• 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 35,000+ people from around the world today!

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

My quest for Sub 8 Petrus

Brest

Super Reconstructor
Staff member
Joined
Sep 30, 2010
Messages
2,310
YouTube
BrestCubing
^ I was a tiny bit surprised to see there aren’t any solves tagged as Petrus on cubesolv.es (although there are likely some tucked away in there somewhere).
There are a couple of old threads on here (sub-20 with 20 methods, sub-10 with 10) which I know had Petrus solves reconstructed by @Brest which might be of interest?
Unfortunately I don't think the tag system was ever fully implemented.
Search by cuber is probably the best way to find solves, if you know who uses Petrus.

www.cubesolv.es/2349
www.cubesolv.es/5379
 

Tao Yu

Member
Joined
Jun 29, 2011
Messages
1,062
Location
Ireland
WCA
2012YUTA01
YouTube
taoxyu
I have a bunch of Petrus example solves in this playlist:

https://www.youtube.com/playlist?list=PLx1IBpMtcY5y-NSiYGmb9gZEGE2q7D0Gm

You can learn a lot from CFOP/Freefop and ZZ example solves as well, especially as all CFOP F2L cases can appear in Petrus, xcross often requires Petrus style blockbuilding, and because Petrus after the EO step is essentially the same as ZZ. There should be plenty of these on youtube.

I've reconstructed most of a 9.91 Petrus avg12 done by me here: https://pastebin.com/yJ9VShQD. The solutions are really bad though - you really don't need perfect solves for sub 10. If you get rid of many of the inefficient things I do, and develop blockbuilding lookahead further than others have done already, I think sub 8 should definitely be possible.

My biggest piece of advice is to experiment and be open minded. Nobody has gotten sub 8 Petrus before, and nobody knows the best way to practise to achieve it. So it's up to you to analyse your solves carefully, and decide on the best way forward for you. Good luck!
 

PetrusQuber

Member
Joined
Jun 27, 2019
Messages
432
Location
My house, cubing.
YouTube
N/A
Week 1's coming to an end, and I think I'll need some more time on 2x2x2 at the moment. I'll start learning Anti-Sune cases as well. Plus, can I just use the mirror of the Sune cases?

I have a bunch of Petrus example solves in this playlist:

https://www.youtube.com/playlist?list=PLx1IBpMtcY5y-NSiYGmb9gZEGE2q7D0Gm

You can learn a lot from CFOP/Freefop and ZZ example solves as well, especially as all CFOP F2L cases can appear in Petrus, xcross often requires Petrus style blockbuilding, and because Petrus after the EO step is essentially the same as ZZ. There should be plenty of these on youtube.

I've reconstructed most of a 9.91 Petrus avg12 done by me here: https://pastebin.com/yJ9VShQD. The solutions are really bad though - you really don't need perfect solves for sub 10. If you get rid of many of the inefficient things I do, and develop blockbuilding lookahead further than others have done already, I think sub 8 should definitely be possible.

My biggest piece of advice is to experiment and be open minded. Nobody has gotten sub 8 Petrus before, and nobody knows the best way to practise to achieve it. So it's up to you to analyse your solves carefully, and decide on the best way forward for you. Good luck!
Thanks!
 
Last edited by a moderator:
Joined
Oct 7, 2018
Messages
55
to anyone saying he should use roux zz or cfop you have entirely missed the point of this thread. he is not trying to become the quickest cuber in the world, he is trying to push the limits of an unproven method and I respect that.

good luck in your goal and can you put in your oh and feet(if you want to do it) Times it would be cool to see if Petrus is good for them.
 

PetrusQuber

Member
Joined
Jun 27, 2019
Messages
432
Location
My house, cubing.
YouTube
N/A
to anyone saying he should use roux zz or cfop you have entirely missed the point of this thread. he is not trying to become the quickest cuber in the world, he is trying to push the limits of an unproven method and I respect that.

good luck in your goal and can you put in your oh and feet(if you want to do it) Times it would be cool to see if Petrus is good for them.
Thanks! And sure, maybe I’ll start posting times as well.
Edit: Times out up.
 
Last edited:

tx789

Member
Joined
Sep 6, 2010
Messages
1,872
Location
New Zealand
WCA
2010HUNT02
YouTube
thetx789
Contrary to what some believe I know what I am talking about.

With COLL you reduce your PLLs from 21 to 4, the EPLLs. Yes U perms are very fast (the fastest for some) but Z and H perm aren’t as fast(for some they are slow because of M moves, U has fast RU alternatives) and there are many non-EPLLs that are faster. You have a 1/12 chance of H and 1/6 of Z (the slowest of these 4).

The three broad cateries of PLL are adjacent swap, diagonal swap, and EPLLs. Diag is the slowest on average because of N-perms (widely considered the worst PLL), the V-perm is a slower PLL too but the E-perm and Y-perm are fast.

The majority of PLLs are adj swap (which will appear 2/3 of the time). The fastest of these being Jb-perm, T-perm the A-perms. None of the adj are really bad PLLs unlike diag PLL. The R-perms are average, F perm is a bit slow (at least for me) and RUL Ja-Perm is a bit arkward. The G-perms are an average speed too despite what newer cubers think they are hard cause there is four of them and the recoginition looks hard compared to the rest of PLL.

The time difference between your adj PLLs and EPLLs is not that big. People think it is much bigger. Yes U-perms can be the fastest PLL but what is the next fastest? Not H and Z there are a lot of adj PLLs that are fast or comaparable in speed. H is faster than Z. Most PLLs have a 1/18 chance the ones that don’t being N-perms, Z-perm, E-perm, H-perm. N and H perm 1/72.

My personal recommendation is too lay off learning LL algs at this time and work on blockbuilding. Learning more LL won't make you much faster. COLL recognition is a useful skill for PLL prediction.

Ultimately do what is most fun for you or else you likely won't reach your goal. OLL/PLL is going to fast enough to get you fairly far. People use 2 look OLL with CFOP and get to 10.xx average because they are too lazy to learn algs.

If you want to quickly work on your what slows you down the most and LL won't be that for a long time.


One thing you could look at is pdesuo block. You likely find it hard but it will probaly we worth it in the long run. Look up FMC tutorial if you don't know what one is.
 

PetrusQuber

Member
Joined
Jun 27, 2019
Messages
432
Location
My house, cubing.
YouTube
N/A
Contrary to what some believe I know what I am talking about.

With COLL you reduce your PLLs from 21 to 4, the EPLLs. Yes U perms are very fast (the fastest for some) but Z and H perm aren’t as fast(for some they are slow because of M moves, U has fast RU alternatives) and there are many non-EPLLs that are faster. You have a 1/12 chance of H and 1/6 of Z (the slowest of these 4).

The three broad cateries of PLL are adjacent swap, diagonal swap, and EPLLs. Diag is the slowest on average because of N-perms (widely considered the worst PLL), the V-perm is a slower PLL too but the E-perm and Y-perm are fast.

The majority of PLLs are adj swap (which will appear 2/3 of the time). The fastest of these being Jb-perm, T-perm the A-perms. None of the adj are really bad PLLs unlike diag PLL. The R-perms are average, F perm is a bit slow (at least for me) and RUL Ja-Perm is a bit arkward. The G-perms are an average speed too despite what newer cubers think they are hard cause there is four of them and the recoginition looks hard compared to the rest of PLL.

The time difference between your adj PLLs and EPLLs is not that big. People think it is much bigger. Yes U-perms can be the fastest PLL but what is the next fastest? Not H and Z there are a lot of adj PLLs that are fast or comaparable in speed. H is faster than Z. Most PLLs have a 1/18 chance the ones that don’t being N-perms, Z-perm, E-perm, H-perm. N and H perm 1/72.

My personal recommendation is too lay off learning LL algs at this time and work on blockbuilding. Learning more LL won't make you much faster. COLL recognition is a useful skill for PLL prediction.

Ultimately do what is most fun for you or else you likely won't reach your goal. OLL/PLL is going to fast enough to get you fairly far. People use 2 look OLL with CFOP and get to 10.xx average because they are too lazy to learn algs.

If you want to quickly work on your what slows you down the most and LL won't be that for a long time.


One thing you could look at is pdesuo block. You likely find it hard but it will probaly we worth it in the long run. Look up FMC tutorial if you don't know what one is.
Learning COLL can’t hurt either. But yeah, you’re correct in that last layer is no the thing I need to work on. After the 2x2x2, I’ll work on my extension.
 
Last edited:
Joined
Oct 7, 2018
Messages
55
Contrary to what some believe I know what I am talking about.

With COLL you reduce your PLLs from 21 to 4, the EPLLs. Yes U perms are very fast (the fastest for some) but Z and H perm aren’t as fast(for some they are slow because of M moves, U has fast RU alternatives) and there are many non-EPLLs that are faster. You have a 1/12 chance of H and 1/6 of Z (the slowest of these 4).

The three broad cateries of PLL are adjacent swap, diagonal swap, and EPLLs. Diag is the slowest on average because of N-perms (widely considered the worst PLL), the V-perm is a slower PLL too but the E-perm and Y-perm are fast.

The majority of PLLs are adj swap (which will appear 2/3 of the time). The fastest of these being Jb-perm, T-perm the A-perms. None of the adj are really bad PLLs unlike diag PLL. The R-perms are average, F perm is a bit slow (at least for me) and RUL Ja-Perm is a bit arkward. The G-perms are an average speed too despite what newer cubers think they are hard cause there is four of them and the recoginition looks hard compared to the rest of PLL.

The time difference between your adj PLLs and EPLLs is not that big. People think it is much bigger. Yes U-perms can be the fastest PLL but what is the next fastest? Not H and Z there are a lot of adj PLLs that are fast or comaparable in speed. H is faster than Z. Most PLLs have a 1/18 chance the ones that don’t being N-perms, Z-perm, E-perm, H-perm. N and H perm 1/72.

My personal recommendation is too lay off learning LL algs at this time and work on blockbuilding. Learning more LL won't make you much faster. COLL recognition is a useful skill for PLL prediction.

Ultimately do what is most fun for you or else you likely won't reach your goal. OLL/PLL is going to fast enough to get you fairly far. People use 2 look OLL with CFOP and get to 10.xx average because they are too lazy to learn algs.

If you want to quickly work on your what slows you down the most and LL won't be that for a long time.


One thing you could look at is pdesuo block. You likely find it hard but it will probaly we worth it in the long run. Look up FMC tutorial if you don't know what one is.
my h and u perms (mu) are my fastest plls and coll also increases oll skip chance
 

Nilsibert

Member
Joined
May 25, 2013
Messages
412
Location
Switzerland
WCA
2014FREY01
YouTube
Nilsibert
My personal recommendation is too lay off learning LL algs at this time and work on blockbuilding. Learning more LL won't make you much faster. COLL recognition is a useful skill for PLL prediction.
While you are right in that coll/epll is not really something that will greatly improve his times, I disagree with laying off LL algs. He already stated before that he expects himself to take years to reach his goal. So then, sure, working on blockbuilding exclusively will make him get better times faster, but is that the goal?
Say he does that and ignores LL for the most part, and gets sub 10 for example. Say this takes 2 years(just an example). If he then decides to learn coll and ultimately zbll, while ofc still doing timed solves to practice what he learned before, he has to basically start from nothing. So let‘s say zbll will take a year to get to a decent level provided learning and drilling the algs is the main focus. That‘s 3 years in total.
I wonder, why not learn blockbuilding and all the other non LL stuff as usual, while also learning coll and actually using it? Sure it might not help his times, but getting good times asap seems to not be the goal.
If he practices for 1 year, and for all that time he‘s using coll his recognition will get really really fast. Once he‘s at that point, he can then gradually add more and more zbll subsets to his arsenal. So in the end, for those total of 3 years, he will get to practice coll/zbll constantly, which will help greatly in the long run.
Gradually learning zbll over a long time is also not as frustrating/boring/difficult as doing it all at once.

tl;dr: if getting faster asap is not the main focus, why not use a certain % of your cubing time to gradually learn algs?
 

tx789

Member
Joined
Sep 6, 2010
Messages
1,872
Location
New Zealand
WCA
2010HUNT02
YouTube
thetx789
Learning COLL can’t hurt either. But yeah, you’re correct in that last layer is no the thing I need to work on. After the 2x2x2, I’ll work on my extension.
Ultimately do whatever is most fun or else you won't reach your goal. It depends on what you want to do with the absolute fastest improvement would mean laying off LL. Also in your case experimenting with block building is something you will have to do a lot to get to your goal.

While you are right in that coll/epll is not really something that will greatly improve his times, I disagree with laying off LL algs. He already stated before that he expects himself to take years to reach his goal. So then, sure, working on blockbuilding exclusively will make him get better times faster, but is that the goal?
Say he does that and ignores LL for the most part, and gets sub 10 for example. Say this takes 2 years(just an example). If he then decides to learn coll and ultimately zbll, while ofc still doing timed solves to practice what he learned before, he has to basically start from nothing. So let‘s say zbll will take a year to get to a decent level provided learning and drilling the algs is the main focus. That‘s 3 years in total.
I wonder, why not learn blockbuilding and all the other non LL stuff as usual, while also learning coll and actually using it? Sure it might not help his times, but getting good times asap seems to not be the goal.
If he practices for 1 year, and for all that time he‘s using coll his recognition will get really really fast. Once he‘s at that point, he can then gradually add more and more zbll subsets to his arsenal. So in the end, for those total of 3 years, he will get to practice coll/zbll constantly, which will help greatly in the long run.
Gradually learning zbll over a long time is also not as frustrating/boring/difficult as doing it all at once.

tl;dr: if getting faster asap is not the main focus, why not use a certain % of your cubing time to gradually learn algs?
Doing what is most fun for him is what he should do. But laying off LL until say 16 second average might be best.

Right now getting some fairly quick progress should be good motivation.

my h and u perms (mu) are my fastest plls and coll also increases oll skip chance
What about your Z-perm? How much slower are you J-perm, T-perm and other fast adj PLLs? How much slower are your COLL from the fastest one in that set? How much slower is your COLL recognition? Are your other PLLs bad because you have neglected them? Yes PLLs skips are more likely but does it save you time to have them? The alternative alg and recognition time could cancel out the benifit of doing what would be more standard. Yes that isn't the case with PLL skips but the other this is not as simple with the other 11/12 cases.
 

WoowyBaby

Member
Joined
May 27, 2018
Messages
351
Location
United States
WCA
2018ISOM02
YouTube
https://www.youtube.com/channel/UCzgIKWl
I would have to say that working on solving your 2x2x2 blind (100% planned) and planning as much as the full 3x2x2 as possible in inspection is more important and will make much more of a difference than changing your last layer, even going from CO->PLL to ZBLL or something like that. Learning dozens or hundreds of LL algs, which takes a LOT of effort, could have the same improvement as just 10 or so minutes of block blind practice. Just my opinion though.
 
Last edited:
Joined
Oct 7, 2018
Messages
55
Ultimately do whatever is most fun or else you won't reach your goal. It depends on what you want to do with the absolute fastest improvement would mean laying off LL. Also in your case experimenting with block building is something you will have to do a lot to get to your goal.



Doing what is most fun for him is what he should do. But laying off LL until say 16 second average might be best.

Right now getting some fairly quick progress should be good motivation.


What about your Z-perm? How much slower are you J-perm, T-perm and other fast adj PLLs? How much slower are your COLL from the fastest one in that set? How much slower is your COLL recognition? Are your other PLLs bad because you have neglected them? Yes PLLs skips are more likely but does it save you time to have them? The alternative alg and recognition time could cancel out the benifit of doing what would be more standard. Yes that isn't the case with PLL skips but the other this is not as simple with the other 11/12 cases.
my h perms are 1 second u and j are 1.2 t and a are 1.3 1.4 z about 1.8
 
Joined
Feb 8, 2017
Messages
1,050
Location
Utah
WCA
2016BAIR01
I have a tip for you for OH.

Table abusing should help you achieve sub-1:00. Don’t know what table abusing is? Check out some of Max’s OH solves. I did it and went from averaging over 1:00 to sub-50. I also use a block building method(Roux) as my main method.
Or just learn how to turn decently at OH and learn a few algs?
 

tx789

Member
Joined
Sep 6, 2010
Messages
1,872
Location
New Zealand
WCA
2010HUNT02
YouTube
thetx789
my h perms are 1 second u and j are 1.2 t and a are 1.3 1.4 z about 1.8
Look at the times of all your PLLs and COLLs. Your Z-perms are quite a bit slower than the other EPLLs. If you do COLL you avoid a lot of good PLLs faster than Z-perms for you.
I have a tip for you for OH.

Table abusing should help you achieve sub-1:00. Don’t know what table abusing is? Check out some of Max’s OH solves. I did it and went from averaging over 1:00 to sub-50. I also use a block building method(Roux) as my main method.
There are videos on cubing world about Petrus OH. Also look at Kian Mansour's video about Roux OH. Max's OH isn't what most use and isn't better just because he uses it. He is an outlier.
 
Top