StachuK1992
statue
And he'd like to document it.
I'm sorry, but I just had to.
This really seems like an interesting method.
-
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] Tony Snyder solves the cube
- Thread starter irontwig
- Start date
And he'd like to document it.
I'm sorry, but I just had to.
This really seems like an interesting method.
Tony Snyder
Member
As to the fnial CE, I orient at a minimum one LLE as I put it in, plus I line up one LLE edge.
As to the 20% claim, it is approximately the number of times I use one look for the last layer. However, I only use a shortest alg some of these times, while the rest are combos planned in advance. Also, most of my algs were human originated, so I have a ways to go in switching to the very best algs, I'll be working on that over the coming months.
As to the 20% claim, it is approximately the number of times I use one look for the last layer. However, I only use a shortest alg some of these times, while the rest are combos planned in advance. Also, most of my algs were human originated, so I have a ways to go in switching to the very best algs, I'll be working on that over the coming months.
Kirjava
Colourful
I don't think anyone learns algs rote with the intention of finding fingertricks for them in a later session.
People generally learn fingertricks for an alg when they... learn the alg.
joey
Member
Tony: Can you show us an example of 1/2 an alg + other alg + finish first 1/2.
Tony Snyder
Member
You read it wrong. What I'm saying is that for about an hour or two per week (on average), I find I can finger extremely fast regardless of the scenarios that I encounter.
StachuK1992
statue
Sorry, I was misinterpreting intentionally for innuendo.
I get what you mean; you're in "the zone."
Tony Snyder
Member
yes Kirjava, that's true about learning fingering as you learn algs, but how does that connect with what I said?
joey
Member
Examples, please.
Kirjava
Colourful
You said many algs makes it hard to turn fast.
You said this is because you have to learn 'fingering' for the many algs.
I'm saying that if you know the algs, you already know the fingering - so this isn't why many algs make it hard to turn fast.
You said this is because you have to learn 'fingering' for the many algs.
I'm saying that if you know the algs, you already know the fingering - so this isn't why many algs make it hard to turn fast.
Last edited:
irontwig
Member
Sounds like an simple insertion that he has memorized for a particular case to me.
Edit: A quick example:
Setup: R U R' U' B' R' F R B F'
F R U R' U' F' leaves three corners, so you can do: F R U F' L' F R2 F' L F R U' F' where the bold part is a corner cycle with a move cancelled. Yeah, yeah it's three moves from optimal, whatevah.
Last edited:
Tony Snyder
Member
Learning an alg, and learning the best fingering for it are two separate phases for me. In the past I have focused a lot more on the former and less on the latter. However, I am now attempting to improve both.
There are some subconscious switches that I don't have a good handle on yet that are somehow involved in the fingering technique I use when I go fast. Most of the time it is frustrating as the cube will get stuck or I'll turn a side too far at least twice per solve. I think I'm naturally clumsy and just occassionally get past that to finger it fast.
I have made liberal use of moving two layers at once to reorient the cube as I turn it, and this has helped in getting around some of the awkward fingering scenarios. On occassion when I'm real precise it doesn't seem to matter, but with nerve damage in my right wrist just about anything interferes. For example, I recently tightened all the screws on my F2 one quarter turn each then glued them down, and discovered it was just perfect, and I could go super fast for about 2 days. At one time I was averaging about 15 seconds based on a hand clock, though I didn't have a working digital timer on me so I can't say for sure that its a new best average. Then after a couple days it was back to struggling again, the cube didn't feel perfect anymore.
There are some subconscious switches that I don't have a good handle on yet that are somehow involved in the fingering technique I use when I go fast. Most of the time it is frustrating as the cube will get stuck or I'll turn a side too far at least twice per solve. I think I'm naturally clumsy and just occassionally get past that to finger it fast.
I have made liberal use of moving two layers at once to reorient the cube as I turn it, and this has helped in getting around some of the awkward fingering scenarios. On occassion when I'm real precise it doesn't seem to matter, but with nerve damage in my right wrist just about anything interferes. For example, I recently tightened all the screws on my F2 one quarter turn each then glued them down, and discovered it was just perfect, and I could go super fast for about 2 days. At one time I was averaging about 15 seconds based on a hand clock, though I didn't have a working digital timer on me so I can't say for sure that its a new best average. Then after a couple days it was back to struggling again, the cube didn't feel perfect anymore.
I guess he means basically:
More algs = more algs to get fast at = lower TPS overall.
For example, think you had 1 hour to get as fast as you could with a new alg. Now think you had one hour to get as fast as you could with 20 new algs. Since there is more algs, there is more to learn and to get good at. You would most likely be slower on with the 20 algs average compared to the single alg, because you would not have as much time to practice each alg.
You said somewhere that you could turn 8 times per second with this method, Tony. Did I misunderstand that? 8 TPS? I use Petrus, and even after my 2x2x3 phase, when the insanity begins and I'm RURUR * 1,000,000 I generally only hit 5-6 TPS on a simple solve. Also, I don't see much of a point to your final step. Most decent cubers will average 3-4 seconds on LL, and I doubt you'd average anything under 5 using this approach. Normally I wouldn't speak so negatively of a method like this, but you named a method after yourself that is basically 1/2 Petrus + a bunch of stuff that nearly everyone has already thought of. Also, you take like 60 second inspection to demonstrate your solve, but most people only take 5-10 seconds to inspect. If you could replicate your move count with 15 seconds inspection, and average 15~, then I would give you a high five, maybe. Thanks, -KYLOL
qqwref
Member
- Joined
- Dec 18, 2007
- Messages
- 7,834
- Location
- a <script> tag near you
- WCA
- 2006GOTT01
- YouTube
- Visit Channel
You should really give qqtimer a try. It's nice to have accuracy in your times, plus there are a lot more options and better statistics than cubetimer.
I think he means burst tps during an algorithm, which is not bad considering he doesn't use the most modern-style fingertricks.
Tony Snyder
Member
I'm confident that I have hit 8 TPS for partial solves (yes, I meant "burst TPS", thanks for the clarification) on a number of occassions, but not regularly. The fastest I hit on a weekly basis is probably about 6 for a partial solve, and far less for a full solve (I nearly always stumble at some point due to my bad wrist). Also, I realize my method is far from perfect, as I developed it entirely solo and never used a computer, adding only about 10-15 algs total from other sources.
Though I've been thinking about it for a long time, I am just now deciding to work on the serious improvements that this method needs to catch up with the other methods. However, for a 1980's method I'm convinced that it was quite good for that era, and original. At least no one that I encountered or knew about back then were using my techniques.
Most of my practice in the 80's was quite serious, I remember averaging 5 hours per day on it - crazy. I drove everyone nuts. Then most of my practice in the 90's and 00's was more a relaxation thing while I was focusing on other issues like work, so I only made minor improvements here and there during that time. And I did not modernize my method at all. And even in the 80's, though I could have written a program to upgrade the algs I didn't want to because it was so much fun thinking them up.
So I'm not trying to claim that I have anything great now, I'm really only filling in for some missing words in what others have written about my method. However, that got the spark going and I'm determined now to upgrade my method. I'm going to take the next couple years upgrading it in my spare time.
I have often explained to people the advantages of the Fridrich LL method, which is both balanced mathematically, and simplifies the visual recognition process. So whenever I encounter anyone wanting to learn speed solving I point them to Fridrich, as it is clearly an effective speed solving method with a relatively short learning curve. However, to improve beyond the Fridrich LL you need to find ways to approximate a LL direct solve, which is what I'm interested in, and in finding a way to represent this method so that it is easy to remember and execute fast. Instead of memorizing over 1200 algs, I'm convinced there will be a much simpler way to look at it. My current LLE+1C, then L3C method is just part way to that goal.
Here is an example: I now have a way to transfigure one alg into many different algs that solve every LL case where all LLE are in place yet 2 are not oriented, and where there is one corner solved and the other 3 are not in place and oriented any way. Nearly all of these transfigurations are quite easy to execute and to recognize. Granted, these are not direct solves, but it is a neat memory trick that cuts down the LL scenarios, producing many approximate direct solves with a single look and from a relatively simple memory pattern. I plan on coming up with similar memory tricks for all the LL scenarios.
In terms of inspection time that was a bad video, I agree - I normally take about 5-10 seconds to inspect.
I might be sub 15 on the solve now, I'm going to take some times this week. Switching to the F2 really made a difference. If I'm not sub 15 now then figure I should be there within the next few weeks.
Though I've been thinking about it for a long time, I am just now deciding to work on the serious improvements that this method needs to catch up with the other methods. However, for a 1980's method I'm convinced that it was quite good for that era, and original. At least no one that I encountered or knew about back then were using my techniques.
Most of my practice in the 80's was quite serious, I remember averaging 5 hours per day on it - crazy. I drove everyone nuts. Then most of my practice in the 90's and 00's was more a relaxation thing while I was focusing on other issues like work, so I only made minor improvements here and there during that time. And I did not modernize my method at all. And even in the 80's, though I could have written a program to upgrade the algs I didn't want to because it was so much fun thinking them up.
So I'm not trying to claim that I have anything great now, I'm really only filling in for some missing words in what others have written about my method. However, that got the spark going and I'm determined now to upgrade my method. I'm going to take the next couple years upgrading it in my spare time.
I have often explained to people the advantages of the Fridrich LL method, which is both balanced mathematically, and simplifies the visual recognition process. So whenever I encounter anyone wanting to learn speed solving I point them to Fridrich, as it is clearly an effective speed solving method with a relatively short learning curve. However, to improve beyond the Fridrich LL you need to find ways to approximate a LL direct solve, which is what I'm interested in, and in finding a way to represent this method so that it is easy to remember and execute fast. Instead of memorizing over 1200 algs, I'm convinced there will be a much simpler way to look at it. My current LLE+1C, then L3C method is just part way to that goal.
Here is an example: I now have a way to transfigure one alg into many different algs that solve every LL case where all LLE are in place yet 2 are not oriented, and where there is one corner solved and the other 3 are not in place and oriented any way. Nearly all of these transfigurations are quite easy to execute and to recognize. Granted, these are not direct solves, but it is a neat memory trick that cuts down the LL scenarios, producing many approximate direct solves with a single look and from a relatively simple memory pattern. I plan on coming up with similar memory tricks for all the LL scenarios.
In terms of inspection time that was a bad video, I agree - I normally take about 5-10 seconds to inspect.
I might be sub 15 on the solve now, I'm going to take some times this week. Switching to the F2 really made a difference. If I'm not sub 15 now then figure I should be there within the next few weeks.
Last edited:
qqwref
Member
- Joined
- Dec 18, 2007
- Messages
- 7,834
- Location
- a <script> tag near you
- WCA
- 2006GOTT01
- YouTube
- Visit Channel
As you develop your method, would you mind writing more details down? The wiki may not be the best place for extremely in-depth descriptions, but you could easily get something like a blog or small personal website set up, and then just write out how you would do things. I do suggest using the standard notation (R, (RL), M, r, x, etc.), not because your notation is bad, but because everyone is already used to the standard notation and it would just make your writings easier to follow. Thing is, even though many of your algs were found by hand, I'm sure a good portion of them are quite good (better than what you'd get if you just asked for an optimal solution). Every now and again someone comes up with a cool new way of doing one of the standard PLLs, or a new F2L trick, stuff which you wouldn't notice if you just tried all the optimal sequences that a program would spit out. And many of your techniques may subtly vary from the standard ones in our fewest moves and puzzle theory circles. If you've spent so many years refining your method and thinking of new ideas, you could easily be casually using strategies that nobody else has quite hit on yet.
Joël
Member
Tony Snyder... That name reminds me about my experiments with time dilation.
Kirjava
Colourful
Hah. qq mentioned that to me some time ago and I wondered where he'd seen it.
Tony Snyder
Member
Ok, sorry about the long time not posting, been dealing with work issues and have not yet pursued the new videos. Will try to get to that soon.
One quick comment on fingering. I am still (after 29 years) in experimental mode, after discovering over and over that I can sometimes go very fast, automatically sensing the best fingering as I go, and other times (most of the time), stumbling through it. I've found that the number one reason for this is that I was solving the cube out of frustration, as I was attempting to wind down after stressful work conditions. And so instead of thinking "smooth" "quiet", "flowing like water", "in tune" solves, I was trying to push it to go faster, and as a result the turns got choppier and with less precise control. This was like a compulsion, which is about 100x harder to master than the other mind-sets (not impossible). Sometimes in this mode I could make it go fast, but it is not at all elegant, so I'm working away from that compulsion now.
I think a lot has to do with the condition the mind, cube and fingers are in, for example, I usually bike to Star Bucks, and once there the blood has left my hands, the cube is cold and not evenly tightened. After an hour warm-up my hands work better but the cube may still be out of kilter. And then coffee often has a shaking effect on the hands, and once in awhile does not - then if the cube is also in perfect condition I will then go super fast. And I'm certain of the 8 TPS (bursting), even recently, though not regularly. Even 25+ years ago I would do the cube in 8 or 9 seconds once or twice per week, primarily due to fast fingers I had back then. And these weren't lucky solves, just more a psychological state then, and a physical+psychological state now.
Aside from these issues, there are also issues involving how to hold the cube, how to turn the sides, etc. These techniques are still in experimentation, as I find it hard to let go of any of the methods that I've come up with, even the bad ones - a little discipline is needed to retrain many of my algorithms to better fingering methods. And that means holding the cube in a different orientation, possibly changing algorithms, limiting fingering to specific turn methods, etc. My understanding of the modern fingering method (guessing here) is that I have to restrict my hand motions to specific ones that function more like pushing the side, and all-the-while keeping the cube in one position. I assume the modern fingering is a set of rules along these lines, and I'll try to read up on that (could someone provide a good liink for my research?), then I'll incorporate some things I've come up with on my own to broaden it a bit.
The super fast method is where a maneuver seems like one thought - kind of like on a typewriter when you get fast a word is just one thought, then your subconscious mind pulls together the details to facilitate that thought.
The time I did it in 7.0 seconds I had walked 10 miles, then sat down and drank about 4-5 pepsi's (a bit thirsty). Then I had one of those moments where it felt like everything around me was in perfect synchronization. At that moment I tested my time dilation theory - which is to trick my mind into believing there was a life threatening event, but not to focus on panic, to instead get the adrenalin rush that seems to slow down time and make the mind clearer (actually had this happen a few times before in real events, so I tried to reproduce those events in my mind). It worked, those 7 seconds seemed like plenty of time, I could form all the thoughts necessary to perfectly execute the solve. Normally I don't average anywhere close to that, and believe me 7 seconds is way way harder to do than 8 seconds, which I do on average once or twice per month even now.
Because of my bad wrist it is a lot harder for me to get my averages down, but I have never shy'd away from a challenge. I think it is possible I can overcome all that.
Sometime in coming weeks/months I'll setup a private website to talk about my thoughts on the cube.
One quick comment on fingering. I am still (after 29 years) in experimental mode, after discovering over and over that I can sometimes go very fast, automatically sensing the best fingering as I go, and other times (most of the time), stumbling through it. I've found that the number one reason for this is that I was solving the cube out of frustration, as I was attempting to wind down after stressful work conditions. And so instead of thinking "smooth" "quiet", "flowing like water", "in tune" solves, I was trying to push it to go faster, and as a result the turns got choppier and with less precise control. This was like a compulsion, which is about 100x harder to master than the other mind-sets (not impossible). Sometimes in this mode I could make it go fast, but it is not at all elegant, so I'm working away from that compulsion now.
I think a lot has to do with the condition the mind, cube and fingers are in, for example, I usually bike to Star Bucks, and once there the blood has left my hands, the cube is cold and not evenly tightened. After an hour warm-up my hands work better but the cube may still be out of kilter. And then coffee often has a shaking effect on the hands, and once in awhile does not - then if the cube is also in perfect condition I will then go super fast. And I'm certain of the 8 TPS (bursting), even recently, though not regularly. Even 25+ years ago I would do the cube in 8 or 9 seconds once or twice per week, primarily due to fast fingers I had back then. And these weren't lucky solves, just more a psychological state then, and a physical+psychological state now.
Aside from these issues, there are also issues involving how to hold the cube, how to turn the sides, etc. These techniques are still in experimentation, as I find it hard to let go of any of the methods that I've come up with, even the bad ones - a little discipline is needed to retrain many of my algorithms to better fingering methods. And that means holding the cube in a different orientation, possibly changing algorithms, limiting fingering to specific turn methods, etc. My understanding of the modern fingering method (guessing here) is that I have to restrict my hand motions to specific ones that function more like pushing the side, and all-the-while keeping the cube in one position. I assume the modern fingering is a set of rules along these lines, and I'll try to read up on that (could someone provide a good liink for my research?), then I'll incorporate some things I've come up with on my own to broaden it a bit.
The super fast method is where a maneuver seems like one thought - kind of like on a typewriter when you get fast a word is just one thought, then your subconscious mind pulls together the details to facilitate that thought.
The time I did it in 7.0 seconds I had walked 10 miles, then sat down and drank about 4-5 pepsi's (a bit thirsty). Then I had one of those moments where it felt like everything around me was in perfect synchronization. At that moment I tested my time dilation theory - which is to trick my mind into believing there was a life threatening event, but not to focus on panic, to instead get the adrenalin rush that seems to slow down time and make the mind clearer (actually had this happen a few times before in real events, so I tried to reproduce those events in my mind). It worked, those 7 seconds seemed like plenty of time, I could form all the thoughts necessary to perfectly execute the solve. Normally I don't average anywhere close to that, and believe me 7 seconds is way way harder to do than 8 seconds, which I do on average once or twice per month even now.
Because of my bad wrist it is a lot harder for me to get my averages down, but I have never shy'd away from a challenge. I think it is possible I can overcome all that.
Sometime in coming weeks/months I'll setup a private website to talk about my thoughts on the cube.
