World Record predictions in the year 2020?

wow how did I miss that?

I'm sorry

It's funny looking back and seeing how hilariously inaccurate the predictions are. They were talking about consistently sub-9 as a lofty expectation, and look where we are now: 3 people have gotten sub-6 averages. People then said Roux was finished at 12 seconds, and Sean-Patrick Villenauva got a sub-6

Tomk said:

CitricAcid said:

3x3: 5.55

And, the new WR holder who get's this record will
use ZZ. I CAN SEE INTO THE FUTURE!!! OOO!!!

Click to expand...

I beg to differ, they will use petrus...

Click to expand...

Well, Yusheng Du did technically use Petrus, if not intentionally. Looks like someone got the prediction right

ProStar said:

It's funny looking back and seeing how hilariously inaccurate the predictions are. They were talking about consistently sub-9 as a lofty expectation, and look where we are now: 3 people have gotten sub-6 averages. People then said Roux was finished at 12 seconds, and Sean-Patrick Villenauva got a sub-6



Well, Yusheng Du did technically use Petrus, if not intentionally. Looks like someone got the prediction right

Click to expand...

You can say that about any solve.

Zeke Mackay said:

You can say that about any solve.

Click to expand...

No, not every solve us an XXcross +EO skip, which was what Du got.

Etotheipi said:

No, not every solve us an XXcross +EO skip, which was what Du got.

Click to expand...

Here's a Petrus solve I just did to prove my point.

Zeke Mackay said:

Here's a Petrus solve I just did to prove my point.

alg.cubing.net

alg.cubing.net

Click to expand...

you just did a CFOP solve, not a Petrus solve. You say you can call any solve a Petrus solve, but I think your logic is that at the end of every solve, 2x2x2 is solved, and 2x2x3, and EO, and F2L, and LL. Petrus is a different method. If I did a solve with domino reduction, could you call it petrus?

Etotheipi said:

you just did a CFOP solve, not a Petrus solve. You say you can call any solve a Petrus solve, but I think your logic is that at the end of every solve, 2x2x2 is solved, and 2x2x3, and EO, and F2L, and LL. Petrus is a different method. If I did a solve with domino reduction, could you call it petrus?

Click to expand...

2x2 block with everything skip!

Saying a solve used a certain method when the solver does not use that method is stupid lol

I don't see any wrong in saying that technically his solve was petrus.

Mike Hughey said:

It would be outrageous to think someone is likely to get 60/60 by 2020, but I don't see it as outrageous by 2025.

Click to expand...

FYI

That70sShowDude said:

What do you guys think the WR's are going to be in the year 2020?

Click to expand...

Time flies

That70sShowDude said:

Time flies

Click to expand...

Two more days!

Tim Major said:

hubingjushi said:

2x2: 0.40 / 1.15
3x3: 3.42 / 4.97
4x4: 16.05 / 19.40
5x5: 34.53 / 40.91
6x6: 1:16.51 / 1:20.55
7x7: 2:05.05 / 2:10.38
OH: 5.64 / 8.27
BLD: 14.92 / 17.99
4BLD: 1:38.66
5BLD: 3:40.12
MBLD: 66/70 1:00.00
Pyraminx: 1.11 / 1.74
Megaminx: 30.62 / 33.90
Square-1: 5.70 / 7.72
Clock: 4.07 / 5.08
Skewb: 1.26 / 1.98

Click to expand...

I only see a couple of these happening ever, let alone in 5 years. Some of the averages just aren't going to ever happen. Rami's predictions are pretty good estimates I think

Click to expand...

Heh heh

ProStar said:

Heh heh

Click to expand...

most of those didn't happen though so he was right in the 5 years bit. Obviously they will probably all happen within the next 2 years so his I don't think these will ever happen is wrong.

I was talking about the "I only see a couple of these happening ever, let alone in 5 years. Some of the averages just aren't going to ever happen." Most of there have either been accomplished or are very close(within a second or two). And some of them are way higher than WR, like 7x7, 4BLD, and 5BLD

Competition Cuber said:

Okay, lets update a bit (single/average):

UPDATE 12-31-19: CURRENT RECORDS ARE IN BOLD
2x2: 0.43/1.21 0.49/1.21
3x3: 3.97/5.44 3.47/5.53
4x4: 18.01/20.72 17.42/21.11
5x5: 37.62/42.51 36.06/39.95
6x6: 1:16.51/1:24.65 1:13.82/1:17.10
7x7: 1:59.87/2:09.56 1:40.89/1:50.10
Skewb: 1.21/1.96 0.93/2.03
Squan: 5.26/7.48 4.95/6.54
Pyra: 0.98/1.78 0.91/1.86
Mega: 28.74/31.51 27.22/30.89
3x3 OH: 6.65/9.87 6.82/9.42
3x3 WF: 20.57/23.56 (IDK how the transition to average of 5 will affect his though) 15.56/19.90, but it is being removed as an event.
3x3 FMC: 18/23.33 16/22.00
3x3 BLD: 16.82/21.54 15.5/18.18
4x4 BLD: 1:19.87 1:02.51/1:08.76, Means are now recognized.
5x5 BLD: 3:31.56 2:21.62/2:27.63, Means are now recognized.
3x3 multi: 45/46 59:59.71 59/60 in 59:46
Clock: 3.38/4.98 3.29/4.38

EDIT March 14, 2018: changed one of these, and I will no longer update these predictions.

Click to expand...

As the WCA's 2020 regulations are officially in effect, might as well see how we did. I haven't edited these since March 14, 2018 aka Pi Day.

2x2: Single: Kinda surprised that the single didn't drop since 2016 Average: RIGHT ON THE DOT!
3x3: Single: Shocked at the 3.47 single. Congrats to Yusheng. Average: Close, considering that the current WR is 5.53. Only 0.09 Off.
4x4: Single: ~Half a second off. Not great, not terrible either. Average: 0.39 off. I'll take it.
5x5: Single: Decently far off. Average: Even farther off.
6x6: Single: Almost 3 seconds off lol. Average: Way off.
7x7: Single: Waaaaay off. Average: Almost 20 seconds off lol.
Side note: Big cubes really progressed a lot over the past couple years. I am positively surprised at how fast they are getting.
Skewb: Single: Pretty far off for being such a fast event. Didn't expect the single to be sub-1. Average: Less then 0.1 off (0.07), I'll take it.
Squan: Single: Jackey's single really surprised me. I did NOT expect the single to be sub-5 by 2020. Average: Similar story.
Pyra: Single: Kinda close. Average: Again, kinda close. Margin of <.1 on both.
Mega: Single: Meh, over a second off. Average: Also over a second off.
3x3 OH: Single: Kinda thought that it would drop a little more. Average: Kinda thought that it would drop a little less.
3x3 WF: Single and Average: Feet got a LOT faster in the past couple of years, just like big cubes. I was way off on both of these.
3x3 FMC: Single: Pretty far off. Average: Little closer, but still pretty far off.
3x3 BLD: Single: Was I close? Kinda sorta not really. Average: Over 3 seconds off.
4x4 BLD: Single: This event improved tremendously. I was clearly way off on my predictions. Average: Was not counted when I wrote these predictions.
5x5 BLD: Single: Over a minute off on the single lmao. Average: Was not counted when I wrote these predictions.
3x3 Multi: Single: Was way off on the points and number of cubes solved. At least I was kinda close on the time it took?
Clock: Single: Actually pretty close! Average: Not really close.

Events I was most wrong about: 6 and 7, 4 BLD and 5 BLD, Multi.
Event I was closest to: 2x2 average, also pretty close on 3x3.

turtwig said:

I'll jump on this bandwagon and see how long these take to break...

3x3: 3.61/5.23 (Single broken by Yusheng Du on 11/24/18)
2x2: 0.39/1.12
4x4: 18.63/22.84 (Both broken by Max Park on 5/27/18)
5x5: 35.49/39.49 (UWR is 32.70/36.40)
6x6: 1:09.91/1:20.12 (Average broken by Max Park on 7/15/18)
7x7: 1:48.92/1:58.10 (Both broken by Max Park on 7/14/18)
3BLD: 15.50/18.67 (lol hope the WR single stays what it is)
FMC: 16/21.67 (Single tied by Sebastiano Tronto on 6/15/19)
OH: 5.75/8.52 (UWR is 5.06/7.36)
Feet: 15.26/22.43 (Average broken by Daniel Rose-Levine on 7/19/18)
Clock: 2.98/3.99 (UWR is 2.15/3.55)
Mega: 26.12/30.91 (Average broken by Juan Pablo Huanqui on 8/11/19)
Pyra: 0.81/1.61 (UWR average is 1.13)
Skwb: 0.87/1.70 (UWR average is 1.27)
Sq1: 4.67/6.99 (Average broken by Vicenzo Guerino Cecchini on 10/20/18)
4BLD: 1:13.23 (Broken by Stanley Chapel on 4/28/19)
5BLD: 3:29.41 (Broken by Stanley Chapel on 12/1/18)
MBLD: 59/60 59:52 (Broken by Graham Siggins on 11/9/19)

Click to expand...

Current record better than prediction: 3x3 Single, 4x4 Single, 4x4 Average, 6x6 Mean, 7x7 Single, 7x7 Mean, 3BLD Mean, Feet Average, Megaminx Average, Sq1 Average, 4BLD Single, 5BLD Single, MBLD Single.

So 13 out of 33 records are better than I expected. It's worth noting that Stanley Chapel's 4BLD and (especially) 5BLD means are both better than my single predictions.

Current record worse than prediction: 3x3 Average, 2x2 Single, 2x2 Average, 5x5 Single, 5x5 Average, 6x6 Single, OH Single, OH Average, FMC Average, Feet Single, Clock Single, Clock Average, Megaminx Single, Pyraminx Single, Pyraminx Average, Skewb Single, Skewb Average, Sq1 Single.

18 out of 33 records are worse than my prediction.

Predictions that were exactly right: FMC Single, 3BLD Single

Obviously FMC Single is the easiest record to predict (it's pretty cool that someone managed to get 16 though). A lot of luck on the 3BLD prediction though. Also worth noting my MBLD prediction was only off by 6 seconds.
BigBLD was probably the most unexpected ( average absolute error for singles drops to ~7% without 5BLD). I doubt anyone else is gonna do calculations like this, but I wonder how accurate my predictions were compared to others. I feel like I got closer on many events than I expected. Regarding the predictions that have not been broken yet, the current UWRs are all better than those predictions (couldn't find 2x2, Pyra, Skewb UWR Singles but I doubt people care very much about those), so I think it is likely that all of my predictions will be broken relatively soon (perhaps in 2020?)

EDIT: It's pretty interesting reading some of the posts that happened around the same time as mine. I think the general trend with these threads is that people are pessimistic. Especially with big cubes and bigBLD basically everyone (including me for 5BLD) was way above what the records actually turned out to be. It's also interesting looking back at posts saying that so-and-so result will never happen when the current record is lower than it.

turtwig said:

Current record better than prediction: 3x3 Single, 4x4 Single, 4x4 Average, 6x6 Mean, 7x7 Single, 7x7 Mean, 3BLD Mean, Feet Average, Megaminx Average, Sq1 Average, 4BLD Single, 5BLD Single, MBLD Single.

So 13 out of 33 records are better than I expected. It's worth noting that Stanley Chapel's 4BLD and (especially) 5BLD means are both better than my single predictions.

Current record worse than prediction: 3x3 Average, 2x2 Single, 2x2 Average, 5x5 Single, 5x5 Average, 6x6 Single, OH Single, OH Average, FMC Average, Feet Single, Clock Single, Clock Average, Megaminx Single, Pyraminx Single, Pyraminx Average, Skewb Single, Skewb Average, Sq1 Single.

18 out of 33 records are worse than my prediction.

Predictions that were exactly right: FMC Single, 3BLD Single

Obviously FMC Single is the easiest record to predict (it's pretty cool that someone managed to get 16 though). A lot of luck on the 3BLD prediction though. Also worth noting my MBLD prediction was only off by 6 seconds.

Event	Predicted Single	Current Single	Error	Prediction would rank	Predicted Average	Current Average	Error	Prediction would rank
3x3	3.61	3.47	4.03%	2	5.23	5.53	-5.42%	1
2x2	0.39	0.49	-20.41%	1	1.12	1.21	-7.44%	1
4x4	18.63	17.42	6.95%	4	22.84	21.11	8.20%	4
5x5	35.49	36.06	-1.58%	1	39.49	39.65	-0.40%	1
6x6	1:09.91	1:13.82	-5.30%	1	1:20.12	1:17.10	3.92%	2
7x7	1:48.92	1:40.89	7.96%	2	1:58.10	1:50.10	7.27%	2
3BLD	15.50	15.50	0.00%	1	18.67	18.18	2.70%	4
FMC	16	16	0.00%	1	21.67	22.00	-1.50%	1
OH	5.75	6.82	-15.69%	1	8.52	9.42	-9.55%	1
Feet	15.26	15.56	-1.93%	1	22.43	19.90	12.71%	4
Clock	2.98	3.29	-9.42%	1	3.99	4.38	-8.90%	1
Mega	26.12	27.22	-4.04%	1	30.91	30.39	1.71%	2
Pyra	0.81	0.91	-10.99%	1	1.61	1.86	-13.44%	1
Skewb	0.87	0.93	-6.45%	1	1.70	2.03	-16.26%	1
Sq1	4.67	4.95	-5.66%	1	6.99	6.54	6.88%	3
4BLD	1:13.23	1:02.51	17.15%	2	N/A	1:08.76	N/A	N/A
5BLD	3:29.41	2:21.62	47.87%	4	N/A	2:27.63	N/A	N/A
MBLD	59/60 59:52	59/60 59:46	-0.003%	2	N/A	N/A	N/A	N/A
Average Absolute Error	N/A	N/A	9.19%	N/A	N/A	N/A	7.09%	N/A
Average Error	N/A	N/A	0.14%	N/A	N/A	N/A	-1.30%	N/A

BigBLD was probably the most unexpected ( average absolute error for singles drops to ~7% without 5BLD). I doubt anyone else is gonna do calculations like this, but I wonder how accurate my predictions were compared to others. I feel like I got closer on many events than I expected. Regarding the predictions that have not been broken yet, the current UWRs are all better than those predictions (couldn't find 2x2, Pyra, Skewb UWR Singles but I doubt people care very much about those), so I think it is likely that all of my predictions will be broken relatively soon (perhaps in 2020?)

Click to expand...

Wow, you put effort into this. How do you calculate your percentage error? And what do Average Absolute Error and Average Error mean?

Competition Cuber said:

Wow, you put effort into this. How do you calculate your percentage error? And what do Average Absolute Error and Average Error mean?

Click to expand...

For percentage error, I did (Prediction-WR)/WR (for example, (3.61-3.47)/3.61=0.0403). I think this should be consistent with normal percentage error which is (experimental-theoretical)/(theoretical). Positive error means the current record is lower than my prediction and negative error means it is higher. Absolute error means I take the absolute value of the error (ex. |-2|=2), so it measures the "distance" between my prediction and the WR. I think average absolute error is a better measure of how accurate results are since it measures the average amount my prediction was off by (so on average my predictions were off by 7%-9%). Just taking the average of the percentage errors means that positive and negative values can cancel each other out (which is why they are much smaller).

It's now 2020. I am a little late.