

more than one third pieces are red - another kind of torment maybe some compromising schemes are better

also i think 12 should be more prioritized than 5 - but maybe i shouldn't consider so much when designing a scheme

[eggyaytso]

@jrj5423 maybe you can make it the idea is just do lighter/darker version for top/bottom sides

[eggyaytso]

@Daanbe also what do you think

[jrj5423]

looking forward to your return, I'll take my handcam video as a welcome gift, I promise

[eggyaytso]

doing both fringes at the same time = new meta 🥚

[eggyaytso]

@jrj5423 ❤️

[jrj5423]

@jrj5423 maybe you can make it the idea is just do lighter/darker version for top/bottom sides

Yes, but I probably won't practise fmc for a while because of my busy schoolwork

[jrj5423]

time to sleep bye

[eggyaytso]

cya

[Daanbe]

i think the biggest problem is colors, there's a lot of useful information but if you try to display it all you get overwhelmed, i guess different colors for top left/ bottom right parts works if you already do one color for top and left

ideally you'd have something that helps you solve to every side and still retain which side you're going for you can get over the same color for different sides thing while solving by going slow, but it makes resetting for a good scramble really hard

with the flexibility of LM color schemes i was thinking of something like top left fringe on background and bottom right fringe on text color but i haven't tried yet

the big issue i see with jrj's scheme idea is that you have so many colors that it gets overwhelming and still have the same color for far away pieces

the dream would obviously be if you could switch between color schemes mid solve

[jrj5423]

two immature schemes

[jrj5423]

i prefer the first one - text colors in the second scheme still make it chaotic for me

[eggyaytso]

text colors is a bad idea

[eggyaytso]

first looks great sad that i can't try it rn