3x3x3 Blind, changing to 3/5?

[Kirjava]

So what disadvantage does best of 5 have that best of 3 doesn't?

 

[tim]

So what disadvantage does best of 5 have that best of 3 doesn't?

Oh, come on. Your "only advantages" is so close to "by introducing the average format", that (at least) i read it as: "best of 5 has only advantages over 3/5". And i'm sure you knew that for the last two posts...

 

[Kirjava]

It's more fun to let people discover their own mistakes!

 

[tim]

It's more fun to let people discover their own mistakes!

So you've discovered the ambiguity in your post?

 

[Kirjava]

Not at all, it was quite clear.

 

[EVH]

I am a noob, so what occurs if you get more than one DNF does the entire average become a DNF?

 

[Ville Seppänen]

ya

 

[qqwref]

I guess all this discussion supports my suspicion that we're just not ready for average of 5. I think if we have this discussion 5 years from now, it will likely be very different. I suspect that if Haiyan or Ville continue working at it for those 5 years, by then they'll be getting times just as good as or better than they get now, but with 80% or better accuracy. At that point, we can have this discussion again, and it will be more likely to be considered.

I really do think that with people who practice a lot, 80% accuracy will eventually be common. It just hasn't happened yet.

And by the way, Gabriel Alejandro Orozco Casillas (who I mentioned earlier in this thread) has a success rate of 8/9 for 3x3x3 BLD, and 3/3 for 4x4x4 BLD. And he's 3rd in the world at both. So certainly it is possible to be very fast and accurate simultaneously.

I don't think the issue is that the community isn't ready; I think it's that 4 of 5 in 3BLD is difficult and is a lot to ask. I'm sure people can practice enough at 3x3 to get under 10 seconds every time but that doesn't mean we should DNF people who are over that mark. Cubing isn't supposed to be for only the most elite, but for everyone. You will never reach the point where for a small local competition everyone who wants to do BLD is capable of doing 4 out of 5 solves at full speed every time. We shouldn't make it artificially hard to just get a success, in any round. 4BLD and 5BLD are very difficult (not because of the format but because it's a hard event to do at all) and look at how popular they are - and look at how many competitions have no winners. It's a lot more exciting to see people get one good solve and two mistakes than to see someone get a few very fast solves and then lose to someone who did four 7 minute (or even 3 minute) solves.

 

[Daniel Wu]

I am a noob, so what occurs if you get more than one DNF does the entire average become a DNF?

Right now the format is best of three. So as long as you get one non DNF solve, the average won't be a DNF because the format is best of three.

Supporting what qq said, I'm one of those people who would find it hard to get 4/5 solves. Officially, I'm 1/6 for my BLD success rate. I saw no real point in going slowly and making sure I made no mistakes in execution or memorization. I probably could have done 3/3s in both of those BLD rounds but I went as fast as I could in order to take 2-3 minutes instead of 4-5 minutes. Anyhow, I think that average of 5 format for competitions is impractical because of the limited time of competitions as well as the high DNF rate. I'd rather watch a person do a 40 second BLD solve and a few DNFs than 5 perfect solves that are slower than what said person is capable of. Pretty much speed and accuracy of a balance and I think average of 5 tips it too much toward accuracy.

 

[AvGalen]

Wow, a lot of new posts and lots of interesting content in them to address. Since I opened this can of worms I will try to comment on everything. In order to keep my post from becoming a meter long I will not quote most things, but instead use the @Person#Number format (my opening post being @AvGalen#1)

@Tim Reynolds#19: As I mentioned earlier, I do not believe that going slower gives a higher chance of success. I personally DNF fast and DNF slow and solve "fast" and slow as well. If media sees many DNF's they will think it is too difficult and they will not report on it (media doesn't report on the people dropping out of a marathon). If DNF's are too common media and the audience loose interest.

@Olivér Perge#21: The increade amount of time needed is absolutely a drawback, as I mentioned already in the opening post

@tim#22: Too bad you seem to like it just because it would benefit you personally
@Ville Seppänen#23: Too bad you seem to dislike it just because it would harm you personally (I am guessing since you didn't list any reason)

@Sakarie#25: Most competitions only have regular blindfolded, but not multi or bigcube. And just like so many others you assume that this format would lead to much slower times without providing any rational behind that assumption. And yes, the WCA cares most for the competitors, but they also like to entertain the audience, media and sponsors.

@qqwref#26: Why do you assume that people that need 6 minutes CAN solve at least 4/5 cubes while people that need 1 minute CAN'T? And if you make 2 or 3 mistakes in a round of 5 I see no reason not to kick someone out. You get disqualified from most other sports at least that fast (although I do see parallels with weightlifting and highjumping-like sports that allow for more errors, which annoys me as well to see so many failures). Currently, fast people that rarely have success will either take 1st place or last which seems too extreme for me. I am claiming that getting an average at a good speed isn't extremely hard. I am claiming that some people are getting crazy fast times because they are taking extreme risks and don't care for the DNF's. If Ville or Haiyan does a 40s solve on the first solve they already know that they have won and will take every risk in their other solves to break the WR. And Haiyan doesn't use his practice solves to try to do an average, he uses it for practice (and a chance to break the WR for blind "unofficially")

@Pedro#27: You are basically saying that you are doing solves about 30% faster than you can actually do them "normally" (meaning not taking risks). We gave up on the idea of perfection for multi-blind indeed, but if you look at the current top 100 there are almost no solves in there that have > 1 unsolved cube. Multi-blind was black-and-white and became grey. I am not proposing to turn blind from grey into black-and-white. I am proposing to make it like everything else 1 good and 1 bad removed. I also don't like the idea of your (extreme) example but find that very unlikely to occur and will propose a (radical) solution for that at the end of this post. I am also wondering if your "go 40% slower to get a guaranteed solve"-strategy worked out every time. If you could provide me with some statistics of fast vs safety solves (success-rate and time) that would be very useful

@Kirjava#29: Best of 5 doesn't have only advantages over the current format (it requires extra time) and it doesn't have only advantages over 3/5 (it promotes extreme risk-taking even more and wouldn't allow for the cut-off procedure I mentioned)

@aronpm#30: Why is it too difficult? Because they don't practice for accuracy but for optimal single speed, which makes perfect sense under the current rules. The point of rankings in competitions is not to get the best times you can. It is to compare your skill to others. Again, the link between slower and more succesful isn't proven.

@bobbyfearfactor: My reasoning is that people will practice more with focus on accuracy. All the other issues you mention are solvable with cut-offs as I mentioned in the start post. Your "I personally prefer speed and taking chance" basically sums it up. You go faster than you can actually go while hoping to solve it anyway

@maggot#32: That happens with blind as well and people like that should get a warning, a ban and ridicule from the community. They make a joke out of the competition and waste the competitions resources.

@Quadrescence#33: Best of 5 would provide a good way to represent someone's OPTIMAL, not average capabilities. Nothing wrong with that, but not what I think we should focus on and that is the crux of this topic. Do we want to measure single, or average performance?

@qqwref#34: Are "gods of consistencies" necessarily slower? Mike's example of Gabriel Alejandro Orozco Casillas seems to disprove that. You seem to be convinced that fast people could only get consistent enough by going slower. I am of the opinion that speed and accuracy aren't linked very much, except for going beyond one's true capabilities to reach WR. I hope to get some clearance about this in this topic

@tim-Pedro#35-37: So we agree that best of 5 would enhance the results of lucky cases and that results would be different depending on group. That seems like proof to me that best of 5 is BAD

@Mike Hughey#40: I was thinking it would be to early to have this discussion, but I am actually surprised that I got so much support. I was expecting to get my head bitten of. Thanks for providing the example of Gabriel Alejandro Orozco Casillas. I have another example, just from FaceBook: Kai Jiptner "now I can tell: I did sub-1 avg of 12 with 57.xx". He didn't provide more details (yet) unfortunately

@qqwref#49: I still don't understand why you think that slow people CAN do 4/5 but fast people can't. If I look at your results your times decreased while your accuracy increased. This seems perfectly normal to me as a result of more practice and experience.

@rickcube#50: So you are saying that you have a 1/6 accuracy while taking 2-3 minutes and at least 6/6 (3/3, multiple rounds) when you take 4-5 minutes? You might be right, but I would like to see evidence. You might even surprise yourself. So I am asking you the same as I asked Pedro (others, feel free to do the same): Please provide some statistics of fast vs safety solves (success-rate and time)

That is it for now, made it all the way to 50, refreshed, and saw no new posts. I am going to post a radical proposal soon after this

 

[Faz]

Interesting. Although I could never get an average of 5 in comp, unless I memorise really carefully (Still, I would probably DNF 2) I think it would add a whole extra level to BLD rankings, and would favour accuracy, rather than recklessness. I think the average world record would be very interesting to follow. People could still go all out for the single WR, and then in a different competition, they could go for the average of 5 WR.

 

[AvGalen]

This topic is doing so well that I don't want to complicate it with a second proposal, but I think it would be best to discuss it here for now. If it is interesting enough, we can always open another topic and continue there.

So let me propose an idea that is very rough and that I really would like to see discussed as well:
What about making blind a "how many successes can you get in xx amount of time" format?
Time planning for blind has always been a major problem. This idea would solve that while not favoring speed over accuracy. Let me list the basic idea I have:

- You get a total time for the event of xx minutes (let's assume 10 for now)
- You solve 1 cube at a time with a different scramble every time (just like we do know, not like multi-blind)
- You get a point for every cube you solve
- You loose a point for every cube you don't solve OR you loose time (let's assume 30 seconds for now) for every cube you attempt. (This is to avoid cherry-picking of easier scrambles. We don't want people to inspect a cube for 10 seconds, put it down and say "give me the next scramble"). I am not sure which of these 2 systems (deduct point or deduct time) would be he best. Let's discuss
- Scoring is done on most points first, time second (so 10 points in 9:50 beats 9 points in 8:45, making the 9-pointer make a strategical decision if he could solve a 10th cube and do it fast enough)

A fast person with low accuracy would be able to score 4 points after doing 12 attempts
A slow person with high accuracy would be able to score 4 points after doing 5 attempts
slow people with low accuracy would be able to score 1 point after doing 2 attempts
fast people with good accuracy would be able to score 8 points after doing 10 attempts

This scoring system
-doesn't seem to favor speed over accuracy
-would allow fast people to do more solves (more fun) than they do now
-allows for > 1 DNF
-could still be used to list individual (fastest single) times as well
-and would make blind events more predictable for the organisors

They challenges and disadvantages I can see are that
-the total time would have to be fixed among competitions to allow results to be comparable
-the "deduct point or deduct time for an attempt" would have to be chosen wisely
-if "deduct time" is chosen, it's value should be carefully chosen
-total used time for all attempts has to be kept track of by the judge
-more scrambling is involved in some cases

A radical idea? I would say so. Please read the proposal a few times, think about it and then let me know
- what you think about the basic idea
- what you hate about it
- what you like about it
- if you would actually like to see it in WCA competitions
- and if you have ideas about the challenges and disadvantages I mentioned

 

[tim]

@tim#22: Too bad you seem to like it just because it would benefit you personally
@Ville Seppänen#23: Too bad you seem to dislike it just because it would harm you personally (I am guessing since you didn't list any reason)

Nah, i do have more reasons why i like it, but you've already posted all of them in your first post (like removing lucky solves and more successful solves). I guess the same is true for Ville.

So let me propose an idea that is very rough and that I really would like to see discussed as well:
What about making blind a "how many successes can you get in xx amount of time" format?

...

A radical idea? I would say so. Please read the proposal a few times, think about it and then let me know
- what you think about the basic idea
- what you hate about it
- what you like about it
- if you would actually like to see it in WCA competitions
- and if you have ideas about the challenges and disadvantages I mentioned

First: Yes, it's radical and seems to solve the drawbacks of 3/5. Second: It's way too complicated. Especially for spectators. You have to keep track of the total time + solved cubes to have an idea who's leading and who's not. That's impossible without a screen containing live results.

For example: Person A has 4P in 8:10 and averages about 1:20 for a solve. Person B also has 4P in 8:10 and also averages about 1:20. Depending on Person A's next solve Person B's strategy is completely different.
Basically what i'm trying to say is: All competitors need either all information or none. None informaion is impractical as long as there are spectators who applaud or happy/unhappy competitors.

 

[Sakarie]

A radical idea? I would say so. Please read the proposal a few times, think about it and then let me know
- what you think about the basic idea
- what you hate about it
- what you like about it
- if you would actually like to see it in WCA competitions
- and if you have ideas about the challenges and disadvantages I mentioned

Sure, have it as an event, but I really hope you don't mean that this is also to exchange todays 3x3 blind for?

BUT, you (or whoever it was) say that most competition doesn't have big cube or multi. But this? Should this be prioritated over multi, or 4x4?

Also, VERY hard for administration I think. I wouldn't want to be the one responsible for that everyone gets the correct cube. It's an event for a very small selection of people, and my first naive opinion of it is that it is way too complicated.

 

[Pedro]

@Pedro#27: You are basically saying that you are doing solves about 30% faster than you can actually do them "normally" (meaning not taking risks). We gave up on the idea of perfection for multi-blind indeed, but if you look at the current top 100 there are almost no solves in there that have > 1 unsolved cube. Multi-blind was black-and-white and became grey. I am not proposing to turn blind from grey into black-and-white. I am proposing to make it like everything else 1 good and 1 bad removed. I also don't like the idea of your (extreme) example but find that very unlikely to occur and will propose a (radical) solution for that at the end of this post. I am also wondering if your "go 40% slower to get a guaranteed solve"-strategy worked out every time. If you could provide me with some statistics of fast vs safety solves (success-rate and time) that would be very useful

Ok, from my official results:

Worlds 07: 2:08 success in the first round, after a DNF, so I wanted to at least get one. Final was 3 DNFs, some really close.

Brazil Open 07: 1:53 success on first attempt, then 2 DNFs, because I tried to go faster (and I could).

Minas Open 08: 1:23 success on first attempt, 2 DNFs. Same story.

São Paulo Open 09: Two over-2-min times (both on second attempt). Was trying a different strategy, which was getting a success (a slower but secure attempt) and then trying to go fast. Didn't work really well...

Santiago Open 09: 1:29 sucess on first attempt, 2 DNFs...I think I've seen that before.

Brasília Open 09: 2 DNFs, so I had to go slower to make sure I'd get one and qualify. Then failed the final (as everybody else) (Didn't want to go slow just to win).

Worlds 09: 1:47 sucess (bleh) then a faster DNF, qualified in 16th place. Final was: DNF (1:14, trying fast), 2:08 (to make sure I'd have a success and don't fail like last time) and DNF again (don't remember the time). Good enough for 9th place. I think this is a good example. Everyone but Rafal, Yumu and Guillain did worse (or a lot worse (Norbert 1:30 - 3:30)) than first round, which was only 2 attempts...Why is that?

UNESP Open 09: 1:20 and 1:19, first (of 3 so far) time that I got 2 successes in a single round. Third was 1:18, but fail. Final was 2 DNFs and a slow solve to make sure I'd win.

ABC Open 10: DNF, 1:46 to make sure I'd qualify, DNF. Both DNFs were faster than 1:46. Final: 1:27, 1:18, DNF. Second time ever doing 2 solves.

SESC/Santos 10: 1:18, DNF, magical-wtf-easy-scramble-let's-go-fast 1:03 . Then 2 DNFs and a slow solve in the final...didn't want to screw up yet another time.

So, my official record is 18/49 = 36.73%, because I try to beat my best time. My "average" is 1:39.82. I could have a 2:00 average with maybe 70% or more accuracy, but I never saw the point. And I would probably never got that 1:03 (or any of the sub-1:20).

(I'll talk about your other proposal on the next post.)

 

[Kirjava]

@Kirjava#29: Best of 5 doesn't have only advantages over the current format (it requires extra time)

You can still hold best of 3 or less if time is an issue. Time requirement is not a disadvantage.

and it doesn't have only advantages over 3/5 (it promotes extreme risk-taking even more and wouldn't allow for the cut-off procedure I mentioned)

Risk-taking is already what BLD is about, and this brings it closer to what it's like to do BLD at home.

As for cut-offs, it can always have it's own style of them - they have already been mentioned! (Did you listen to the latest podcast? )

 

[Pedro]

So, about your new proposal, I don't really like it. We'd have to fix times for the bld event to be able to compare competitions around the world. Some people just squeeze in half an hour of bld because of demand, some have 2 hours and 50 people. Some do cut-offs for the first time, then for the second one, while some let people do 3 attempts in 10 minutes each (we do, most of the time). Also, as Tim pointed out, it's difficult to the audience to follow, more work for judges (people don't really enjoy holding a piece of paper for 5 minutes), more complication.

If I understood correctly, your main idea behind this is to make bld the same format as (almost) everything else. But it's not a speed event. You don't practice the same way. If you can sit down and do 50 solves, non-stop, and get your best time in the 46th solve, I can consider (I can do that in speed events, even OH, which involves more tiredness...some people do 200 3x3 solves PER DAY).

 

[AvGalen]

@tim#54: Keeping track won't be as hard as it seems. Most people don't keep exact track of all solves (not even in 3x3x3 final) but they have a good idea who is doing well and who isn't. If I see A do a few solves then I know he is doing well, just like when B would do a few sub 11's in a final. Also, it isn't my idea to introduce a huge tactical game, but the top blindfolders could try anticipating and analysing others results indeed which isn't a problem for me (can happen in multi-blind or bigcubes blind or even speedsolves as well). As long as a competitor solves at his best, he will put pressure on others and increase his chances of winning.

@Sakari#55: Yes, I did mean as a replacement of current 1/3 blind. And it isn't complicated at all for administration. All the judge needs to do is determine how many time is left and what the results for all attempts were (duh). It could look something like this (assuming deducting points for DNF):
Attempt, Result, Time spent on this solve, Total time left, Score
1, DNF, 59.16, 9:00.84, -1
2, Solved, 1:14.25, 7:46.59, 0
3, Solved, 1:36.37, 6:10.22, 1

and when the scramble table gets his cube they apply the 4th scramble just like they would determine that a speedsolver has done the first 3 scrambles but not the 4th one yet.

 

[AvGalen]

@Pedro#56: Thank you for taking the time to write all of that. I can't really draw a conclusion from your competition results and analysis though. It seems that you got faster AND more accurate over time. It seems that safety solves only work sometimes but DNF's are more likely to happen when you push beyond your limits. But when you push beyond your limits you also did your best time ever, confirming to me that your best time was more a result of tactic and luck than of skill. I am assuming the results from Worlds 09 in the final were worse because of pressure that some people handles better than others?

@Kirjava#57: I think it would be very unfair to have competitions with 1/3 (or 1/2 sometimes) and other competitions with 1/5. Records wouldn't be comparable anymore. According to you we should make competitions more like home which is just silly. Make your home solves more like competitions instead! I am not going to a competition to scramble my own cubes, start/stop the timer by pressing spacebar while listening to my boombox!
I haven't listened to the podcast and won't. Podcasts are easy for the producer (no need to type, just talk) and easy if you prefer listening over reading but they are too inconvenient for searching. Text rules!

@Pedro#58: I think there is too much variation in blind currently making the winner of a competition the rightful winner, but making records less comparable. I understand how difficult it is to estimate how much time you need, which is why my proposal aims to solve that as well. If you have 50 competitors for blind you would need 50*10*(1+overhead_factor) / nr_of_timers minutes (10 being the total time as I proposed, overhead_factor being the percentage of time that you loose with scrambling, getting to your timer, changing groups, etc). I do agree that we need to find a solution for "holding the piece of paper", but that is independant of my proposal here.
And I agree about the fatique factor, which is why I proposed the total time to be only 10 minutes. Surely practising 10 minutes (maybe even multiple times per day) isn't too hard for the top-blinders? Haiyan told me he did 200 blind solves per day, I have seen Dennis Strehlau do 5/5 (small multiblind) between ordering food and getting the soup, and I see more and more people doing warm-up blindsolves during competitions. Maybe you are so inaccurate because you focus on going fast, but undertrained your memo?

 

[Kirjava]

@Kirjava#57: I think it would be very unfair to have competitions with 1/3 (or 1/2 sometimes) and other competitions with 1/5. Records wouldn't be comparable anymore.

This already happens anyway? Some compitions have multiple rounds of events and other competitions only have one round. Why aren't you complaining about them, too?

According to you we should make competitions more like home which is just silly. Make your home solves more like competitions instead! I am not going to a competition to scramble my own cubes, start/stop the timer by pressing spacebar while listening to my boombox!

The point I was making is that you're trying to treat BLD like speed, when the two are very different.