Loading ...
Sorry, an error occurred while loading the content.
 

What's in an algorithm?

Expand Messages
  • cmhardw
    Hey I ve been thinking about something a lot lately. I m starting to really believe that a whole LOT of what makes a speedcuber is the algorithms you use for
    Message 1 of 3 , Feb 29, 2004
      Hey I've been thinking about something a lot lately. I'm starting
      to really believe that a whole LOT of what makes a speedcuber is the
      algorithms you use for the final steps of your method. I've always
      heard that people should find algorithms that fit your hands, but to
      me it seems odd that Peter Jansen's algs seem to fit EVERYONE's
      hands. Perhaps there are algorithms that are better suited to be
      performed by human hands in general, and Peter has apparently
      compiled many of these. Ever since I've started one by one slowly
      replacing my slowest LL alg with Peter's alg I speed up around 0.5
      seconds on average per alg. It just seems odd to me that I was
      stuck at just above 20 seconds for almost 3 years using a certain
      set of LL algs, and within a month or two of starting to replace my
      algs I broke 20 seconds on average for the first time. Now that
      I've learned a number of his algs I can break 20 seconds
      consistently.

      It seems that eventually all of us speedcubers will use the same
      algs and everyone will average around 15-17 seconds.

      I don't know maybe I'm just being weird, but I think a LOT of what
      makes a speedcuber (once you've figured out how to optimize the F2L)
      is simply how fast you can solve the LL. And I think the algs you
      use are what make or break you here.

      Any opinions on this? I'm curious if there are people out there who
      don't like Peter's algs? I don't mean to demean any of the work
      that Peter did to compile his list of LL algs, and so far I LOVE the
      algs of his that I've learned, I'm just curious if anyone anywhere
      doesn't like one of Peter's algs. If there isn't anyone that
      doesn't like Peter's algs, perhaps they are the ones that are
      optimized for human hands in general, and therefore anyone using
      them would find them to be very fast.

      Just wanted to bring up a topic for discussion,
      Chris
    • Kyle Bryant
      Ok, since Im the one that likes to construct arguments as a separate hobby, I would like to say: Makes sense doesn t it? about your theory Chris. lets divide
      Message 2 of 3 , Feb 29, 2004
        Ok, since Im the one that likes to construct arguments
        as a separate hobby, I would like to say:

        Makes sense doesn't it?

        about your theory Chris.

        lets divide methodology of the LBL methods into two
        submethods

        f2L - LL

        now further into four methods (COx = Cross on D,L,R,
        or U)
        COx F2L - OLL PLL

        now, analyzing these two sub groups in terms of
        algorithms:

        1 41 - 57 21

        since the Cross step is usually intuitive, we'll give
        it a 1 since the "algorithm" is usually something you
        make up during inspection

        adding it all up

        1 + 41 + 57 + 21 = 120

        42/120 = F2L
        78/120 = LL

        now, just by standard mathematics, we see that the LL
        has more memorizable situations...therefore more
        algorithms to learn to accomplish your goals.

        since the LL clearly is more heavy concerning the
        amount you have to learn (and therefore the amount you
        must cycle through in your head [algorithm
        recognition]), it is understandable that any
        reduction, or balancing in the ease of those
        algorithms or quality would affect your time in a
        positive way.

        I may not make any sense to you...but I do to me and
        thats all that counts I suppose.
        -=K-=

        --- cmhardw <no_reply@yahoogroups.com> wrote:
        > Hey I've been thinking about something a lot lately.
        > I'm starting
        > to really believe that a whole LOT of what makes a
        > speedcuber is the
        > algorithms you use for the final steps of your
        > method. I've always
        > heard that people should find algorithms that fit
        > your hands, but to
        > me it seems odd that Peter Jansen's algs seem to fit
        > EVERYONE's
        > hands. Perhaps there are algorithms that are better
        > suited to be
        > performed by human hands in general, and Peter has
        > apparently
        > compiled many of these. Ever since I've started one
        > by one slowly
        > replacing my slowest LL alg with Peter's alg I speed
        > up around 0.5
        > seconds on average per alg. It just seems odd to me
        > that I was
        > stuck at just above 20 seconds for almost 3 years
        > using a certain
        > set of LL algs, and within a month or two of
        > starting to replace my
        > algs I broke 20 seconds on average for the first
        > time. Now that
        > I've learned a number of his algs I can break 20
        > seconds
        > consistently.
        >
        > It seems that eventually all of us speedcubers will
        > use the same
        > algs and everyone will average around 15-17 seconds.
        >
        > I don't know maybe I'm just being weird, but I think
        > a LOT of what
        > makes a speedcuber (once you've figured out how to
        > optimize the F2L)
        > is simply how fast you can solve the LL. And I
        > think the algs you
        > use are what make or break you here.
        >
        > Any opinions on this? I'm curious if there are
        > people out there who
        > don't like Peter's algs? I don't mean to demean any
        > of the work
        > that Peter did to compile his list of LL algs, and
        > so far I LOVE the
        > algs of his that I've learned, I'm just curious if
        > anyone anywhere
        > doesn't like one of Peter's algs. If there isn't
        > anyone that
        > doesn't like Peter's algs, perhaps they are the ones
        > that are
        > optimized for human hands in general, and therefore
        > anyone using
        > them would find them to be very fast.
        >
        > Just wanted to bring up a topic for discussion,
        > Chris
        >
        >
        >
        > ------------------------ Yahoo! Groups Sponsor
        >
        >
        > Yahoo! Groups Links
        >
        >
        http://groups.yahoo.com/group/speedsolvingrubikscube/
        >
        >
        > speedsolvingrubikscube-unsubscribe@yahoogroups.com
        >
        >
        >


        __________________________________
        Do you Yahoo!?
        Get better spam protection with Yahoo! Mail.
        http://antispam.yahoo.com/tools
      • Eric Johanson
        assuming a general fridrich approach, i think what makes a speedcubist is 2/3 F2L and 1/3 LL. the LL is relatively simple compared to F2L, in my opinion.
        Message 3 of 3 , Mar 1, 2004
          assuming a general fridrich approach, i think what makes a
          speedcubist is 2/3 F2L and 1/3 LL.

          the LL is relatively simple compared to F2L, in my opinion.

          ironicly, when you look at a list of algs, there are
          fewer "theoretical" cases listed for F2L compared to the theoretical
          cases listed for LL, but there are way more "real-life" cases in
          F2L. it requires improvisation to map the real-life case to the
          correct theoretical case by apply a move or two and/or changing grip
          position. it is this improvisation that makes (or breaks) a
          speedcubist, in my opinion.

          furthermore, a top speed cubist will place all 4 corner-edge pairs
          in 8 seconds total. a top speed cubist will solve the LL in about 6-
          7 seconds total. now suppose you get a whole bunch of beginners and
          have half of them work *only* on CE pairs, and the other half work
          *only* on LL. i believe it will take far less practice for those
          working on LL to reach expert-level LL times (6-7 seconds), than for
          those working on F2L to reach expert-level F2L times (8 seconds).

          again, all in my opinion.

          -eric


          --- In speedsolvingrubikscube@yahoogroups.com, Kyle Bryant
          <craptastic_crap@y...> wrote:
          > Ok, since Im the one that likes to construct arguments
          > as a separate hobby, I would like to say:
          >
          > Makes sense doesn't it?
          >
          > about your theory Chris.
          >
          > lets divide methodology of the LBL methods into two
          > submethods
          >
          > f2L - LL
          >
          > now further into four methods (COx = Cross on D,L,R,
          > or U)
          > COx F2L - OLL PLL
          >
          > now, analyzing these two sub groups in terms of
          > algorithms:
          >
          > 1 41 - 57 21
          >
          > since the Cross step is usually intuitive, we'll give
          > it a 1 since the "algorithm" is usually something you
          > make up during inspection
          >
          > adding it all up
          >
          > 1 + 41 + 57 + 21 = 120
          >
          > 42/120 = F2L
          > 78/120 = LL
          >
          > now, just by standard mathematics, we see that the LL
          > has more memorizable situations...therefore more
          > algorithms to learn to accomplish your goals.
          >
          > since the LL clearly is more heavy concerning the
          > amount you have to learn (and therefore the amount you
          > must cycle through in your head [algorithm
          > recognition]), it is understandable that any
          > reduction, or balancing in the ease of those
          > algorithms or quality would affect your time in a
          > positive way.
          >
          > I may not make any sense to you...but I do to me and
          > thats all that counts I suppose.
          > -=K-=
          >
          > --- cmhardw <no_reply@yahoogroups.com> wrote:
          > > Hey I've been thinking about something a lot lately.
          > > I'm starting
          > > to really believe that a whole LOT of what makes a
          > > speedcuber is the
          > > algorithms you use for the final steps of your
          > > method. I've always
          > > heard that people should find algorithms that fit
          > > your hands, but to
          > > me it seems odd that Peter Jansen's algs seem to fit
          > > EVERYONE's
          > > hands. Perhaps there are algorithms that are better
          > > suited to be
          > > performed by human hands in general, and Peter has
          > > apparently
          > > compiled many of these. Ever since I've started one
          > > by one slowly
          > > replacing my slowest LL alg with Peter's alg I speed
          > > up around 0.5
          > > seconds on average per alg. It just seems odd to me
          > > that I was
          > > stuck at just above 20 seconds for almost 3 years
          > > using a certain
          > > set of LL algs, and within a month or two of
          > > starting to replace my
          > > algs I broke 20 seconds on average for the first
          > > time. Now that
          > > I've learned a number of his algs I can break 20
          > > seconds
          > > consistently.
          > >
          > > It seems that eventually all of us speedcubers will
          > > use the same
          > > algs and everyone will average around 15-17 seconds.
          > >
          > > I don't know maybe I'm just being weird, but I think
          > > a LOT of what
          > > makes a speedcuber (once you've figured out how to
          > > optimize the F2L)
          > > is simply how fast you can solve the LL. And I
          > > think the algs you
          > > use are what make or break you here.
          > >
          > > Any opinions on this? I'm curious if there are
          > > people out there who
          > > don't like Peter's algs? I don't mean to demean any
          > > of the work
          > > that Peter did to compile his list of LL algs, and
          > > so far I LOVE the
          > > algs of his that I've learned, I'm just curious if
          > > anyone anywhere
          > > doesn't like one of Peter's algs. If there isn't
          > > anyone that
          > > doesn't like Peter's algs, perhaps they are the ones
          > > that are
          > > optimized for human hands in general, and therefore
          > > anyone using
          > > them would find them to be very fast.
          > >
          > > Just wanted to bring up a topic for discussion,
          > > Chris
          > >
          > >
          > >
          > > ------------------------ Yahoo! Groups Sponsor
          > >
          > >
          > > Yahoo! Groups Links
          > >
          > >
          > http://groups.yahoo.com/group/speedsolvingrubikscube/
          > >
          > >
          > > speedsolvingrubikscube-unsubscribe@yahoogroups.com
          > >
          > >
          > >
          >
          >
          > __________________________________
          > Do you Yahoo!?
          > Get better spam protection with Yahoo! Mail.
          > http://antispam.yahoo.com/tools
        Your message has been successfully submitted and would be delivered to recipients shortly.