My Solve Breakdowns (help please)

[retep]

Too long? Read only the blue stuff

Background info: Ok, so I am quite new to cubing, having only really started Jan 21, 2012 (I have solved the cube before this a couple years ago, but only like 10 times, never got below 5 mins). I currently average about 35 seconds and my current cube is a Guhong V2 (I just switched from what I thought was a decent rubik's brand and my times dropped about 10 seconds ). I use the Fridrich method (2 look oll, I know full pll).

I decided to break down my solves (and make use of qqtimer's session breakdown things). I timed how long it took me to do the 1) Cross 2) 4 F2L pairs 3) 2 look OLL 4) PLL. To do this I basically just double tapped the space bar so my times would include inspection and I wouldn't bias my results by getting a "free" look ahead. Here are my results:

1) Cross (this time I used a more appropriate/standard 15 sec inspection instead of figuring out the whole cross, although this means I was usually only able to solve for 2-3 pieces immediately and had to track and do cube rotation(s) for the other piece(s)).
avg12: 4.13 (σ = 0.74)
best time: 2.66
worst time: 5.85
4.72, 3.39, 4.34, 5.13, 4.80, 4.15, 4.15, 2.66, 5.85, 2.94, 3.16, 4.53

2) 4 F2L pairs (Inspection from cross to 1st pair is included in this time, which is generally about 3 seconds, I know I need to work on this)
avg12: 18.07 (σ = 2.18)
best time: 12.30
worst time: 20.93
18.38, 14.42, 17.82, 18.35, 18.64, 18.08, 12.30, 20.72, 14.28, 20.93, 20.14, 19.82

3) 2 look OLL (in brackets is the number of edges correctly oriented already (I try to use partial edge control if I see the opportunity, but that is rather rare), inspection time for OLL is generally very small as there is limited cases which are easy to spot just by looking at the top of the cube)).
avg12: 5.61 (σ = 0.66)
best time: 2.28
worst time: 6.91
6.91[0E], 4.66[4E], 2.28[4E], 6.39[2E], 6.39[2E], 6.10[2E], 5.60[2E], 5.32[2E], 6.27[2E], 4.79[2E], 5.64[2E], 4.96[2E]

4) PLL (as mentioned I know full PLL, these times including inspecting, which unlike my 2 look OLL, is rather long, I know from badmephisto's pll recognition video how to determine which case it is, but I spend like 2 seconds looking around the cube for the distinguishing features, then another like half second doing a cube rotation (or sometimes adjusting U layer) to get it properly aligned and then finally a few seconds to execute. I included in brackets which PLL it was, although not overly specific).
avg12: 6.75 (σ = 0.90)
best time: 4.71
worst time: 9.33
5.54, 6.98[G], 4.71[T], 6.51[R], 5.86[F], 6.36[G], 8.31[G], 7.27[G], 5.81[F], 9.33[N], 7.04[Y], 7.83[H]

Conclusion: Clearly my F2L needs work as more than half my time is coming from there, I used to be able to look ahead fairly well with my old cube, but I am in love with the speed of my new one and just going as fast as I can, in a few days I will work on look ahead with my new cube (just enjoying it mostly for now ). My PLL times are also kind of low, considering I know full pll, I will be working on recognizing as well as possibly posting times of purely my PLL execution, just to compare how much is actually recognition and if my gut feeling is correct. I also want to be able to plan out my entire cross in the 15 seconds as well as be able to look ahead to first F2L pair, but my cross times seem to be at least decent and it will kind of be a side note for now (can't try to do everything at once).

Reliability: There is of course some error involved in these times but adding the averages for each part gives: 34.56 seconds per solve on average, which is quite quite comparable to my average solve (within about .3 seconds of my average of 100)

So anyway I had a couple questions:
- Does this seem typical for some one averaging my time?

- Is there something glaringly bad/good about these times?

- I was wondering, does recognizing how to solve the F2L pairs simply come with practice? I kind of look ahead and can often spot a pair I can potentially do next, however I spend like 3ish seconds just looking at or playing with it until I get it to a familiar spot and know what to do (ie. separated into the top layer, one of the 3 basic cases, not counting when it is already paired).

- Should I bother learning full OLL, from my times I noticed my 2 look OLL is generally better than my full PLL, and I think it is largely due to having increased possibilities = increased inspection time to figure out what I need to do during the PLL.

- How long does it take to get used to a new cube? I just recently got a Guhong V2 (switching from rubik's brand), and while my times did drop about 10 seconds, this new cube feels very awkward. It feels like it is held together with a piece of fine string and so fast that I have to hold it differently and can't do some "finger tricks" I did previously with my other cube because the layers will "slip", which was never an issue previously. Although this does open up some new possibilities as well (eg. for H perm I can now move the middle layer, whereas that was impossible with my previous cube (I did M2 as r2 R2)).

- Should I practice my PLLs, I have only performed each about 20 times while memorizing it and then only when it comes up in my solves after that, I am just wondering if it would be worthwhile as I am pretty sure I could shave like 2 seconds off a few of the algs and while using Badmephisto's PLL trainer to time 10 of each PLL (see below) I could see and feel my times getting better, just from less than 10 times doing it.

- What is a good/typical F2L:LL ratio? (mine from the times above is 22.2/12.36~1.8:1

I also want to say, please don't just tell me to look ahead, I already know it is something I need to work on and this forum is flooded with that comment already.

EDIT:
Ok so I finally got my Guhong V2, and WOW, it is SOO fast and feels SOO different. With my first avg of 12 my times dropped by about 8 seconds and I beat my old PB 3 times Although like I mentioned it feels very different, it feels like it is held together with a very thin string and turns almost too easily, some old things that I did with my old cube I can't actually do with this one because other sides like slip. Anyway, going to have to update this thread.

EDIT 2:
Thread has been updated with new times, old times are in white (below) if you are curious (hold mouse button and drag (highlight) to see).

EDIT 3:
OK so I timed my PLLs and was actually a little surprised. I did 10 solves of each PLL (generated randomly by Badmephisto's PLL trainer and disabled once 10 were reached). All the PLLs and names can be found here. Here are my results (from worst to best, in seconds):
N(b) - 5.43
N(a) - 5.33
Z - 5.09
G(b) - 4.71
F - 4.54
G(c) - 4.44
G(a) - 4.34
E - 4.27
R(a) - 4.07
V - 4.01
H - 3.97
Y - 3.91
J(a) - 3.80
R(b) - 3.77
G(d) - 3.71
A(b) - 3.35
U(a) - 3.14
A(a) - 2.76
T - 2.62
U(b) - 2.52
J(b) - 2.44

I know my H and Z perms are particularily bad considering how easy they are to execute, this is because on my old cube I couldn't move just the middle layer (it was too stiff) so I instead did r2 R2 and after getting my new cube I am working on using the middle layer to perform this alg, but am still new to doing so (only had this cube for 1 day). I think I would benefit most from maybe finding a better way to do the G perms (not necessarily new alg, was thinking more about new finger tricks), would help if some one could link a video of G perm finger tricks please and thanks.

1) Cross (I used more than the 15 sec inspection time several times, like i said I am still quite new, although I generally use 8 or less moves to solve it):
Times:
6.63, 8.43, 5.54, 6.00, 4.52, 2.89, 1.59, 2.84, 5.34, 5.99, 4.81, 4.34
best time: 1.59
worst time: 8.43
avg12: 4.89 (σ = 1.28)

2)F2L (I included inspection time (I don't yet look ahead from cross to F2L pair), so like the first 3ish seconds are me looking for and figuring out how to solve a pair):
Times:
25.61, 26.29, 28.09, 17.93, 18.95, 23.49, 27.87, 18.34, 19.98, 18.98, 23.30, 25.62
best time: 17.93
worst time: 28.09
avg12: 22.84 (σ = 3.52)

3) 2-look OLL (I included how many edges were already oriented correctly, the time also includes inspection time, but using 2 look oll, that is minimal as there isn't as much to look for):
Times:
5.55[2E], 5.45[2 E], 5.37[2E], 2.90[4E], 3.40[4E], 5.91[2E], 7.35[2E], 6.78[2E], 6.77[2E], 5.74[2E], 7.59[2E], 4.20[4E]
best time: 2.90
worst time: 7.59
avg12: 5.65 (σ = 1.19)

4) PLL (again, I included inspection time, which is actually, imo, quite bad, like 2-5 secs and also included the perm, although not very specific (not which G perm for example)):
Times:
8.49[G], 9.75[G], 7.36[A], 8.98[G], 7.29, 9.11[G], 8.50[G], 9.22[G], 5.46, 7.96[Y], 8.35[G], 7.88[G]
best time: 5.46
worst time: 9.75
avg12: 8.31 (σ = 0.69)

So in total, adding the averages of 12 gives an average solve time of 41.69, which is quite close to my average, so I feel pretty good about these results, that they are fairly accurate.

 

[Hodari]

I'd say the thing I noticed most is that your full PLL is taking much longer than 2 look OLL, so you should probably work on that a bit. If it's mostly inspection time as you mentioned, I'd say watch some videos on PLL recognition(Badmephisto's is especially good) first of all and try to get that down a bit. Otherwise, obviously F2L is more than half of your total time, so that's the next place to work on(putting PLL first here only because it's probably a fairly quick improvement while F2L will take much longer)

 

[balloon6610]

You are pretty good actually my advice is practice recognising PLL, Although you know all of them but take so much time to recognise which PLL you need to apply then it's useless. For full OLL it's depend on how fast you want to get , Full OLL will reduce half of your 2 look OLL time (since you need to apply 1 algorithm) but you still can get sub 30 with 2 look OLL as well . For F2L is quite good enough but i would say try to slow down and practice look ahead this will help you a lot when you speed up and can solve consecutively rather than Finding first pair - Solve- Pause and finding another pair- Solve and repeat on all 4 pairs just practice
For Cross your time is quite good enough i am sub 30 but my still need 10 moves to solve the cross
For a new cube guhong and zhanchi is very good and can push my times down about 3-5 seconds BUT you should get average about 30 seconds first simply because those cube are as fast as hell and if your look ahead are bad then it's worthless for beginner i recommend try some DIY cube first since you can maintain your speed while look ahead,Then you think you ready then buy guhong or zhanchi or whatever you want don't give up

Hope this help

 

[skeletonboy]

Your PLL. A lot of people can be using 2 look OLL and 2 look PLL and beginner cross and intuitive F2L without lookahead and still average around 30 seconds.

But, since you are already that far ahead in algorithms, like your Full PLL, I don't see why it would take an average of 8 seconds to execute an algorithm. Work on your inspection and try out different algorithms to see which alg fits you best. The algorithm really shouldn't be over 4 seconds, and a lot of the time in your PLL seems to be recognition. I recommned learning some recognition methods, especially for G perms. To be honest, after I learned the G Perm recognition technique, the G-Perm seems to be easier than Y-Perm for me.

Like for the G-Perm, make sure the "headlight" is on the left side of the cube, now see the 2x1x1 block? Depending on where that block is, you know what algorithm to execute. Further information can be told using the Youtube tutorials, I can't really explain it verbally.

As for the F2L, try practicing the executions of specific cases until you can do it subconsciously, maybe while drinking milk...(hehe). Then, when you feel that you basic F2l is sturdy, try to look ahead around the cube for the other F2L pieces/pair. Technically, substitute the milk for F2L pair. (hehe). You may need to turn slower because if you turn at your full speed, the pieces will be flying around the cube too fast for your eyes to recognize.

So practice until comfortable in F2L, then try out lookahead if you want.

Good Luck!

 

[retep]

Ya, thanks guys, just wanted to comment on a few things:

- I did watch badmephisto's PLL recognition video, although it was before I started learning full PLL (lol), so I may take another look, perhaps I missed something useful. I do implement at least most of the things, but for example, with the G perm I look around the cube (I tilt though, maybe that is bad?) until I find the headlights, I then look for the connected 2x1x1 block, after that I reposition the U layer (or do cube rotation(s), kind of just whatever happens, i don't think about it), then recall and execute the algorithm. (It is generally several steps like this for all my PLLs, I almost always look for the headlights or a 3x1x1 (bar) first, then look for a second defining feature (2x2x1 block, connected corner block (with a correct edge on both sides of the corner, and extended bar etc.), then rotate the cube/U Layer, then recall and execute alg.
Wheras, with 2 look OLL I just merely have to look at the top of the cube to see what I need to do, and maybe look at the layer facing me to distinguish like Sune from Anti-Sune, so the recognition is very quick, then I just need to position the cube (through U layer or cube rotation) and execute the algorithm.

- I am getting a Guhong V2 hopefully this week, should help a lot (I have looked at some videos and it seems like my cube is garbage, I watched a one hand solve and laughed, I can only turn with my index finger one handed, it is too stiff to turn with my pinky, and my hand gets tired quick. Even doing 2 handed solves I generally have to stop after like 10-15 solves because my hand is rather sore and tired.)

- Also as I mentioned I do kind of look-ahead during my F2L solves (due to my slow cube it isn't overly hard and probably is the only reason I have the times I do), however I still don't know all cases from all angles and most cases I do far from optimally. I also use like 1 cube rotation on average per pair (I sometimes use 0 or 2, but rather rarely). I know some people say don't look ahead until you are good with F2L and know all the cases and such, and i kind of see where they are coming from, but I think learning it as early as possible will just make it that much easier in the future, both can be developed with time, why does it matter which I choose to do first?

Anyway, thanks again for your replies, keeping it in mind

 

[Carson]

Just a tip for breaking down your times: Don't do it by individual step. You will never be able to recreate the conditions of a full solve by starting in the middle. The best way to get an accurate idea of your breakdown is do an average of 12 stopping the timer after f2l, then an another average of 12 stopping after f2l, then oll, then pll. Do the math to figure out the times for the individual steps.

 

[retep]

Just a tip for breaking down your times: Don't do it by individual step. You will never be able to recreate the conditions of a full solve by starting in the middle. The best way to get an accurate idea of your breakdown is do an average of 12 stopping the timer after f2l, then an another average of 12 stopping after f2l, then oll, then pll. Do the math to figure out the times for the individual steps.

Ya, I was considering doing this, but double tapping the space key seemed good enough, I mean there might be like a 0.2 second error, which isn't particularly large, also the error should be fairly consistent between the times. In fact, the error in your method may be greater, I mean if you get a very easy cross and do F2L and OLL well, you might end up with a time lower than your average of 12 for the F2L, what then? You can't have a negative time for your OLL (for example my best cross, F2L and OLL times added together equal 22.42s whereas my average F2L time was 22.84s, I can't really record an OLL time of -0.4s...)

 

[Rubiksfreak]

<snip>

Well for a beginner i dont think your times are terrible but theyre not amazing either! Yes, if you practice enough you will be able to do F2L without pausing to look for pieces. About the playing around with it to get it to a familiar spot to solve it, id suggest just playing with the cube. Once you get more familiar with the cube and how to pieces move youll be able to solve an F2L pair irregardless of the position of the pieces. Will buying a different cube help my times? It really depends on how bad your cube is, get some silicone spray it works best. I can easily solve a rubiks brand cube in 10 seconds if its modified correctly, but then again ive played with some rubiks cube brands that were horrible. Should you bother learning full Fridrich method? Well this question is totally up to you, but yes this would greatly increase your times. Just keep working on it and before you know it youll find your self solving the cube without even having to think about what your doing.