Mr.Toad's Fifteen puzzle with timer (modified by stannic)

[stannic]

It would save some time to either store the result of getColors (updating when the puzzle size or color scheme is changed), or to just not compute the whole array each time, e.g.

Yes, you're right. Changed the code to not compute the array. Also, changed a bit behavior of scrambling function (in v.5.2, scrambling 16x16 puzzle with Live XXX scheme was very slow). Updated online link.

I'll see if there are other ways to optimize speed. Also, the code is rather long and could be shortened (e.g., right now I have three separate functions for Ao5, Ao12 and Ao100).

Edit: should work more or less fast now. Also, uploaded 5.2 upd 2 offline.

- stannic

 

[stannic]

16x16 pb: 34:07.285, 10393 moves, mean tps 5.076.

 

[qqwref]

Very nice! I thought I should beat that, so I gave it a try and got... 28:16.555 in 9940 moves

And no, I don't do this weird row/column shift thing every time, I just felt like doing it on this solve. In fact, I'm thinking I should have started the columns later.


EDIT: I wonder... how does the movecount on an NxN puzzle grow? N^2? N^3? N^4? Something else entirely?

 

[stannic]

Very nice! I thought I should beat that, so I gave it a try and got... 28:16.555 in 9940 moves

And no, I don't do this weird row/column shift thing every time, I just felt like doing it on this solve. In fact, I'm thinking I should have started the columns later.

Well, I've did not reduction to (N-1) x (N-1) too. I've just used Live Rows mode.

This is to show how it is unfair to use it on competitions...

Edit:

EDIT: I wonder... how does the movecount on an NxN puzzle grow? N^2? N^3? N^4? Something else entirely?

Probably as N[SUP]3[/SUP]. Basically, you need O(N) moves to solve one tile, and there are O(N[SUP]2[/SUP]) tiles.

I've looked at my PBs (Mo10). They were not established yet, though. Also, it's not FMC but just the best movecounts from speedsolves.

2x2: 4.0 = 2[SUP]3[/SUP] * 0.5
3x3: 42.5 = 3[SUP]3[/SUP] * 1.574
4x4: 100.7 = 4[SUP]3[/SUP] * 1.573
5x5: 219.8 = 5[SUP]3[/SUP] * 1.758
6x6: 438.2 = 6[SUP]3[/SUP] * 2.029
7x7: 724.8 = 7[SUP]3[/SUP] * 2.113
8x8: 1280.0 = 8[SUP]3[/SUP] * 2.5

16x16 (the best "fair" single): 12932 = 16[SUP]3[/SUP] * 3.157

Assuming that after a certain amount of practice the TPS becomes constant independent of the size of puzzle, times would be also O(N[SUP]3[/SUP]).

Edit 2: My movecounts are not a good data source since I'm newbie. Someone else probably could provide better data. Maybe oranjules?

- stannic

 

[oranjules]

For 4x4 and 3x3, that is what I get in average. The 2x2 isn't representative enough I think, and for bigger ones, my method (row by row) isn't move optimal. But the O(N^3) seems a good approximation.

 

[Owen]

So frustrating, I'm trying to do 2-7 relays, but my 7x7 time is bigger that 2-6 combined...

 

[Jaycee]

Second try on the 16 x 16. 58:03.004 Sub-hour! Around 11.5k moves at around 3.1 TPS

 

[stannic]

Got a lucky 7x7 single: time 1:43.507, moves 606 = 7[SUP]3[/SUP] * 1.767.

Ao5 2:16.160 (774.333 moves), Ao12 2:23.732 (792.7 moves). Rainbow coloring (Multi Rows) used this time, so I believe the results are fair

Edit: Oh wow, 4x4 mo10 16.759. 109.4 moves.
best single 12.357 (83 moves), mean tps 6.528

Edit 2: 4x4 single 11.682, 67 moves, tps 5.570.

 

[qqwref]

The "Multi Fringe" color scheme seems to help a lot, haha. I am tempted to do the 16x16 again.

 

[uvafan]

I'm back into these things...

4x4 pb 13.385 avg12 19.248

 

[stannic]

I'm back into these things...

4x4 pb 13.385 avg12 19.248

Wow, moar speedslidin.

16x16 time 36:41.071 moves 9823 tps 4.463 color scheme Multi Fringe

 

[oranjules]

First try on 16x16 : 30:31.261, 10957 moves, 5.983 tps.
With row by row method, the 2 last ones were a hell

 

[qqwref]

Too much competition Okay, I kinda like it.

 

[vcuber13]

46:24.121
Moves: 11102
TPS: 3.99


edit:
38:39.515 11479
Mean TPS: 4.949

finished in 28 min but had 2*2 numbers switched which took 10 min to find and solve

 

[oranjules]

2-7 relay : sub-3 ! 2:46.812 PB, and i almost always do 2:5x.xx.
I optimized my last line solve (i do it with 3-cycles, so it's a bit long, especially for 16x16)

 

[qqwref]

38:39.515 11479
Mean TPS: 4.949

finished in 28 min but had 2*2 numbers switched which took 10 min to find and solve

That hurts ._.

 

[vcuber13]

2-7 relay : sub-3 ! 2:46.812 PB, and i almost always do 2:5x.xx.
I optimized my last line solve (i do it with 3-cycles, so it's a bit long, especially for 16x16)

whats your breakdown?

 

[oranjules]

something like : 1:20 , 50 , 20 , 8 , 3 ,0.xx

 

[qqwref]

16x16 in 24:59.699 (LOLOLOLOL), 9329 moves, 6.221 tps. I used Multi Fringe and did one row, one column, etc. the whole time.

 

[uvafan]

2-7 relay 5:46.915 1547 moves

times:
2:26.750, 1:46.407, 57.576, 27.215, 6.662, 2.305 = 5:46.915

moves:
665, 454, 266, 126, 27, 9 = 1547