• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

[Unofficial] Sq-1 19.61 AO5 (parity control)

Joined
Apr 23, 2010
Messages
1,391
Location
Scotland, UK
WCA
2009SHEE01
YouTube
Visit Channel

Edit:
18.56, 19.82, 20.44, (20.52), (17.84) = 19.61 AO5

Wanted a decent AO5 on video with successful parity control, didn't take long to roll this out. Also wanted to test recording with my phone. Quality is good, the angle isn't great so I'll fix that in future.

Not practiced much so untimed inspection for now, mostly that goes ok, sometimes it doesn't like the 4th solve here where I took too long to remember how I handled the case (but it was the slowest solve anyway).

Wish someone else would decide to learn this stuff, it's so damn cool.

Not the best video I've ever made, but it's something. Felt like trying to make this eventually today, the original plan was to pull this off in comp a few months ago, but I failed in comp with 7/10 success rate.

Sq-1 hax is awesome.
 
Last edited by a moderator:

Lucas Wesche

Member
Joined
Mar 25, 2013
Messages
289
WCA
2012WESC01
This is so awesome. I'm going to learn this now, but I have a last question.
Do you think it would be faster for easy cubeshapes like fist/fist to just do the normal cubeshape and then solve parity with CP parity algs?
The CP parity algs are pretty fast and I think doing a longer cubeshape alg for a case which requires 2 slice moves isn't profitable enough.
 
Joined
Apr 23, 2010
Messages
1,391
Location
Scotland, UK
WCA
2009SHEE01
YouTube
Visit Channel
Some interest, cool. I've got a couple of minor tweaks to make to the info I uploaded, but it's not much of an issue if anyone wanted to have a look. Questions are welcome!

This is cool. I kind of want to learn this even though I suck at square 1.
I didn't really solve sq-1 when I started this. I think my PB was around several minutes with very nooby techniques, I only started solving it so I could do this stuff.

This is so awesome. I'm going to learn this now, but I have a last question.
Do you think it would be faster for easy cubeshapes like fist/fist to just do the normal cubeshape and then solve parity with CP parity algs?
The CP parity algs are pretty fast and I think doing a longer cubeshape alg for a case which requires 2 slice moves isn't profitable enough.
I guess? Fist/fist is 2 moves with good parity, 5 without, so that's 3 extra moves. Parity CP algs are 7 moves minimum, but I don't know how the average movecount compares to non-parity CP, could be less than 3.

Would it be better to learn a parity alg which is a lot shorter than the others that doesn't preserve anything minus CS?
62/90 cases can be done with the same number of moves regardless of parity, a further 18 take only one extra twist (though 14 of those are annoying IMO). There's 2 cases which take 2 extra moves for bad parity, but they're both nice. The rest take more moves because they're too close to cubeshape, like fist/fist - Lucas just suggested parity CP instead for these cases and he's faster than I am, so maybe worth considering.

So overall, for most cases it's beneficial to take care of parity directly rather than doing a short parity alg once you get to cubeshape.

Best short parity alg only preserving CS: /-3,0/4,-1/2,-2/-2,2/-4,1/3,0/

Edit: I worked out the number of extra moves needed in R, I wrote some code in my free time to map out all the states as a graph (I'm a PhD student in network theory, which is similar to graph theory). Now that I have that I managed to improve a couple of cases, still got a couple left to do though before I update the document here. If anyone has specific requests regarding this, I'm happy to help.
 
Last edited:

not_kevin

Member
Joined
Jul 23, 2008
Messages
495
Location
California, USA
WCA
2008YOUN02
YouTube
Visit Channel
As a person who used to use CP+P until switching back to full Vandenbergh, I think that Lucas' idea of saving parity for CP+P is valuable here. I switched to Vandenbergh because I found that the inspection was too costly for me - although I was able to finally get sub-16 avg100, I would have to spend a second or two to recognize the PBL, and then decide what to do - even though I had already figured out the CP while doing EO. Now, I'm sub-15 avg100 because half the time, I don't have the extra recognition time (it's a parity-less solve), and the other half, I have enough EP influence to not significantly slow down my average.

However, if you can go into your solve knowing if it's going to be parity, then you have no extra transition time - although it's true that you can "save" 4+ slices by doing the 5-move parity fist-fist case, doing the 2-move fist-fist case allows you to predict your OBL much easier, and therefore have a smoother transition into the rest of your solve, and now that you _know_ you have parity, you don't have to bother with the CP+P recognition step - just ABF as necessary to go straight into your CP+P alg from the CP you saw during EO, as usual.

But yeah, for sure CS+P is on my list of things to learn once Nats is over (the paradigm shift is a bit too much for me at the moment :p). Really cool stuff!
 
Joined
Apr 23, 2010
Messages
1,391
Location
Scotland, UK
WCA
2009SHEE01
YouTube
Visit Channel
You're probably right. I solve differently after cubeshape for the fun of it, so I've no interest in learning CP+P. People using normal methods who are mostly interested in being fast can do the recognition for CS+P and use that information in whichever way they think best, and CP+P now sounds to me to be the best option for 'bad' CS+P cases.
 
Top