Possible 2011 WCA Regulations Changes

[hr.mohr]

I think he meant that 1/3 of people break this rule by accident and 2/3 don't, but that it's a silly rule and we shouldn't punish one of those two groups.

I know that. I was just pointing out how ridiculous it is to use random numbers as arguments for anything.

 

[DavidWoner]

There is no difference between ranking by sum and ranking by average to the thousandths place, I don't see why people seem to be distinguishing them.

 

[qqwref]

But why would it be appropriate to round to the 3 decimal places if we are only measuring the solves in 2. Unless we decide to change the entire ranking system by using 3 decimal places, this doesn't make sense to me.

You agree that it's right to keep the sum to two places, and not immediately round it to the nearest .03, right? So when we divide by the integer 3 (it's important that this is an exact number, there's no uncertainty in that one) keeping a 0, 3, or 7 in the thousandths place is just continuing to keep track of the same amount of information. The point about 3 being exact is that a sum-of-3 of 31.19 and a mean-of-3 of 10.397 are equally accurate as long as you know that the .xx7 is meant as one of 3 possibilities (not 10).

 

[maggot]

I completely agree on this. I feel that the flaws in the Stackmat timer are probably one of the most severe things that should be changed whenever is feasible to do so. Even if that means we can't do it this year, I still think it should be done whenever an opportunity presents itself.

I also think this will solve all the issues if they arise. What is appealing to the speedsolving community is probably even more appealing to the WSSA. To my knowledge, none of the events that they have take more than 1 minute (hence 10min dead time). Also, many of their events are much like 2x2 or magic, in which the variances of .xx is sometimes significant. I wonder if we were to request this to the Speedstacks company, since they are also well aware that we use the same timer, if they would be willing to make a new timer. I see the WSSA benefiting more than the WCA for .xxx or even .xxxx, but I'm sure adding 10:00+ would not be difficult either. Not only would it be welcomed with open arms by any sport to have a more accurate timer, but also to have a timer that would nearly eliminate the issue of idential times. Or, we could develop our own timers. I also wonder if some company (QJ?) would be interested in creating a timer for the cubing community if presented with a schematic or even better, a prototype.

 

[AvGalen]

@Dan: who would win this according to the WCA rules, and who should win according to logic?
Person that has the best 3/5; 1.01 1.33 1.33 .133 1.99
Person that has the fastest single; 1.00 1.34 1.33 1.33 1.99
So simply said: Base ranking on sum of total 3/5 first, and only when still equal on fastest solve

About 2x2x2 blind: I already stated that I think almost nobody can do 2x2x2 speedblind succesfully with getting a maximum of 1 DNF, especially because there is no inspection time. So no, 2x2x2 would not be the same as normal 2x2x2. And if somebody can do it, he deserves the WR.
@Michael: For single 2x2x2 I have a 1.xx single and 6.xx average. Should we stop doing 2x2x2 for that reason? I already stated that the single 2x2x2 WR is still held by the best and I expect the same for 2x2x2 single blind. It might be held by a lucky (guess) solve for a while, but so has the 2x2x2 single WR been. I have adressed all of this before, why are you giving arguments that have already been countered?

 

[masterofthebass]

The thing is, take those 2 averages and separate them. Looking at the average result of 1.33, you would consider that to be correct, wouldn't you? Now why should we add another number into the equation, when we are only looking at the final average result calculated when rounding the average of the counting times? When looking at an average result, the individual counting times technically do not matter, or even have to be considered as relevant. I would consider them more as a "fun fact" when looking through results. When you look at a world record, you don't think about the individual times necessarily (other than perhaps the fastest), so why should that change when comparing two averages next to each other.

 

[qqwref]

@Michael: For single 2x2x2 I have a 1.xx single and 6.xx average. Should we stop doing 2x2x2 for that reason? I already stated that the single 2x2x2 WR is still held by the best and I expect the same for 2x2x2 single blind. It might be held by a lucky (guess) solve for a while, but so has the 2x2x2 single WR been. I have adressed all of this before, why are you giving arguments that have already been countered?

I'm saying that 2BLD would have the same problem as 2x2 already does, but it would be *even worse*. And I don't think we need to augment problems like this. Maybe you get as low as 1/4 of your regular average on 2x2 speed, but on a similar scramble on 2BLD it is not hard at all to imagine someone getting 1/10 of their average. I might take 40 seconds to do a normal BLD solve, but if I immediately see a 4- or 5-move speed BLD solution I'm going to go for it.

EDIT:

The thing is, take those 2 averages and separate them. Looking at the average result of 1.33, you would consider that to be correct, wouldn't you? Now why should we add another number into the equation, when we are only looking at the final average result calculated when rounding the average of the counting times? When looking at an average result, the individual counting times technically do not matter, or even have to be considered as relevant. I would consider them more as a "fun fact" when looking through results. When you look at a world record, you don't think about the individual times necessarily (other than perhaps the fastest), so why should that change when comparing two averages next to each other.

What is the difference in saying that the difference in two means-of-three of 31.19/3 and 31.20/3 does not matter, and saying that the difference in two single times of 0.80 and 0.81 does not matter? If we can measure them both, they are both valid differences in time.

 

[Radu]

I completely agree on this. I feel that the flaws in the Stackmat timer are probably one of the most severe things that should be changed whenever is feasible to do so. Even if that means we can't do it this year, I still think it should be done whenever an opportunity presents itself.

I can't agree more on that! And the same goes for displays too...it's embarassing to use such an obsolete and heavy display, which uses 8 type D batteries in the year 2011, when we have such great technologies. One can make a better display at home with a few LEDs. I posted actually about this, here a few weeks ago: http://worldcubeassociation.org/forum/viewtopic.php?f=4&t=915&sid=294d8419bf03f3aee55955a8fa9358dc

Besides that, they are quite expensive for what they do... I just hope the Chinese industry will come quickly with something that will help us all.

 

[AvGalen]

I'm saying that 2BLD would have the same problem as 2x2 already does, but it would be *even worse*. And I don't think we need to augment problems like this. Maybe you get as low as 1/4 of your regular average on 2x2 speed, but on a similar scramble on 2BLD it is not hard at all to imagine someone getting 1/10 of their average. I might take 40 seconds to do a normal BLD solve, but if I immediately see a 4- or 5-move speed BLD solution I'm going to go for it.

EDIT:

What is the difference in saying that the difference in two means-of-three of 31.19/3 and 31.20/3 does not matter, and saying that the difference in two single times of 0.80 and 0.81 does not matter? If we can measure them both, they are both valid differences in time.

I can do a real 2x2x2 blind in about 45 seconds, but I have never done a speedblind in 11.25 seconds (1/4) let alone 4.5 second (1/1). You are forgetting that there is no inspection time for blindfolded, that a DNF is very likely and that predicting an entire solve is MUCH harder than seeing a skip and stopping the timer quickly? Did you ever see anyone do a speedblind on a 4 or 5 move scramble in real life? And even if all of these arguments of mine could be countered, that would still just mean that 2x2x2 blind is about average and that the single shouldn't be taken too seriously. It isn't a reason not to have the event.

As for the rounding problem: 3/5 (and other averages) should be COUNTED as the sum of the middle values, but DISPLAYED rounded. If the sum of the middle values is identical the fastest solve should count to determine the winner, but not if there is already a winner based on the counting solves. I really don't see a reason not to do this

 

[qqwref]

I can do a real 2x2x2 blind in about 45 seconds, but I have never done a speedblind in 11.25 seconds (1/4) let alone 4.5 second (1/1). You are forgetting that there is no inspection time for blindfolded, that a DNF is very likely and that predicting an entire solve is MUCH harder than seeing a skip and stopping the timer quickly? Did you ever see anyone do a speedblind on a 4 or 5 move scramble in real life?

I'm not forgetting anything - I've tried it. I tried some 4-5 move BLD solves earlier today and all of them were in the 4-9 second range (if I could solve it at all, which did happen most of the time). Same event, but an easy scramble makes a big difference. I can't speedblind hard scrambles, so if I tried a real BLD solve on most scrambles it would take about as long as you.

And even if all of these arguments of mine could be countered, that would still just mean that 2x2x2 blind is about average and that the single shouldn't be taken too seriously. It isn't a reason not to have the event.

Fair enough, but I disagree

 

[MaeLSTRoM]

On the subect of adding events, What do people feel about adding maybe the Gigaminx or Pyramorphix?

 

[Mike Hughey]

I'm not forgetting anything - I've tried it. I tried some 4-5 move BLD solves earlier today and all of them were in the 4-9 second range (if I could solve it at all, which did happen most of the time). Same event, but an easy scramble makes a big difference. I can't speedblind hard scrambles, so if I tried a real BLD solve on most scrambles it would take about as long as you.

If I recall correctly, Faz was practicing this a while ago (like before he broke any world records or something), and he could consistently do sub-10, with almost no DNFs ever. Well, except for the fact that he always did a +2 - he never bothered to figure out the AUF at the end. At the time, he would have had no problem at all doing averages of 5 successfully sub-10.

 

[AvGalen]

If I recall correctly, Faz was practicing this a while ago (like before he broke any world records or something), and he could consistently do sub-10, with almost no DNFs ever. Well, except for the fact that he always did a +2 - he never bothered to figure out the AUF at the end. At the time, he would have had no problem at all doing averages of 5 successfully sub-10.

Well, Faz IS the best of the best, and if 3x3x3 is about 30 seconds with a normal blindfolded method, I think sub 10 for 2x2x2 with a normal blindfolded method should also be possible (7 pieces instead of 20 is about 1/3).

And it would be interesting to determine a strategy:
-Spend first 2 seconds to see if speedblind is a good idea, then find a solution, then execute
-or just start with a regular blind method to save 2 seconds, but need more time for memo and execution
-or somehow even a hybrid between speedblind and regular blind?

Michael, your test seems to not take into account that you HAVE to get 4 out of 5 solves correct AND that you don't know if the scramble can be speedsolved. Also, 4 or 5 move solutions are RARE. I have never encountered 1 while using the WCA 2x2x2 optimal scrambler

 

[Pedro]

On the subect of adding events, What do people feel about adding maybe the Gigaminx or Pyramorphix?

Gigaminx takes too long. And I never really saw someone practicing pyramorphix.

 

[PhillipEspinoza]

Inspecting with hands for 3x3 with Feet should be allowed. If you can inspect with 2 hands for one-handed, then you should be allowed to inspect with hands for Feet.

For OH, I propose that the tabling/surfacing move be made illegal or limited to a certain amount (maybe 3) per solve with any more such moves being counted negatively towards the competitor's time.

 

[keemy]

For OH, I propose that the tabling/surfacing move be made illegal or limited to a certain amount (maybe 3) per solve with any more such moves being counted negatively towards the competitor's time.

no

 

[AvGalen]

Inspecting with hands for 3x3 with Feet should be allowed. If you can inspect with 2 hands for one-handed, then you should be allowed to inspect with hands for Feet.

For OH, I propose that the tabling/surfacing move be made illegal or limited to a certain amount (maybe 3) per solve with any more such moves being counted negatively towards the competitor's time.

yes

I agree about feet. This just makes sense from a consistency point of view and wouldn't be difficult or unfair to change.
I don't know why the table/mat is allowed to be used, but we should either allow it or prohibit it. Because prohibiting seems impossible (a cube could fall out of the hands and then has to be picked up) I would say: Don't change this

 

[Mike Hughey]

Well, Faz IS the best of the best, and if 3x3x3 is about 30 seconds with a normal blindfolded method, I think sub 10 for 2x2x2 with a normal blindfolded method should also be possible (7 pieces instead of 20 is about 1/3).

Considering I can go sub-20 on a good solve and sub-25 on a good average, it seems likely that sub-10 for someone who's actually fast is very likely with a normal blindfolded method, yes. I do think it would be an interesting event. I sympathize with qq's perspective about singles being meaningless, but the averages really would be pretty interesting. It would be kind of fun to have an event where two such very different strategies were often competing to find a winner.

 

[qqwref]

Michael, your test seems to not take into account that you HAVE to get 4 out of 5 solves correct AND that you don't know if the scramble can be speedsolved. Also, 4 or 5 move solutions are RARE. I have never encountered 1 while using the WCA 2x2x2 optimal scrambler

1) Why do you think I would be trying for an average in competition? Considering I'm bad at BLD already (enough that I doubt I could consistently get 4+/5 with normal 2x2 BLD solves) I would certainly be trying for the best single solve I could get.
2) 4 or 5 move solutions are common enough that many people have encountered them in 2x2 speed rounds...

 

[Stefan]

Was curious and did some calculations (hope they're right):
- One in 300 positions for 2x2x2 can be solved in five or fewer moves.
- One in 9095 averages of five has more than one such position.
- There have been 845 rounds of 2x2x2 in WCA competitions (819 were average of five).