[WCA Regulations 2014] Mean of N for FMC

[Ton]

You are right that a DNF would be harsh if the format is Mean of 3. But that's something that also goes for 6x6x6, 7x7x7 and feet. And it doesn't seem to be a problem there. And as I said before, I think that is far easier to not DNF in FMC than at every other event. As far as you don't do crazy last minute finishes, you can finish your solve in FMC and check your solution multiple times before the attempt ends. This is a benefit that you don't have for any other event.

I also don't think that it is a common thing that "better" FMC competitors hand in a solution which is supposed to work and then doesn't. I haven't had a single DNF like this in my whole cubing career.

As I wrote, sure mean of 3 should be allowed, but I judged many FMC attempts and many DNF's where written mistake , where R' it says R like mistakes, still they did check many times. So you might be very accurate in writing down your solve, but you do not want to pay lunch for every DNF in the past by a small write error... But yes a mean of 3 in large competitions would for sure be an addition for the top FMC cubers .
I do not think a mean of 3 will work for smaller competitions, where there is just no room for a 3+ hour event. But than again this is up to the organizers

 

[Sebastien]

As I wrote, sure mean of 3 should be allowed, but I judged many FMC attempts and many DNF's where written mistake , where R' it says R like mistakes, still they did check many times. So you might be very accurate in writing down your solve, but you do not want to pay lunch for every DNF in the past by a small write error... But yes a mean of 3 in large competitions would for sure be an addition for the top FMC cubers .
I do not think a mean of 3 will work for smaller competitions, where there is just no room for a 3+ hour event. But than again this is up to the organizers

Seems that we are in agreement here. Sure, I acknowledge that I also checked lots of DNFs like this. I can't help imagine any other reason though then that those people were too sloppy when writing down their solution and while checking it. I mean, we all know that notation, don't we? I guess those people where just going though their solution like "ok, T-Perm, I do a T-perm now and I wrote down a T-Perm, fine" missing that they wrote down one move wrong within their algorithm, instead of checking the solution move per move. I don't see how it is possible to get a solved cube 2-3 times in a row when checking a solution move per move that is indeed written down wrongly.

 

[amostay2004]

Seems that we are in agreement here. Sure, I acknowledge that I also checked lots of DNFs like this. I can't help imagine any other reason though then that those people were too sloppy when writing down their solution and while checking it. I mean, we all know that notation, don't we? I guess those people where just going though their solution like "ok, T-Perm, I do a T-perm now and I wrote down a T-Perm, fine" missing that they wrote down one move wrong within their algorithm, instead of checking the solution move per move. I don't see how it is possible to get a solved cube 2-3 times in a row when checking a solution move per move that is indeed written down wrongly.

In my last comp, I double and triple checked my solution on the cube, move by move, and I still got 1 move wrong lol

Then again I suck at FMC

 

[Sebastien]

Competitions with FMC event divided by 'Best of'

My small contribution to the FMC Mo3 debate

Thought I should quote this here.

 

[okayama]

Result from Cube Camp in Kanazawa 2013, maybe a good example to see a difference between the formats.

Best of 3

1. 25 (25, 33, 33) Yuki Tanaka
2. 26 (28, 26, 35) Hideki Niina
3. 27 (30, 40, 27) Wataru Hashimura
4. 28 (28, 29, 30) Tomoaki Okayama
5. 29 (29, DNF, DNF) Fumiya Matsui
6. 30 (30, 38, 35) Shuto Ueno
7. 32 (32, 37, 38) Takayuki Ookusa
8. 37 (DNF, DNF, 37) Taku Yanai
9. 38 (44, DNF, 38) Nobuaki Suga
10. 39 (42, 39, 42) Ryo Ito
11. 41 (DNF, DNF, 41) Takafumi Seki
12. 42 (42, DNF, 42) Yuuki Kobayashi
13. 43 (43, 46, 47) Sinpei Araki
14. 43 (43, 47, 53) Takayuki Matsumoto
15. 49 (49, DNF, DNF) Sanae Koseki
16. DNF (DNF, DNF, DNF) Jin Mochizuki
17. DNF (DNS, DNF, DNF) Kou Oobatake
18. DNF (DNF, DNF, DNF) Yuki Matsuba
19. DNF (DNF, DNF, DNF) Yusuke Morikawa

Mean of 3

1. 29.00 (28, 29, 30) Tomoaki Okayama
2. 29.67 (28, 26, 35) Hideki Niina
3. 30.33 (25, 33, 33) Yuki Tanaka
4. 32.33 (30, 40, 27) Wataru Hashimura
5. 34.33 (30, 38, 35) Shuto Ueno
6. 35.67 (32, 37, 38) Takayuki Ookusa
7. 41.00 (42, 39, 42) Ryo Ito
8. 45.33 (43, 46, 47) Sinpei Araki
9. 47.67 (43, 47, 53) Takayuki Matsumoto
10. DNF (29, DNF, DNF) Fumiya Matsui
11. DNF (DNF, DNF, 37) Taku Yanai
12. DNF (44, DNF, 38) Nobuaki Suga
13. DNF (DNF, DNF, 41) Takafumi Seki
14. DNF (42, DNF, 42) Yuuki Kobayashi
15. DNF (49, DNF, DNF) Sanae Koseki
16. DNF (DNF, DNF, DNF) Jin Mochizuki
17. DNF (DNS, DNF, DNF) Kou Oobatake
18. DNF (DNF, DNF, DNF) Yuki Matsuba
19. DNF (DNF, DNF, DNF) Yusuke Morikawa

Median of 3

1. 28 (28, 26, 35) Hideki Niina
2. 29 (28, 29, 30) Tomoaki Okayama
3. 30 (30, 40, 27) Wataru Hashimura
4. 33 (25, 33, 33) Yuki Tanaka
5. 35 (30, 38, 35) Shuto Ueno
6. 37 (32, 37, 38) Takayuki Ookusa
7. 42 (42, 39, 42) Ryo Ito
8. 42 (42, DNF, 42) Yuuki Kobayashi
9. 44 (44, DNF, 38) Nobuaki Suga
10. 46 (43, 46, 47) Sinpei Araki
11. 47 (43, 47, 53) Takayuki Matsumoto
12. DNF (29, DNF, DNF) Fumiya Matsui
13. DNF (DNF, DNF, 37) Taku Yanai
14. DNF (DNF, DNF, 41) Takafumi Seki
15. DNF (49, DNF, DNF) Sanae Koseki
16. DNF (DNF, DNF, DNF) Jin Mochizuki
17. DNF (DNS, DNF, DNF) Kou Oobatake
18. DNF (DNF, DNF, DNF) Yuki Matsuba
19. DNF (DNF, DNF, DNF) Yusuke Morikawa

I like Median of 3, next Mean of 3. Best of 3 is not good.

 

[ThomasJE]

I like Median of 3, next Mean of 3. Best of 3 is not good.

I agree. We are allowed 1 DNF in most events, so if we use Median of 3, we are allowed a DNF in FMC as well; especially as DNF's are very common.

Because of this; could Median of 3 be a possible format for BLD as well? This could be a possibility if we want to place emphasis on accuracy and consistency rather than on one lucky scramble so that you need 2 successes instead of just 1. Maybe not for 4BLD and 5BLD, but for 3BLD and MBLD possibly.

Back to FMC; we're talking about medians of 3, but what for competitions where there is only time for 2 attempts? Do we use best of or median (mean in this case) to decide a winner (if we were to impliment this new format)?

 

[okayama]

I agree. We are allowed 1 DNF in most events, so if we use Median of 3, we are allowed a DNF in FMC as well; especially as DNF's are very common.

Because of this; could Median of 3 be a possible format for BLD as well? This could be a possibility if we want to place emphasis on accuracy and consistency rather than on one lucky scramble so that you need 2 successes instead of just 1. Maybe not for 4BLD and 5BLD, but for 3BLD and MBLD possibly.

Back to FMC; we're talking about medians of 3, but what for competitions where there is only time for 2 attempts? Do we use best of or median (mean in this case) to decide a winner (if we were to impliment this new format)?

NB: I understand that Sébastien prefers Mean of 3, mainly because WCA already has the format,
and thus it's relatively easy to push for the format rather than to make a completely-new format.
However, as ThomasJE and many others said, Median of 3 has such a great advantage (one DNF
is allowed), and furthermore IMO it fits FMC in view of the results above.

As for BLD, it's worth considering, actually I think the idea has been proposed somewhere,
but I couldn't remember the thread. Personally I think Median of 5 fits 3BLD.

I don't think it's good that the mean (or median) result of 2 attempts and 3 attempts go to
the same ranking. If Mean (or median) of 3 is introduced, 2 attempts should be Best of 2.

 

[AvGalen]

I agree. We are allowed 1 DNF in most events, so if we use Median of 3, we are allowed a DNF in FMC as well; especially as DNF's are very common.

Because of this; could Median of 3 be a possible format for BLD as well? This could be a possibility if we want to place emphasis on accuracy and consistency rather than on one lucky scramble so that you need 2 successes instead of just 1. Maybe not for 4BLD and 5BLD, but for 3BLD and MBLD possibly.

Back to FMC; we're talking about medians of 3, but what for competitions where there is only time for 2 attempts? Do we use best of or median (mean in this case) to decide a winner (if we were to impliment this new format)?

We are allowed 1 DNF in all events that are Average of 5, but in not in 666 or 777. Blind events are not averages so they are different

Most competitions don't have FMC
Most competitions with FMC only have 1 attempt
Most competitions with FMC have lots of DNF's for many reasons.

Making FMC a 3 attemps event would take away time from other events, would put a lot of pressure on people that have to check the results and would only benefit the top FMC people. It would also force many people to change the way they do FMC. Instead of going for "1 good solve" it would force people to do safetysolves for 3 hours.
I am an FMC lover and I think this is a bad idea for almost all people

 

[ThomasJE]

We are allowed 1 DNF in all events that are Average of 5, but in not in 666 or 777. Blind events are not averages so they are different

BLD's are not averages, but there has been talk of ranking averages for BLD I think. This could be done in the same way that 3x3 is ranked by average in competitions, but we have a world ranking for single.

Most competitions don't have FMC
Most competitions with FMC only have 1 attempt
Most competitions with FMC have lots of DNF's for many reasons.

Making FMC a 3 attemps event would take away time from other events, would put a lot of pressure on people that have to check the results and would only benefit the top FMC people. It would also force many people to change the way they do FMC. Instead of going for "1 good solve" it would force people to do safetysolves for 3 hours.
I am an FMC lover and I think this is a bad idea for almost all people

The 'Median of 3' only really applies to major competitons such as Worlds where more time is available for an event like this. I think that would also be good if for a major competition, someone who consistently gets a low count (around 25ish) wins rather than someone who gets a lucky 20/21 and wins. The smaller competitions are fine as they are; but an optional format for large competitions is a good idea, even if it is used only 2 or 3 times a year.

 

[AvGalen]

BLD's are not averages, but there has been talk of ranking averages for BLD I think. This could be done in the same way that 3x3 is ranked by average in competitions, but we have a world ranking for single.



The 'Median of 3' only really applies to major competitons such as Worlds where more time is available for an event like this. I think that would also be good if for a major competition, someone who consistently gets a low count (around 25ish) wins rather than someone who gets a lucky 20/21 and wins. The smaller competitions are fine as they are; but an optional format for large competitions is a good idea, even if it is used only 2 or 3 times a year.

Let's do average of 12, but only during Worlds.
And for the 100 meter sprint during the Olympics we make it 125 meters okay?
Rules should be the same for all competitions...that is in the rules!

 

[Sebastien]

It would also force many people to change the way they do FMC. Instead of going for "1 good solve" it would force people to do safetysolves for 3 hours. I am an FMC lover and I think this is a bad idea for almost all people

I think you should question how you do FMC then. In every other event, I finish my attempts with the best result I can achieve within that attempt. Nowadays, when I do FMC, I try to handle it the same way. So I don't do safety solves, I just finish with the best result that I can get within the time limit.

The 'Median of 3' only really applies to major competitons such as Worlds where more time is available for an event like this. I think that would also be good if for a major competition, someone who consistently gets a low count (around 25ish) wins rather than someone who gets a lucky 20/21 and wins. The smaller competitions are fine as they are; but an optional format for large competitions is a good idea, even if it is used only 2 or 3 times a year.

That's the wrong way of thinking. Mean of 3 should be applied if an organiser decides to have 3 attempts. And Major Competitions should have 3 attempt anyway imo. That's it.

 

[Stefan]

The 'Median of 3' only really applies to major competitons such as Worlds where more time is available for an event like this. I think that would also be good if for a major competition, someone who consistently gets a low count (around 25ish) wins rather than someone who gets a lucky 20/21 and wins. The smaller competitions are fine as they are; but an optional format for large competitions is a good idea, even if it is used only 2 or 3 times a year.

I repeat: no world or continental competition has had three attempts.

What makes you think major/large competitions have more time for it than smaller ones? The actual data suggests the opposite - the largest competition with three (or more) FMC attempts had only 68 competitors (overall competitors, not just FMC competitors). And WC2003, WC2007, WC2009 and WC2011 even just had a single attempt each.

Top 200 competitions sorted by attempts, then competitors:

+----------------------------+-------------+----------+
| competitionId              | competitors | attempts |
+----------------------------+-------------+----------+
| MovesN00bs2012             |          20 |        6 |
| TwenteOpen2012             |          68 |        3 |
| UAMOpen2013                |          52 |        3 |
| SpanishChampionship2012    |          51 |        3 |
| MurciaOpen2013             |          50 |        3 |
| BerlinOpen2013             |          48 |        3 |
| CampeonatoSudamericano2013 |          41 |        3 |
| GranadaOpen2012            |          36 |        3 |
| ValenciaSummerOpen2013     |          33 |        3 |
| DualGames2013              |          33 |        3 |
| AlaniaOpen2011             |          32 |        3 |
| VelbertOpen2013            |          32 |        3 |
| PuydeDomeOpen2012          |          31 |        3 |
| MolinaOpen2013             |          29 |        3 |
| DolmenOpen2012             |          26 |        3 |
| DolmenOpen2011             |          25 |        3 |
| ClermontOpen2009           |          24 |        3 |
| GernikaOpen2013            |          21 |        3 |
| ValgameOpen2012            |          21 |        3 |
| CastellonOpen2012          |          20 |        3 |
| CastellonOpen2013          |          19 |        3 |
| PKUFMC2010                 |          16 |        3 |
| ChinaFM2011Xian            |          15 |        3 |
| PuydeDomeOpen2011          |          15 |        3 |
| BeijingFootsteps2011       |          14 |        3 |
| ChinaFM2011Beijing         |          13 |        3 |
| ChinaFM2011Shanghai        |          13 |        3 |
| PKUFM2012                  |          12 |        3 |
| ChinaFM2011Guangzhou       |           6 |        3 |
| ChinaFM2011Shenyang        |           3 |        3 |
| ChinaFM2011Zhengzhou       |           3 |        3 |
| WC2013                     |         580 |        2 |
| USNationals2012            |         259 |        2 |
| USNationals2010            |         227 |        2 |
| USNationals2011            |         204 |        2 |
| WC2005                     |         149 |        2 |
| USNationals2009            |         135 |        2 |
| SpanishChampionship2011    |          93 |        2 |
| MITSpring2012              |          92 |        2 |
| BWOpen2013                 |          76 |        2 |
| PolishOpen2007             |          76 |        2 |
| PolishOpen2006             |          72 |        2 |
| France2013                 |          66 |        2 |
| Poland2009                 |          63 |        2 |
| PolishNationals2010        |          62 |        2 |
| NacionalPerubik2013        |          61 |        2 |
| ZonhovenOpen2013           |          58 |        2 |
| BarcelonaSummerOpen2011    |          57 |        2 |
| GermanOpen2013             |          56 |        2 |
| BWOpen2012                 |          55 |        2 |
| PoznanOpen2009             |          55 |        2 |
| Euro2004                   |          53 |        2 |
| SpolsOpen2011              |          53 |        2 |
| BerkeleyWinter2012         |          53 |        2 |
| DanishOpen2011             |          52 |        2 |
| CzechOpen2012              |          50 |        2 |
| CubeCampKanazawa2012       |          47 |        2 |
| CubeCampKanazawa2011       |          45 |        2 |
| Rubikaz10thAnniversary2012 |          44 |        2 |
| LyonOpen2011               |          41 |        2 |
| MontpellierOpen2011        |          41 |        2 |
| DanishOpen2012             |          41 |        2 |
| RomanianOpen2011           |          41 |        2 |
| Altbier2012                |          38 |        2 |
| CzechOpen2013              |          37 |        2 |
| HelsinkiOpen2011           |          35 |        2 |
| CastellonOpen2011          |          34 |        2 |
| MontpellierOpen2010        |          33 |        2 |
| BasauriOpen2012            |          32 |        2 |
| MIROpen2013                |          31 |        2 |
| LaMontagneOpen2012         |          30 |        2 |
| BarcelonaWinterOpen2012    |          30 |        2 |
| WasedaBrainChallenge2012   |          28 |        2 |
| AlaniaOpen2010             |          28 |        2 |
| LyonSpringOpen2009         |          27 |        2 |
| ErfurtOpen2013             |          26 |        2 |
| DiamondBarSpring2013       |          26 |        2 |
| BlindOpen2011              |          25 |        2 |
| LyonSummerOpen2010         |          25 |        2 |
| FinnishOpen2010            |          24 |        2 |
| MagicCubeDays2012          |          24 |        2 |
| BasauriOpen2011            |          23 |        2 |
| RubikkuChamp2013           |          19 |        2 |
| ShanghaiSpring2011         |          17 |        2 |
| RenoWinter2010             |          16 |        2 |
| DanishSpecial2011          |          16 |        2 |
| GuangzhouFMC2012           |          16 |        2 |
| ZhengzhouOpen2011          |          12 |        2 |
| JavaFMCCubeDay2010         |          12 |        2 |
| WC2009                     |         327 |        1 |
| WC2011                     |         292 |        1 |
| Euro2012                   |         285 |        1 |
| AachenOpen2011             |         230 |        1 |
| Euro2010                   |         227 |        1 |
| WC2007                     |         214 |        1 |
| PolishNationals2013        |         181 |        1 |
| IndianCubeChallenge2013    |         179 |        1 |
| MPEIOpen2012               |         166 |        1 |
| AachenOpen2010             |         154 |        1 |
| JakartaOpen2010            |         152 |        1 |
+----------------------------+-------------+----------+

SELECT tmp2.competitionId, count(distinct personId) competitors, attempts
FROM
(SELECT competitionId, sum(attempts) attempts
FROM
(SELECT competitionId, roundId, max((value1<>0)+(value2<>0)+(value3<>0)+(value4<>0)+(value5<>0)) attempts
FROM Results
WHERE eventId = '333fm'
GROUP BY competitionId, roundId) tmp1
GROUP BY competitionId) tmp2,
Results
WHERE Results.competitionId = tmp2.competitionId
GROUP BY competitionId
ORDER BY attempts desc, competitors desc
LIMIT 100

 

[Lucas Garron]

Sébastien has written a proposal to make this change for 2014.

The WRC is unclear exactly how much of the community would support this, so I've added a poll to the first post. We'd love to see a lot of votes to help make a decision.

 

[Kirjava]

I worry I wouldn't get a chance to do a mean of 3

 

[kinch2002]

I would definitely like to see it in major championships. It's going to be very hard to schedule it in other competitions, so the rankings won't be a great reflection of anything. But the podiums in championships would be a lot more based on skill at FMC than luck and that's a great thing.

 

[Carrot]

I don't see how this would work... at all!

Look at this scenario:
You have a competition with FMC and you decide mean of 3? oh that sounds awesome, let's do that.
Competitors who are fighting for podium will now need to have stability and high success rate, fair enough.
Wait, the competitors needs to do 3 attempts at the same competition? spending up to 3 hours?
Well, so far so good, now we get to winner's ceremony "And as the only person to actually make a succesful mean of 3 we have John Doe!"

Guys, mean of 3 is fine, but I don't think we have enough fmc enthusiasts too make it a feasible format at most competitions. (best of 3 is NOT comparable to mean of 3 as best will make you able to say "ohh I think I'll stop here, because I already got a 29 in my previous attempt". I'm not saying mean of 3 is a bad format, I'm just saying I can't see how it would work around realistically speaking.

 

[kinch2002]

Time-wise, I don't see a problem. I would still only hold one attempt at UK comps, and I assumed that other comps would also. I see the mean of 3 as very useful for major championships only, where there is enough time.

With regards to DNFs, I don't see a problem. Writing stuff down correctly isn't hard imo. I fail to see how people write things down wrong. Even if you do, just check your solution.

My thoughts haven't changed in the last 4 months.

 

[Piotr]

I thinkt 30 minutes per cube for an experienced person is fairly enough. This will give 1.5h in total, which is comparable to the current 1h fmc attempt. Mean of 3 will also lower the impact of lucky solves I think. One will need to be really good at fm to acheive a decent result. Instead of something like 'hmm, let's try with fridrich. maybe i wll get LL skip and win'. I don't mean that a lucky solve is bad or something. But I think that's why in other events averages and means are being used instead of 'best of x' format.

 

[Kirjava]

It's going to be very hard to schedule it in other competitions, so the rankings won't be a great reflection of anything. But the podiums in championships would be a lot more based on skill at FMC than luck and that's a great thing.

Convinced me.

 

[Kit Clement]

I like the idea for championships, but not in general. I know that it has been mentioned that the FMC Mean rankings "wouldn't mean anything," but I still don't like having formats appear on a WCA profile that are so rare, as a significant amount of competitors simply wouldn't have the resources to attend a competition where an FMC Mean of 3 is offered.

Since awards can be given at the organizer's discretion, why not simply award the winners based on Mean of 3? I know the WCA competition results page will not reflect the actual award winners, but I think at a major championship with large prize money at stake, this will be the competitors' main goal rather than appearing on the WCA page.