Should official 3BLD mo3s be ranked? (Read the first post)

[cubizh]

I voted yes. Accuracy must be recognised in BLD. I agree with Noah and support his arguments.

[off]And I'm also pretty sure it would be better if we changed the way we rank competitions to something like "mean of the best 2 of 3 attempts". We need to change the way we think about BLD - "I can solve the cube blindfolded... sometimes" means you can't really.[/off]

Actually I think your [off] is very much [on]

 

[elrog]

@ cubizch: I fail to see how this helps your case. your first chart shows that only 5-10% of mean of 3s get all 3 solves even as the number of competitors went up. Your second chart also just doesn't make sense. You have 1-13 on the left, but it says competitors and not year. These statistics wouldn't make sense even if you put years to them because it doesn't match up with your first chart. It also can't all be for one year because it would contradict itself. Could you please elaborate what this is?

 

[cubizh]

I fail to see how this helps your case.

The information I posted was not meant to help any case, but provide a more in depth look at the results under the current regulations. Nothing else. More information contributes to better informed decisions, regardless of which one it is.

Could you please elaborate what this is?

Absolutely!
Since you talked about the charts, they show the evolution of 3BLD format through the years. You can clearly see that organizers have progressed from doing best of 1 or 2 to a preferred best of 3 since 2007. That has been the case ever since, up to the point of it being very very rare to see a 3BLD event that is not Best of 3. I posted this chart as a way to verify if even thinking about mean of 3 was relevant or not in the global picture. If only 5 or 10 competitions had Mo3's, while all others were best of 1, for instance, it would not be relevant to even discuss this.
Regarding the tables I posted earlier, I focused on all the competitions that had 3BLD with Best of 3 format only.
Out of those, I looked at how many cubes people were actually solving on all their three tries.
So for competitions held in 2004 that had 3BLD Best of 3, no one solved all three.
For competitions held in 2013 so far, that had 3BLD Best of 3, 1580 of people's individual results were total DNFs; 1538 of the results had only one valid solve; 827 results had 2 solved and one DNF/DNS; 294 had completed all three solves accurately.
This gives a total sum that shows that the majority of people out of the 3 attempts they are given didn't have any success in all three of them (6431 results) and so on.
I tried to see if the accuracy changed through the years, which wasn't the case.

On the second table, I checked of all those competitions that held 3BLD Best of 3, how many competitors had actually completed a mean.
There are 1189 competitions so far that had 3BLD Best of 3.
On 667 of those, no one completed all three solves with 3 successes.
On 319 of those, only one competitor completed all three solves successfully.
.
.
.
On 1 competition (actually, Worlds 2013) 13 competitors completed all three solves.

That's what the table wanted to illustrate: that only a few people are accurate BLD solvers, under the current regulations.
The idea of adding Mean of 3 as a possibility to become official would allow organizers that want to, to promote competitions where speedy accuracy would be taken into account.
I hope that made it a bit more clear. If you want to know/discover any other information about the past competitions regarding this subject please feel free to ask, as always.

 

[elrog]

The second table was the one that was confusing me. I get it now though. Thanks.

 

[BaMiao]

As a spectator, I like this proposition because it rewards consistency without overly punishing competitors for DNFs.

Just a quick question to the BLDers: How would this change affect your approach to BLD solving in competition? Would you become more conservative? Should we expect slower times in return for fewer DNFs?

My expectation is that we should, though I'm not an expert.

 

[tseitsei]

As a spectator, I like this proposition because it rewards consistency without overly punishing competitors for DNFs.

Just a quick question to the BLDers: How would this change affect your approach to BLD solving in competition? Would you become more conservative? Should we expect slower times in return for fewer DNFs?

My expectation is that we should, though I'm not an expert.

Now I would solve quickly but not uberfast as long as I get a success, if I get a success on my first or second solve I will crazyfastmemo the remaining solve(s) to possibly get a good single.

If the change was made I would solve my first solve quickly (like I do now for the first solve).

If it was a success I would probably go a little more on the safe side to get the mo3.
Obviously if the second solve is DNF I can just crazyfastmemo my last solve to get a fast single possibly.
And if the first solve would be DNF I would solve quickly the second solve and last solve either normally or crazyfast depending on the result of the second solve.

That's what I would do. I don't know about others

 

[sneze2r]

As a spectator, I like this proposition because it rewards consistency without overly punishing competitors for DNFs.

Just a quick question to the BLDers: How would this change affect your approach to BLD solving in competition? Would you become more conservative? Should we expect slower times in return for fewer DNFs?

My expectation is that we should, though I'm not an expert.

My aprroach to BLD wouldn't change at all . I knew before there is something like "missing averages ranking", so i was happy i was 2-nd for a while. But priority remains the same-do the best 3bld solve as it is possible, 3 chances

 

[Noahaha]

As a spectator, I like this proposition because it rewards consistency without overly punishing competitors for DNFs.

Just a quick question to the BLDers: How would this change affect your approach to BLD solving in competition? Would you become more conservative? Should we expect slower times in return for fewer DNFs?

My expectation is that we should, though I'm not an expert.

My normal approach is:

Solve 1 - Get a success i.e. go two or three seconds slower
Solves 2 and 3 - Full speed.

My new approach would be:

Solve 1 - Get a success.
Solve 2 - Full Speed
Solve 3 - If solve two was a fast success, go for a success.

I don't know if there's much of a difference for me though, because I tend to DNF more when I slow myself down too much.

 

[cubizh]

Since the proposal to "Remove "Best of 3" for "Mean of 3" events." seems to be going to be a reality next year, I see no point anymore in thinking it would be a good idea to consider having 3BLD Mo3 oficially added to 9b2).
The idea would be to allow the choice of type of competition to have, giving an extra option and competitive possibility, ("Best of X", fastest wins vs. "Mean of 3", accurately fastest wins) instead of forcing Mo3 and completely changing the spirit of the event in such short notice.
I think people should think about changing this for 2015, and understand Marcell Endrey's thoughts.
As for mixing unofficial Mo3 results with official results, it may be a little weird, but perhaps will allow people to get used to it for when the consideration to change to official will take place, it won't be too much of a change.

 

[ThomasJE]

Is this being considered for the 2014 regulations?

 

[TDM]

I know I should read through the thread, but it's over 100 posts. Has anyone said anything about average of 3? This still means that you need reasonable accuracy to win (2/3 successes), but you don't have to be 100% accurate and do all safety solves like you would with mean of 3. It also means that you can't win from a lucky scramble: you still have to be fast to win.

 

[Noahaha]

Is this being considered for the 2014 regulations?

I believe it is.

I know I should read through the thread, but it's over 100 posts. Has anyone said anything about average of 3? This still means that you need reasonable accuracy to win (2/3 successes), but you don't have to be 100% accurate and do all safety solves like you would with mean of 3. It also means that you can't win from a lucky scramble: you still have to be fast to win.

I suspect that you did not read the first post carefully. The suggestion is only to RANK mean of 3s, not to change the format of the event. Singles would still win competitions, and there would be two ranking lists: one for single and one for mean of 3, similar to how for 3x3 averages win competitions, but singles are also ranked.

In case you did not make the mistake that I think you did, I will say that I don't see any reason to encourage accuracy for the sake of encouraging accuracy. It's not like someone who only gets one solve out of three correct is any less deserving to place in the competition than someone who gets two. That's the nature of the event. It's meant to encourage fast times rather than accuracy.

The purpose of my proposal is not to encourage people to be accurate. The purpose is to add more depth to the 3BLD rankings. From my point of view, the current 3BLD ranking system is almost as bad as if 3x3 were just ranked based on single. You would still see the best solvers at the top, but most people's rankings would be based on just one lucky solve, rather than a larger sample of their abilities.

 

[TDM]

I suspect that you did not read the first post carefully. The suggestion is only to RANK mean of 3s, not to change the format of the event. Singles would still win competitions, and there would be two ranking lists: one for single and one for mean of 3, similar to how for 3x3 averages win competitions, but singles are also ranked.

I know, but I was trying to suggest something different.

In case you did not make the mistake that I think you did, I will say that I don't see any reason to encourage accuracy for the sake of encouraging accuracy. It's not like someone who only gets one solve out of three correct is any less deserving to place in the competition than someone who gets two. That's the nature of the event. It's meant to encourage fast times rather than accuracy.

The purpose of my proposal is not to encourage people to be accurate. The purpose is to add more depth to the 3BLD rankings. From my point of view, the current 3BLD ranking system is almost as bad as if 3x3 were just ranked based on single. You would still see the best solvers at the top, but most people's rankings would be based on just one lucky solve, rather than a larger sample of their abilities.

Fair enough. Although BLD can't really be compared to sighted events, as some of the best people may not get many Mo3s in a comp. Most people who do 3x3 have a near-100% success rate. That's why I think that even if comps weren't won by an Ao3, I still think that Ao3 would be better for BLD.

 

[Sajwo]

It happened. Mo3 will be official since 1st January 2014