Older cubers discussions

[Mike Hughey]

I can't respond intelligently, but I'm convinced a megaminx with magnets would instantly cut half a minute off my solves. I have so many problems with alignment on megaminx that I'm positive it would have a bigger effect than magnets had with big cubes for me - and magnets made a huge difference in my big cube times.

 

[Mike Hughey]

Awwww man... that’s my home state. I’d LOVE to make a comp in Michigan. Of course, they are always hosted by that “other” school in the state. You know, the crappy blue and yellow one... (for our overseas friends, University of Michigan has yellow and blue school colours... I attended Michigan State Univeristy which uses Green and White—CLEARLY there is a rivalry)

I don't see why you wouldn't go ahead anyway. If you hide your colors until you get into the building, I'm sure you'd survive - no one at the cubing competition would physically attack you. (There might be some words exchanged, but certainly no physical attacks, anyway.)

But I admit you might want to try to be discreet as you get in the building...

I still think you should attend, despite it being the rival school.

 

[SpartanSailor]

I don't see why you wouldn't go ahead anyway. If you hide your colors until you get into the building, I'm sure you'd survive - no one at the cubing competition would physically attack you. (There might be some words exchanged, but certainly no physical attacks, anyway.)

But I admit you might want to try to be discreet as you get in the building...

I still think you should attend, despite it being the rival school.

Oh... not a chance! I would go in FULL MSU glory and Colours!! There could be absolutely NO hiding that rivalry! I’d be on the spot at every moment to do right by MSU. My dad is a U of M alumni too... so there’s more than just a simple school rivalry at stake.

 

[pglewis]

For my own unofficial PB list I pretty much stick to WCA regs (including no long inspections, +2 and DNF rules), csTimer scrambles, and Stackmat only. If I'm away from home I'll use timer.cubing.net on my phone but I'd be very reluctant to count any session not on Stackmat. It's not hard to cheat a second+ under my Stackmat times with one-finger timer stops and not having to drop the puzzle. This is in no way a knock on what anyone else does, it's for procedural reasons for myself: what I do for counting unofficial times at home closely matches solves under comp conditions. I have enough other things going on from the neck up to throw me off at comp, the solve procedure isn't one of them . BTW, once after doing a lot of timed solves on my phone I sat down with the Stackmat, did inspection, set the cube down, and touched the LCD with one finger. True story.

That said, I will accept a stackmatted single from a hand-scramble. It's common for me to watch something streaming while individually timing untracked solves on the Stackmat. I can't watch a show and constantly read notation but I can hand-scramble and follow the plot while getting my spamming in (and 3x extra time spent scrambling is acceptable). Singles are all about some luck and a lightning strike anyway so as long as it isn't something ridiculous... we all know if it felt like a legit solve happened. If I decide to track any more beyond a hand-scrambled single then I'll break out csTimer and the rest will be random state scrambles.

I wish I could go to Michigan, but it's looking like I'm just not going to be able to make it to that one. :-( Good luck there - it should be an awesome competition!

I'm also eyeballing Flag City in Oct., it's too early to commit but that's a perfect comp at the right time if I can make it.

I attended Michigan State Univeristy which uses Green and White—CLEARLY there is a rivalry

I have family that went to Michigan St., am an admirer of Tom Izzo's basketball teams, and don't give a hoot about THE Ohio State University so if I were forced at gunpoint to align myself with one of the three it'd be Spartans

 

[Old Tom]

Awwww man... that’s my home state. I’d LOVE to make a comp in Michigan. Of course, they are always hosted by that “other” school in the state. You know, the crappy blue and yellow one... (for our overseas friends, University of Michigan has yellow and blue school colours... I attended Michigan State Univeristy which uses Green and White—CLEARLY there is a rivalry)

Just go incognito, no colors. It's for a blind solve anyway.

 

[openseas]

Giiker cube (brief) mod

Got this Giiker cube thingy, played a little bit & decided to change magnets (original is too strong)

- Original magnets look to be 4x2, probably N35, magnet strength was similar to my N35 family.
- There are slots for magnets and they are not glued. You can hear loud click sound because all magnets are slightly moving around inside the slot when you turn the cube. Slot size fits to 4x2 for corner, 5x1~2 for edges. Edge magnet slots have a little level(?) thing that pressed magnets inside the slot.
- Used 5x1 N38 for edges, 4x1.5 N35 for corners.
- I still feel I can use weaker magnets (I do lots of slice moves)
- I tried to use 3x1.5 but magnets are too small, just too noisy.

I plan to order 4x1 N35 for corners later to see the difference.

 

[AbsoRuud]

It's such an adorable cube with its headphones. But a shame about that awkward colour scheme.

 

[SpartanSailor]

I was sad to see on some reviews that the app doesn’t “track” M-slices, rather makes it two outside turns to match. Which isn’t too bad, but if I wanted to reconstruct a solve, that would need to be taken into account.

 

[openseas]

It's such an adorable cube with its headphones. But a shame about that awkward colour scheme.

Yes, need a different set of sticker. Plan to use GTS1.

@SpartanSailor / even WCA doesn’t recognize slice moves officially

 

[AbsoRuud]

Yes, need a different set of sticker. Plan to use GTS1.

@SpartanSailor / even WCA doesn’t recognize slice moves officially

I understand that you have to be very careful making the holes for the charger and that you have to follow their colour scheme or the cube in the app will look different from the cube in your hands.

 

[SpartanSailor]

Yes, need a different set of sticker. Plan to use GTS1.

@SpartanSailor / even WCA doesn’t recognize slice moves officially

Really? I know being off by an M-slice is a DNF, rather than +2. What about FMC? I’d think they would have to count M moves...

 

[mark49152]

even WCA doesn’t recognize slice moves officially

Not sure what you mean by that, can you clarify? Are you talking about FMC?

 

[Mike Hughey]

What about FMC? I’d think they would have to count M moves...

No, M moves are technically invalid for FMC; they should be DNFed if properly judged.

 

[SpartanSailor]

No, M moves are technically invalid for FMC; they should be DNFed if properly judged.

Interesting... and good to know!

 

[openseas]

Not sure what you mean by that, can you clarify? Are you talking about FMC?

WCA doesn’t accept any slice moves in any events, fmc, n by n etc. just DNF. Not only that, official scrambles don’t include any of slice moves. Only OBTM is accepted and recognized as legal moves.

 

[Mike Hughey]

Practicing 3BLD: I just got 6 DNFs in a row and was feeling pretty discouraged, then got 14 successes in a row! Best avg12 was 1:34.16, which is pretty good for me lately.

I'm still a bit discouraged by how long it has taken me to get used to putting edges in audio (A year ago, I used images for edges and audio for corners; then I switched to audio for edges and images for corners). I'm now to the point where I can generally do pretty well with a solve that has 10 edges. But at 12, it starts getting much slower - seems to add another 10-15 seconds to my memo time, and at 14 it almost always goes really bad, with a minute or more to memorize. I just can't seem to handle constructing an audio loop that long. Once I get the sequence constructed, it's usually fine, but it takes me too long to construct it. I'm not sure if it's just a limitation I have, or if I'm just not doing it right, or if I simply need more practice.

 

[xyzzy]

I can't find it right now, but from memory I think it showed that a sequence of random moves needs to be at least 40 long before the distribution of good/bad edges converges on what you would get with random state scrambles.

Convergence is a process, so technically it starts converging on the very first move, but the convergence is kinda slow. (And just like every statement qualified with "technically", that means "this is true but maybe not in a useful way"…)

Stefan Pochmann showed, back in the day, that with 25-move scrambles (the standard scramble length then), it was about three times as likely to get all 12 good edges as to get all 12 bad edges, even though these should have the same probability with random-state scrambles. The gap between the two probabilities decreases exponentially as the number of moves increases. At 30 moves, the probability of getting 0 bad edges is still 46% higher than expected; at 40 moves it's 4% higher; at 50 moves it's 0.4% higher; et cetera. It never ever reaches 0, although you can get as close as you want by increasing the scramble length. (source)

Is 4% an acceptable deviation? Maybe if you use a method that doesn't involve EO at the start, i.e. pretty much anything other than ZZ and 3OP. However, it's fairly easy to get random-state scrambles these days, so I don't think it's fair to count anything but solves done in "ideal" scrambling conditions.

(Also singles are pretty dumb and IMO they're just for fun, so however seriously you want to treat them is really up to you. Handscrambles all the way? Sure, whatever. Only random-state scrambles with cube cover and stackmat and all that stuff? That's cool too! The above paragraph mostly applies only if you want to measure how fast you really are.)

My question is whether or not the Magentic version is worth the cost?

This isn't quite answering your question, but I have the Galaxy v2 M and I quite like it. My initial impression of the puzzle was lukewarm at best, but it's really all around better than the Galaxy v1. Is it worth it over the nonmagnetic Galaxy v2? Hard to answer, I don't have both. Some people (including myself) have mentioned that the magnetic "click" is fairly pronounced in this puzzle, so that's something to take into account.

 

[mark49152]

Convergence is a process, so technically it starts converging on the very first move, but the convergence is kinda slow.

Yes. Please excuse my lazy wording but I do of course mean that by 40 moves it "reaches a sufficient level of convergence" as opposed to "starts to converge". How much convergence is sufficient is an open question, as you say, but 4% is way better than 40%. Thank you for finding the source.

 

[JanW]

Is 4% an acceptable deviation? Maybe if you use a method that doesn't involve EO at the start, i.e. pretty much anything other than ZZ and 3OP. However, it's fairly easy to get random-state scrambles these days, so I don't think it's fair to count anything but solves done in "ideal" scrambling conditions.

Keep in mind that in this case 4% more likely means 1.04/2048 instead of 1/2048. In other words you would get on average one extra scramble with all edges oriented every 50 000 solves. That is perfectly acceptable. Even with the 46% increase it would mean getting that case one time too many in about 4500 scrambles. Not really a big deal.

 

[pglewis]

(Also singles are pretty dumb and IMO they're just for fun, so however seriously you want to treat them is really up to you. Handscrambles all the way? Sure, whatever. Only random-state scrambles with cube cover and stackmat and all that stuff? That's cool too! The above paragraph mostly applies only if you want to measure how fast you really are.)

Exactly. I'd personally be more stringent about "counting" my singles if I were 10+ seconds faster, too. At this point I pretty much know that a good one + luck should be roughly .6 * my global and darned if mine isn't about a tenth off that. That fist pump when you actually get it is a great moment though.