[WCA Regulations 2012] Remove +2 penalty for misaligned sides

[keyan]

If a person cares not to get a DNF then they should be more careful with their turning and choose more reliable puzzles.

Isn't that what we've all been saying in favor of removing +2?

Don't you guys understand? You are being way too technical. No duh they intended to solve it. Why go to a comp to not solve it? The solve hundreds to thousands of times and you are questioning whether or not they actually even know how to align a misaligned U face?
...
If I went down the street with the cube looking like this. And asked every single person, regardless of age, if they were capable of solving it. There would would be none who couldn't do it. (Besides infants, Mentally impaired, and few other exceptions. Obviously we won't go into that.) Any person in their right mind could do it.

In general I think it's best to not respond to such a rambling post, but with regard to the two points I've quoted, I want you to explain first how my U perm example (post #161) is unacceptable given your logic here, and second why you aren't arguing to consider as "solved with penalty" all "trivially solvable" cases (defined as three or fewer moves from solved by Stefan, or with your own definition).

Keyan's numbers strictly show the number of penalties that were awarded, nothing more, stop trying to use them to determine the effects of a regulation change.

I think the numbers tell quite a lot. It says to me that your argument that this will change the approach cubers take in competition is making a mountain out of a molehill. In the current environment, where we don't have "harsh punishments" and cubers are supposedly free to not worry about not solving their cube, +2s are a very rare event. Were the rule to change but competitors weren't informed, and so went ahead competing with their current mindset, almost nothing would change. The impact would be tiny.

I'll respond to the 'metagame' bit after I get a chance to compare some competitor's previous averages.

 

[Escher]

I think the numbers tell quite a lot. It says to me that your argument that this will change the approach cubers take in competition is making a mountain out of a molehill. In the current environment, where we don't have "harsh punishments" and cubers are supposedly free to not worry about not solving their cube, +2s are a very rare event. Were the rule to change but competitors weren't informed, and so went ahead competing with their current mindset, almost nothing would change. The impact would be tiny.

I'll respond to the 'metagame' bit after I get a chance to compare some competitor's previous averages.

"[the] numbers strictly show the number of penalties that were awarded, nothing more"
The statement above is completely true. It's impossible to derive the conclusions you've reached.

Also, you're taking a (small compared to the number of solves in the database) sample from every single event where a +2 can be awarded, not where the +2 meta is relevant.

If you didn't tell competitors, and awarded a DNF in a plus2 situation for a fast 2x2er's first solve I can guarantee you'd completely mind**** them and their average. There are plenty of conceivable occurrences whereby some other error is committed early in an average and the rest is damaged due to having to be extremely careful of not overshooting or putting the puzzle down too hard. In an ideal world nobody would make these mistakes, but they do, and they have been doing for the past 4-5 (?) years. The onus is on you to demonstrate to the community that removing +2 wouldn't be damaging to top competitor mindset, or that the damage is minimal enough to warrant removal for the sake of your argument from simplicity.

I don't want to sound elitist but it is the elite that are setting the records and getting into news media and expanding the reach of cubing. It's for the boundary-pushing competitors where every single miniscule reaction and turn makes a huge difference, and it's for these whom the meta change 'would' affect the most.

Maybe it's just me, but in my experience +2s are a regular occurrence at UK competitions, and certainly weren't rare at US Nats or CO. If you want to start making conclusions about meta change, then a separate poll needs to be set-up where a reasonable proportion of the community votes.

 

[Pedro]

Huh, why? One more move and it's a different topic? O_o

No, what I meant is that 2 moves away is a DNF, no matter the rule changing or not, and filtering scrambles is another thread/discussion.

 

[FatBoyXPC]

Rowan: Thank you for clarifying my point. You saw what I was attempting to say and made it much clearer.

Dene: I agree with you 100% about "people seem to not know the regulations that well, including those you'd expect to know them." I do disagree with you on this same point, though. The only competitors I see that do not know the +2/DNF rules are new competitors. That being said: Most (my opinion, obviously) competitors with more than one competition on their WCA profile page will have a clear enough (relative, I know) understanding of the difference between a +2 and DNF (at least when it comes to misalignments). Rowan is completely right: Take a competitor who might not have seen the regulations update (say they were without the ability to see the rule change(s) and the delegate/organizer at their next competition doesn't mention a ruleset update) and put them in a 2x2 event, they will have a significantly worse average than usual if they get a DNF instead a +2 on their first solve. This could very well be true for 3x3 solves. The truth is knowing that you cannot afford another DNF will affect your times. I've already seen impedance created from one DNF or +2 in a 2x2 round with some competitors (to be fair they aren't a sub3 average, but still).

I agree with Rowan on his point about the stopping procedure being more time of the solve. Large cubes obviously won't matter near as much to take that extra half second to verify you are aligned correctly. Look at the current 2x2 WR: an extra 0.5 second would obviously blow this time (and I believe I already mentioned this).

 

[Dene]

Now that you've speculated, I thought I would check some facts. I don't have any programming skills so I can't do any sort of thorough analysis, but I decided to do a quick look over the profiles of Feliks Zemdegs and Rowe Hessler, and Michal Pleskowicz, the top three cubers in 3x3 speedsolve at the 2011 World Championships, and also acknowledged as among the best cubers in the world. There weren't a lot of examples to go by, but I will post some of their averages that included DNFs to see how destroyed they were. Except they weren't destroyed at all:

Feliks Zemdegs:
3x3: 7.78, DNF, 11.27, 6.77, 7.58
Average: 8.88. WR single, 6.77
NOTE: The DNF was the second solve in the average.
Not only did he get what was a reasonable average for him, certainly not considerably better or worse than anything else he was getting (at the same competition his other averages were 9.07, and 7.91, which became the WR average), but two solves later he went and got the new WR single! This, of course, happened just after he had beaten Erik's long-standing 7.08 WR earlier in the competition.

4x4: 38.28, 40.91, DNF, 36.84, 37.00
Average: 38.73
In this case the DNF came in the middle of the average, and after that he proceeded to get the best two times of the average. The average was not the best because of the solves that came before the DNF, not after.

2x2: 3.31, 2.80, DNF, 1.61, 2.93
Average: 3.01
A normal average for him at the time (which, incidentally, was followed in the final round by the OcR), Feliks once again proceeded to get his best solve after the DNF.

Megaminx: 1:02.86, DNF, 1:05.47, 1:09.80, 1:02.58
Average: 1:06.04
Even though megaminx isn't one of his main events, I thought I'd put this one in there too for the sake of it. This is the current OcR.

Also from Feliks, the old OcR in Square-1 had a DNF on the second solve. The current OcR for OH has a DNF on the final solve (this doesn't show much re. the meta thing, but still shows that a DNF is not necessarily devastating).

Rowe Hessler:
3x3: DNF, 9.96, 9.44, 10.40, 9.56
Average: 9.97
A brilliant example of why all this "meta" talk is a load of crap. Rowe's most recent attempt, a DNF on the first solve, followed by a decent sub10 average.

3x3: 11.13, 8.88, DNF, 9.44, 8.38
Average: 9.82
Just another example of a fairly recent average. This could have been significantly better if not for the 11.13 on the first solve, which is unfortunate. I'd be interested to know if that was a +2 or not.

4x4: 41.66, 39.91, DNF, 44.59, 36.40
Average: 42.05
Rowe's PB 4x4 average in competition, the best solve comes after the DNF, once again.

2x2: I won't pick out any specific cases, but if you look through all his averages there are a few with DNFs in them and they are mostly normal averages, and certainly no exceptional cases of a DNF completely destroying an average except one occurrence of two DNFs. But this only shows to me that he obviously didn't care enough to be careful after the first DNF (which is what you meta-argument defending people seem to think would happen, that is, people would slow down a lot and be more careful after an early DNF).

OH: 13.03, DNF, 13.86, 18.19, 16.05
Average: 16.03
Just one of a few examples from Rowe's OH averages. This is once again a normal average for him in comparison to other competition averages. The DNF is followed by a very fast 13.86.

Michal Pleskowicz:
Our current World Champion is a good example of how you can avoid DNFs if you put your mind to it. He has a very good record regarding these. Unfortunately there is no information on his +2 penalty history for misaligned slices. I'll just point out a couple of averages that show that DNFs do not destroy averages.
3x3: 10.21, 8.33, 9.93, 9.68, DNF
Average: 9.94

And an old average of his from 2009, which was his PB at the time:
3x3: DNF, 15.09, 14.38, 17.27, 14.90
Average: 15.75


Because all of this was such a successful endeavour I thought I would have a quick check of the averages of two previous World Champions and also highly respected cubers.

Breandan Vallance:
3x3: DNF, 9.28, 10.13, 9.21, 9.27
Average: 9.56
A recent average for him, and the best he had done in quite a few months.

3x3: 9.00, DNF, 8.88, 9.46, 10.77
Average: 9.74
Done at WCs

3x3: 9.33, DNF, 9.21, 9.78, 10.63
Average: 9.91
The winning average at USNats 2011.

Yu Nakajima:
Another example of how you can solve the cube when you put your mind to it, not a single 3x3 DNF!
4x4: DNF, 1:08.68, 48.44, 48.41, 46.06
Average: 55.18
This is the only case I have seen of a DNF "destroying" an average, except my guess is that it was actually a big old POP that destroyed the average, and not his meta, seeing as he proceeded to get what were good times for him after that.

5x5: DNF, 1:23.09, 1:29.05, 1:19.11, 1:16.41
Average: 1:23.75
The NR at the time.


So what can we conclude from all this? Firstly it should be made clear that we cannot conclude that changing a +2 to a DNF will have no affect on cubers. All of this had no information on +2, nor on how changing the rule would affect the meta of elite cubers. However what it does show is that elite cubers are not concerned in the least by getting a DNF. Therefore I think we have good reason to believe that the meta of cubers would not change significantly in the face of a change of rule, as penalties in general are a rare occurrence (thanks to the data provided by keyan), and even when a penalty is incurred the elite cubers simply don't let it affect them.

 

[bobthegiraffemonkey]

Dene, I think your inference is rather flawed. You have picked out data which appears to suit your argument, but which doesn't include why the solves were DNFs or how easy the solves were that were fast after the DNFs. There is also the fact that there is naturally some variation in solve time, so it is hard to tell what the solve times 'should' be to compare against what happened after the DNF. Logically, you can conclude pretty much nothing from the data you posted.

 

[Robocopter87]

The same is true for cubes off by M2 which is a DNF. So what's your point? We consider a cube "almost solved" (aka +2) if the average joe can solve it?

The M2 thing is not part of this discussion. However I don't see why this is only for the U face. Anywho, no, thats not my point. He asked me how we knew they intended to solve it. he asked me how I knew he was capable of solving it. The point of the going down the street was to show that a cuber is obviously capable of aligning that face. Except in +2 scenarios, for whatever reason, the cuber failed to finish the revolution.

 

[Dene]

Dene, I think your inference is rather flawed. You have picked out data which appears to suit your argument, but which doesn't include why the solves were DNFs or how easy the solves were that were fast after the DNFs. There is also the fact that there is naturally some variation in solve time, so it is hard to tell what the solve times 'should' be to compare against what happened after the DNF. Logically, you can conclude pretty much nothing from the data you posted.

I picked out *every* relevant average with a DNF, except for in the cases where I mentioned. I didn't pick out specific cases to suit my argument. If you don't believe me go and check their WCA profiles.

As for the rest of the stuff you said, that applies to any average, therefore generalises across all situations and all rules. If that applies to past averages with the +2 rule it would apply to future averages with the DNF rule, therefore it is completely irrelevant. Logically, I can conclude exactly what I did from the data I provided.

 

[PhillipEspinoza]

Feliks Zemdegs:
3x3: 7.78, DNF, 11.27, 6.77, 7.58
Average: 8.88. WR single, 6.77
NOTE: The DNF was the second solve in the average.
Not only did he get what was a reasonable average for him, certainly not considerably better or worse than anything else he was getting (at the same competition his other averages were 9.07, and 7.91, which became the WR average), but two solves later he went and got the new WR single! This, of course, happened just after he had beaten Erik's long-standing 7.08 WR earlier in the competition.

This is a good point in seeing whether or not DNF affects the meta of top cubers, but this is pre-+2-regulation change. Wouldn't the regulation change essentially double the chance of a DNF? And wouldn't knowing that affect your meta differently? I dunno if the data provided would really be indicative of how meta would be affected if the +2 regulation change were to go into effect. It's kinda like changing it to a mo3, where you would in fact be more afraid to DNF because that would ruin your average, then using avg of 5 to show how DNF doesn't effect meta.

On another note, imagine if that 11.27 was a +2, under the rule change it would DNF the average. The real question is how many good averages would turn into DNF averages given the proposed regulation change? An example (not so good average but w/e) in AZ Showdown this was the winning average.

10.86, 11.31, (41.47), 13.03+, 10.21 = 11.73 (sucky I know)

Had it been under the regulation change, this average would've went from winning the competition to a 21.21 avg, all because it was 1° off for the 11.03 solve. How many good or winning averages would this new rule have affected? I'm too lazy to research it myself, DNF me I guess.

Point: It seems like all the +2 regulation change would do is increase the chances of a competitor getting a DNF for some arbitrary measure of "laziness"? (At 46° you're lazy and deserve a DNF but at 45° you're okay). It is therefore meaningless to use old results to show how DNF will not affect a top cubers meta/performance under the new regulation change.

 

[FatBoyXPC]

Dene: I was reading your post and was about to reply, but Phillip said pretty much what I wanted to say. Since all of your data is pre +2 -> DNF, this only shows that their DNFs are due to not performing algorithms correctly (I'd speculate this is probably the case for the 3 cubers you mentioned in 3x3, at least), most likely pops for 2x2, and I will go ahead and speculate that it's probably less relevant in 4x4 (especially the longer you go down the line as the 4x4 curve at least appears steeper than 3x3).

The whole point of the argument is that getting a +2 instead of a DNF means you have to be even more careful than you are now, and you cannot account for that in the current data set (no matter how much you try to display said results). You already know that we don't have a way to see what solves were a +2, and more specifically which +2's were misaligned layers vs other +2 penalties.

I can tell you this right now: I'm quite the nervous competitor, especially when I'm averaging better at home than what I did at the previous competition, but the only thing a single DNF does to me is irritate me because that means any +2's I get will raise my average. I get annoyed at me being silly enough to get a +2, but it doesn't really affect the "metagame" that much. I am interested in your +2 -> DNF experiment though, and will do an average of 50 and post you the results in IRC or something. I don't think this thread is the place for everybody posting experiment results. I would like to see this experiment conducted with a lot more people, though.

I could probably argue with your logic about +2's that we should never have a +2, including time penalties, timer/hand penalties, etc. You might already be on this path though, and I've missed it since I'm looking at this specifically in regards of misaligned layers vs +2's altogether.

 

[Dene]

Basically everything you two have said is irrelevant. The examples I gave were directly to do with how getting a DNF affects the meta of cubers, causing it to destroy their averages. The data clearly shows that the meta of cubers is not affected after a DNF, and no one has provided any reason why that should be different if we give out a DNF for a misalignment instead of a +2. For example, if that 11.27 solve by Feliks was a +2 (it wouldn't be hard to find that out, I'm sure there's a video), if he knew he would DNF the average he probably would have gone to the extra effort to fix the misalignment before slamming down on the timer, which would have cost him less than 2 seconds, and made the average better.

Also, I don't recommend doing my experiment, it was more of a thought experiment than anything else, and I'm pretty sure I got the point across. Whether you agree with that or not, I don't know (probably not, tbh ).

Phillip: As the regulations currently stand the difference between no penalty and a +2 is between 45 degrees and 46 degrees. The difference between a +2 and a DNF is between 45 degrees and 46 degrees. What's your point?

 

[FatBoyXPC]

Dene, I was watching a tv show after I posted that and realized that doing 50 solves here at home won't change my meta (but it should show how often I'd +2/DNF). Doing solves at home I care far less than solves in competition, and I am certain I'm not the only one who is like this. I am also certain that I'd be much more careful about my +2's (that would now be DNFs) in competition, because I have a hefty amount of +2's in competition. At either UIndy Open or Dayton the month before that, I had a 3x3 round where I had 3 +2's and a DNF. I realize that I'm nowhere near "good" by any means, but it doesn't change the fact that my times would have been slower had it been DNF's under my current mind set (and that my times would all be increased by at least 0.5 by ensuring I place the cube correctly and have it aligned properly).

With 2x2 (for me) it's even worse. I for sure slow down when I get a DNF in 2x2, for fear of popping (most likely the cause of the first DNF), so it adds easily 1s to my time (I'm slow at 2x2, sue me).

 

[Cool Frog]

Basically everything you two have said is irrelevant. The examples I gave were directly to do with how getting a DNF affects the meta of TOP cubers, causing it to destroy their averages. The data clearly shows that the meta of TOP cubers is not affected after a DNF, and no one has provided any reason why that should be different if we give out a DNF for a misalignment instead of a +2. For example, if that 11.27 solve by Feliks was a +2 (it wouldn't be hard to find that out, I'm sure there's a video), if he knew he would DNF the average he probably would have gone to the extra effort to fix the misalignment before slamming down on the timer, which would have cost him less than 2 seconds, and made the average better.

As with the last statement that is a pure assumption.
You can't generalize all cubers into the Top cuber group.

 

[Dene]

As with the last statement that is a pure assumption.
You can't generalize all cubers into the Top cuber group.

Don't look at me, Escher was the one that thinks the top cubers are more important than the rest of us, not me. Do please try and keep up if you're going to post. And of course it's an assumption. Almost every argument that everyone is making is based solely on assumptions and nothing else. I mean, nobody has any real data to base their claims on, do they?

@fatboyxpc maybe you should be more careful and solve your cubes properly to save yourself all of those +2s.

 

[FatBoyXPC]

@fatboyxpc maybe you should be more careful and solve your cubes properly to save yourself all of those +2s.

If only you knew how many +2's I had at home I have some real bad habits I need to break

 

[jincronics_07]

i dont agree what if you are solving 4x4 and 5x5 +2 doest matters unlike for 2x2 and 3x3.

 

[Escher]

Don't look at me, Escher was the one that thinks the top cubers are more important than the rest of us, not me.

Lolz I never said that.

 

[keyan]

"[the] numbers strictly show the number of penalties that were awarded, nothing more"
The statement above is completely true. It's impossible to derive the conclusions you've reached.

Your statement is of course true. However, numbers mean nothing without analysis. Would you like me to add a disclaimer to my post: "I am not omniscient, all interpretation provided reflect my own views, and are not endorsed by SpeedSolving.com or its sponsors."
I'm curious, what's your interpretation of those numbers?

If you didn't tell competitors, and awarded a DNF in a plus2 situation for a fast 2x2er's first solve I can guarantee you'd completely mind**** them and their average.

That's not what I was saying. Hold a competition exactly as we do now, but simply record +2s as DNF. I expect almost nothing would change.

The onus is on you to demonstrate to the community that removing +2 wouldn't be damaging to top competitor mindset, or that the damage is minimal enough to warrant removal for the sake of your argument from simplicity.

Or is the burden on your side, to prove that this psychological boogeyman does exist?

I don't want to sound elitist but it is the elite that are setting the records and getting into news media and expanding the reach of cubing. It's for the boundary-pushing competitors where every single miniscule reaction and turn makes a huge difference, and it's for these whom the meta change 'would' affect the most.

I remember in previous discussions of this issue, some said that removing this rule would only hurt slower cubers. That such a change would 'obviously' have opposite results is interesting.

I'm glad that I quoted Phillip's post before it was edited. Is the little 'edited by' note at the bottom automatic, or can that be changed as well? Who has authority to edit other people's posts?

 

[Godmil]

The examples I gave were directly to do with how getting a DNF affects the meta of cubers, causing it to destroy their averages. The data clearly shows that the meta of cubers is not affected after a DNF,

Hahaha. Your examples make no sense whatsoever. How do you know those solves wouldn't have been a little faster if they didn't have a DNF? Just cause a time is still quite good doesn't mean it is as good as it could have been. You would need some large averages to prove the point.

 

[Pedro]

Phillip: As the regulations currently stand the difference between no penalty and a +2 is between 45 degrees and 46 degrees. The difference between a +2 and a DNF is between 45 degrees and 46 degrees. What's your point?

The point is that the difference between no penalty and DNF would be between 45 and 46 degrees, with your proposal.

And your examples are flawed. "Top cubers are not affected at all by a DNF". Well, maybe that's one of the reasons why they are top cubers...and how do you know the rule change will affect them more or not? We can't know.

As I said before, this debate is going nowhere. I'm wondering what the WRC action will be. Since there is no agreement here, the rule shouldn't be changed. But how exactly will WRC do this? What if the people in there are in favour of changing the rule?