Proposal - Average of 5 for 6x6

[uberCuber]

I am proposing just what the title says -- that we change the format of the 6x6 event from Mean of 3 to Average of 5. **EDIT: My preferred idea is now to have both mo3/avg5 available, with sufficiently fast mo3's allowing competitors to continue to an avg5, while "slower" competitors do a mo3 as before. See the following posts.**

My motivation for proposing this change is probably clear, but I'll spell it out anyway. Mean of 3 is a format that has an inherent, frustrating issue, which is of course the fact that there is absolutely no room for accident or mistake. Cube pops one time? Average ruined. Make one bad mistake in one solve? Average ruined. Before someone else points it out, yes, as it happens, every 6x6 (and 7x7, but that's another story) mean I have done in competition has been ruined by a single terrible solve, so I certainly have a personal bias in this, but so do plenty of others. And besides, why would I be proposing it if I didn't have reason to care myself?

Let's take a quick look at a bit of data.

5x5 used to be a mean of 3 event. Looking at past results in the WCA database tells me that it was the beginning of 2006 when the event format was changed to Average of 5. Let's look at what the 5x5 results page looked like just before this change was made.

- The world record single was 1:51.41; the world record average was 2:06.15.
- The single at the 50% mark was 4:44.49; the average at the 50% mark was 4:31.37.

Now let's look at what the 6x6 results look like right now.

- The world record single is 1:40.86; the world record average is 1:51.30.
- The single at the 50% mark is 4:22.73; the average at the 50% mark is 4:10.44.

If the times listed above were apparently fast enough to switch 5x5 to an Average of 5 event, why can't 6x6 be given the same designation now?

 

[Kirjava]

the times at the 50% mark may be a certain thing, but time variance is way way higher.

 

[Hays]

I am in complete support of this idea. The mean of 3 format for 6x6 drastically changes the mindset and strategy I have to have when competing at competitions.

For example, I didn't just happen to get a 5 second better MO3 and 11 second better single at Vancouver Open than at worlds, it was because there was no real consequences if I popped and ruined the average. For me, at competitions where the main goal is to place, rather than get good times, it is a much better idea to turn and solve a little slower in order to reduce the risk of popping. The fact that one pop can ruin an entire average forces me and I'm sure others to have slower overall times when competing at events where placing is the top priority. And while it could be said that going slower in order to have a safe average is part of the "strategy" of cubing, I think that the entire point of competitions and cubing in general is to have people achieve their best times possible.

That being said, I know there is a large time issue when changing 6x6 to MO3, since it will take significantly longer to do 5 solves instead of 3. Would it be a possibility to have those under, lets say the 3:00 cutoff to complete an average of 5, while everyone else does a mean of 3? If someone that does a mean of 3 finishes in under 3:00, he would simply do an average of 5 at the next competition he attends. I know this would be very strange considering that there has never been a split standard for an event before, but just an idea.

And also, mean of 3 sucks.

 

[uberCuber]

the times at the 50% mark may be a certain thing, but time variance is way way higher.

Yeah, you can probably tell from the choice of data I presented that I haven't taken a statistics class. I noticed that for each number of minutes from 4 to 10, the percentage of current 6x6 results below that barrier is greater than the percentage of 5x5 results below that barrier at the end of 2005, but I don't have the knowledge base to draw any other type of conclusion. I wonder if someone here can briefly make it clear to me what it means when you mention the variance of these sets?

That being said, I know there is a large time issue when changing 6x6 to MO3, since it will take significantly longer to do 5 solves instead of 3. Would it be a possibility to have those under, lets say the 3:00 cutoff to complete an average of 5, while everyone else does a mean of 3? If someone that does a mean of 3 finishes in under 3:00, he would simply do an average of 5 at the next competition he attends. I know this would be very strange considering that there has never been a split standard for an event before, but just an idea.

I see no reason why this couldn't be done. The fact that changing 5x5 from mo3 to avg5 didn't "reset" the rankings shows that there is no problem with means of 3 and averages of 5 coexisting in the same ranking. In fact, that 2:06.15 5x5 MO3 world record from 2005 is still standing in the rankings (now at 1075th ).

 

[MaeLSTRoM]

So I agree and disagree with some stuff in your post, so I'll just break it down and give you what I think on each bit.

I am proposing just what the title says -- that we change the format of the 6x6 event from Mean of 3 to Average of 5.

Ok so first thing, I agree with you on this, although I do think that it shouldn't happen. Allow me to explain why.

My motivation for proposing this change is probably clear, but I'll spell it out anyway. Mean of 3 is a format that has an inherent, frustrating issue, which is of course the fact that there is absolutely no room for accident or mistake. Cube pops one time? Average ruined. Make one bad mistake in one solve? Average ruined. Before someone else points it out, yes, as it happens, every 6x6 (and 7x7, but that's another story) mean I have done in competition has been ruined by a single terrible solve, so I certainly have a personal bias in this, but so do plenty of others. And besides, why would I be proposing it if I didn't have reason to care myself?

Yes, Mean of 3 is a format that does have its challenges. I agree that it can be frustrating having your average ruined by a bad time, although I also think you can save it with a good one. But when it comes down to choice, I would prefer to move away from Mo3 as a format, but I don't think there is the demand for Ao5 just yet.

Let's take a quick look at a bit of data.

5x5 used to be a mean of 3 event. Looking at past results in the WCA database tells me that it was the beginning of 2006 when the event format was changed to Average of 5. Let's look at what the 5x5 results page looked like just before this change was made.

- The world record single was 1:51.41; the world record average was 2:06.15.
- The single at the 50% mark was 4:44.49; the average at the 50% mark was 4:31.37.

Now let's look at what the 6x6 results look like right now.

- The world record single is 1:40.86; the world record average is 1:51.30.
- The single at the 50% mark is 4:22.73; the average at the 50% mark is 4:10.44.

If the times listed above were apparently fast enough to switch 5x5 to an Average of 5 event, why can't 6x6 be given the same designation now?

I would answer this with just one word: Potential.
This change was made 7 years ago. There weren't many choices of cube for 5x5 and the event was fairly new. Given that it was so long ago, it's easy to imagine that looking at a solve from that time, you could see that it is much slower than todays 6x6 solves, both in terms of TPS and in terms of solve efficiency. I would say that 6x6 is approaching a limit in terms of speed, and that its unlikely to drop to near 5x5 times soon. Also, although its been mentioned by Kirjava already, it would take up too much time. Even with fairly harsh cuts, the time taken to do 5 solves is still fairly long. If we assume that at least 1 person is at that 50% mark or higher, then they would have 9 -> 25 minutes of solving time, depending on cutoffs, whereas now they would have 5->15. Adding that amount of time can affect the run time of the round significantly.

However, I still agree with the change, so I will suggest this as a possible, if unfeasible option:

Allow organisers to have 2 cutoffs for rounds. So a combined round, which now has the format Bo1/Mo3, would become Bo1/Mo3/Ao5. Allowing the numbers to be cut even further like this would help to stop the time overflow problem, and also encourage organisers to still allow people to get averages, even on less harsh cutoffs because the option for Mo3 is still there.
The problem of course with this is that you could get a better Mo3 than Ao5, and the rankings would be mixed together, which could cause some conflicts in how the results are managed.

 

[Nathan Dwyer]

Justin, you say that the 6x6 times now are comparable to the 5x5 times when the format was changed, but you have to remember that at the biggest competition in 2006, there were only 17 competitors in 5x5. There are sooooo many more 6x6 competitors now that it's not really a valid argument for changing it since the 6x6 event would take so much longer.

 

[uberCuber]

@scrambling, the first thing to come to mind:

Forget my references to 2005/06 for the moment. As you pointed out, there are quite a few 6x6 competitors nowadays; of course the same is true of 5x5, which also can take up quite a bit of time. Compare 5x5 vs 6x6 right now. At worlds, ~945 5x5 scrambles were done in the first round. For 6x6, 88 people made the cutoff with their first solve, 83 did not. Keeping those numbers with the avg5 format would give 88*5+83*2 = 606 scrambles. Of course, it's likely that several people who barely missed cutoff could have made it with a second solve, let's just arbitrarily say the 8 people whose one solve was sub-4:10 make cutoff with a second solve, so add the extra 8*3 solves = total of 630 scrambles.

A 6x6 scramble takes (very roughly) twice as long to perform as a 5x5 scramble.* So then that 630 6x6 scrambles should take a similar amount of time as 1260 5x5 scrambles, clearly more than would have been done at worlds if 6x6 were a plain average of 5 event. But, taking the idea suggested by Hays/Maelstrom of combining both mo3 and avg5, this number is dropped to be more reasonable. Pulling the number suggested by Hays, let's say that anyone who got a sub-3 mean with their first three solves gets to do the last two solves for the avg5, while the cutoff for finishing the mo3 remains 4 minutes. 24 people got a sub-3 mean, 64 people got a mean that was not sub-3, and then still the 83 with only one solve. Then with this proposal that would be 24*5+64*3+83 = 395 scrambles performed. Compared to the 945 5x5 scrambles performed, 395 6x6 scrambles would not take up more time, especially since then the majority of scrambles are being given to faster cubers, who are more likely to have better 6x6's that won't be so difficult to scramble. EDIT: I should point out here that putting this proposal into practice would have only added an extra 48 scrambles on top of what was already performed at worlds, not a hugely drastic change.


*With my own main cubes, 5x5 takes ~20 seconds to scramble and 6x6 takes ~35. I also did a scramble on my awful unlubed, unmodded v-cube to test and it took 50 seconds. Since both my 5x5 and 6x6 are better than the average cube, I'll assume it would take closer to 25 seconds for me to scramble an average 5x5, and closer to 45 for me to scramble an average 6x6 (my 6x6 is way better than the average 6x6; I actually like it better than many 5x5's). Of course there is also the factor with scramblers who don't practice bigger cubes who'll find the 3-layer turns to make 6x6 scrambles inherently more difficult for them than 5x5, and so on. I'll say that my "twice as long" estimate seems as good as I can get right now.

The problem of course with this is that you could get a better Mo3 than Ao5, and the rankings would be mixed together, which could cause some conflicts in how the results are managed.

I'm wondering how much conflict it would actually cause, though. Remember, it would only be slower cubers doing means of 3, while all of the fastest would be doing averages of 5. Means of 3 would gradually get sifted out of the top rankings, and the slower ranks would still be filled exclusively with means of 3. It could take time before the divide between mo3's and avg5's in the rankings became pretty solid, but even for 5x5, two full years after the shift was made, there was still a small handful of relatively decent means of 3 standing in the ranks.

Oh yeah, and my PB avg5 is actually lower than my PB mo3 no I'm not presenting this as a serious argument don't take it as one thx

Feet has a better case for average of 5 if it's staying.

Lol I didn't even know feet was a mo3 event. Someone who cares about the event go ahead and ask for it to be changed, too. :s

 

[Sa967St]

Thread moved to Constructive Speedcubing Discussion. If the discussion gets out of hand, we'll move it back to Private Forum. Have fun.

 

[PeelingStickers]

Why not simply have a cutoff of one solve, as it already is. If you fail to get a solve below say: 4 minutes then you obviously cannot continue. Those that qualify can do the ao5.

 

[Benyó]

i support the avg 5 format. i don't think wca changed 5x5 from mo3 to avg5 because it had potential. wrs always seem unbeatable for most of the people, which means noone was expecting that in 4-5 years the good times would be less than the half of the 2005 wr. nobody has mentioned one of the main reason why should 6x6 become avg5: it has parities. and parity means that in a close race most of the time the luckier will win. that was the problem with square-1 too years ago. avg5 reduces the luck-factor quite well (like on 4x4 i had about 2 'lucky' averages in about 50 comps, while on 6x6 my means are 'lucky' in 20%, normal in 60 and 'unlucky' in 20% where lucky means i have only about 2 parities and at least one of them is not oll, and unlucky means the opposite).
i think we shouldn't mess things up with the mean-for-the-slower-idea. it is really simple: who wants to have an average, just practice. sub3 can be reached by anyone, even sub2:40 i think.
and don't forget that slow bld-solvers still have 2x10 minutes on most of the competitions like many years ago however the average results halved and top 3x3 blders can do a whole round in 3 minutes including scrambles. maybe we should take time away from slow cubers in order to give more attempts to the fasts, since competitive cubing isn't for people who have just learned how to solve the cubes a day before, but for cubers who are skilled at least a bit.

 

[Dene]

I'm not convinced about the arguments for changing to average of 5, but I don't see any good reasons why not to. Certainly the extra time taken up is not a good reason to remain mean of 3; you cannot tell me multibld takes less time. And what about the average of 3 for FMC that is being proposed? That's 3 hours of a schedule taken up.

 

[tx789]

the only mean of events left are 6x6,7x7 and feet. 6x6 and 7x7 take a long time and feet most people hate. You change them to keep the events with averages uniform. But then most comps wouldn't hold 6x6 or 7x7 or have a 4:00 cut off for 6x6 and 5:00 for 7x7 or something. And feet well...

 

[Benyó]

i was talking about 3bld because you are stopped only if the stack had turned off, but you can still get another try. but yes, multi needs some restrictions too, since barely noone gets even close of being stopped by the judge only those who tries more than 6 cubes (but this belonhs to another topic)
and fmc mo3 takes too much time as i wrote in that topic

 

[Akash Rupela]

I do not see any sort of harm in converting to an ao5.
I propose something of the format - cutoff on 1st solve -> then a lower cutoff on first 3 solves->get an average of 5
Say cutoff to attempt is 7 minutes and lower cutoff is 4 minutes. So someone slower than 7 minutes gets only 1 solve, someone between 4 and 7 minutes for the best solve in first 3 solves gets a mo3. And someone who gets a sub4 single in first 3 solves qualifies for an avg5. Both these cutoffs should be decided by the organisers rather than a standard for them so that there are no time problems in any case.
What is wrong with this is that its making things very untidy and non standardized, all for just safety solves to mess up? Average of 5 is certainly somewhat exhausting too on big cubes. As bence stated a very good point , the luck factor due to parity is a thing which should make us consider ao5 rather than mo3. There are both pros and cons. Directly calling it must have ao5 or must stick to mo3 wont be wise.

About the giving best and winning competitions part, I dont think one should really worry about winning if he gives his best. It may pop one day but sometime wen it doesn't, he wil be in for a show. It happened to me for example at my last competition. In final round, I was just trying to beat my official PB, i had a big internal pop and so couldn't win it but not trying just because its mo3 is not a good argument(its understandable at worlds, but generally i dont think of it as an argument).

 

[Benyó]

creating new types of cut offs doesn't help at all. if we would have cut off for the first solve, you would still do a safety solve instead of a full speed one, then what is the point of it? and the administtation would become more difficult. the solution only can be the restriction of limits

 

[uberCuber]

oh look there's a poll

 

[Antonie faz fan]

If whe change it whe would make 2 cuttofs 1 for mean and 1 for average lets say 6.00.00 for mean 4.00.00 for avg5.
What I do think wil be weird is how while the records go are whe going to reset it or something

 

[AvGalen]

I agree that 666 now has reached a level where it should be ao5 exactly because of the reasons in the OP. Everything else seems irrelevant and overthinking it.

 

[cubizh]

It's not a simple call, but I lean more on not liking this.

1) I am not really sure a new 9b) subitem will be added just for 6x6, so I can assume this is basically moving this event from 9b2) to 9b1), which makes the acceptable formats be "Best of X"( where X is 1,2 or 3) and "Average of 5" (means no more Mo3 possible). This will make organizers either have a best of (no average ranking) only or go for the average of 5 with eventual deeper cutoffs. As much as I like an average of 5, where the competitor has room to fail, I lean to believe this will only push more people away from even bothering registering/paying for the event that has cutoffs they won't likely achieve (more than what currently happens) for the sake of time. Time is, of course, important and fast people deserve to have better tries, but I am not sure everyone would be happy (the more with more time consuming events). Right now, I think reducing even more the existing Mo3 cutoffs to Avg5 cutoffs (that I guess will always happen: scrambling a 6x6 correctly takes some time) will not be a good move for the event's popularity.

2) If moving from 9b2) to 9b1) goes forward, I strongly suggest at least reviewing 4g1) for consistency.

3) I have to agree with Thom when saying other events are probably more deserving of change, like "Rubik's Cube: With Feet" from 9b2) to 9b1) or "Rubik's Cube: Blindfolded" from 9b3) to 9b2), which, of course, should be discussed in their own particular threads.

 

[Frubix]

Some years ago Megaminx was also mo3, and now it's ao5