Speed-Cubjectives: The tool that computes statistics-based speedcubing objectives

[kotarot]

Hello,

I've created a tool named "Speed-Cubjectives",
that automatically generates speedcubing objectives (=cubjectives).
Once you input your record or your WCA ID, you'll get list of computed objectives for all WCA events.

For example, if you input 15.0sec as record of Rubik's cube, it tells that you are likely to solve 4x4 in 1:14.75, 5x5 in 2:22.93, ... and so on. These can be your speedcubing objectives.
These estimated values are derived from the statistics based on the WCA database.
You can also view charts of relationship between WCA events.

I believe that there has been no dynamic statistics tools like this.
Give it a try!

http://www.terabo.net/cubjectives/

 

[AlphaSheep]

Interesting tool. It's estimated times are very close to my actual times for 4x4, 5x5, blindfolded, one-handed, pyraminx, square-1 and clock. I'm much better at 2x2, skewb, and fewest moves than it says, but a little slower at 6x6, 7x7 and megaminx than it says I should be... Not surprising that those match the events I practice most and least often.

 

[SpeedCubeReview]

So are the comparable times what my average "should" be or my record?

 

[obelisk477]

Depends on what you input. If you input your global average, it should output global averages. If you input PB single, it will ouput something more like PB single

 

[lerenard]

This is so cool! Apparently I suck at 3x3... Because I put in my global average and it gave me times that were too high for everything...

 

[StachuK1992]

Neat!
Any way you could use all of the user's times in the estimates, rather than just 1 event?

Edit: my want might not be obvious.

From all of my official WCA stats, tell me what I'm doing WELL in and what I'm doing POORLY in.

So,
- get some base 'score' of me as an overall cubers
- order the events I participate in from my "best" event to my "worst" event, probably with some color-coding.

 

[Hssandwich]

Apparently I should average 15.5 on Skewb, 1:10 on 4x4 and 12.99 on Pyraminx after putting in my global average.
I average 5.3, 58 and 5.5 respectively

 

[lerenard]

Neat!
Any way you could use all of the user's times in the estimates, rather than just 1 event?

Edit: my want might not be obvious.

From all of my official WCA stats, tell me what I'm doing WELL in and what I'm doing POORLY in.

So,
- get some base 'score' of me as an overall cubers
- order the events I participate in from my "best" event to my "worst" event, probably with some color-coding.

I like this idea.

 

[Mike Hughey]

I'm very disappointed to discover that 3x3x3 speedsolve is no longer my worst event. By a slight margin, skewb is. I may have to practice skewb some more to fix that. (Actually, I suspect all I need to do is get a decent skewb and that would probably take care of it; I still use a rather bad LanLan.)

At least 3x3x3 speedsolve is by far my second worst event. I actually average better than my predicted PB in all the other events.

 

[okayama]

Great tool!

Quick question about the records:

single for fmc/333bld/444bld/555bld/mbld, and average/mean for the rest?
Or single for only 444bld/555bld/mbld?

 

[DGCubes]

Rubik's Cube 15.51
4x4 Cube 1:17.18: I average around 1 minute.
5x5 Cube 2:27.15: About my average; I'm a little faster.
2x2 Cube 6.26: I'm more 4-5 range.
Rubik's Cube: Blindfolded 3:46.48: Around 3 minutes
Rubik's Cube: One-handed 34.87: Around 40-45 seconds.
Rubik's Cube: Fewest moves 42.91 moves: IDK
Rubik's Cube: With feet 2:49.30: Pretty dang good.
Megaminx 2:32.59: A little less.
Pyraminx 13.44: 5 second average... but Pyra is my favorite event.
Square-1 56.11: Yup yup.
Rubik's Clock 25.02: IDK
Skewb 16.18: Sub-10.
6x6 Cube 4:32.65: Yeah.
7x7 Cube 6:31.44: My PB is 8:59...
4x4 Cube: Blindfolded 13:09.13: IDK
5x5 Cube: Blindfolded 24:38.82: IDK
Rubik's Cube: Multiple Blindfolded 3.14 cubes: Definitely possible, but 2/2 is my best (3/3 almost happened once.)

Overall, this was quite impressive, even though a bit inaccurate. Good job!!

 

[kotarot]

Thanks for the many comments!

So are the comparable times what my average "should" be or my record?

The estimated times are not your real time.
You may have the ability to solve them in the shown times.
It depends on you that you aim the shown times or you just treat them as references.

 

[kotarot]

Neat!
Any way you could use all of the user's times in the estimates, rather than just 1 event?

Edit: my want might not be obvious.

From all of my official WCA stats, tell me what I'm doing WELL in and what I'm doing POORLY in.

So,
- get some base 'score' of me as an overall cubers
- order the events I participate in from my "best" event to my "worst" event, probably with some color-coding.

Nice idea!
This might be my next work.

 

[kotarot]

Quick question about the records:

single for fmc/333bld/444bld/555bld/mbld, and average/mean for the rest?
Or single for only 444bld/555bld/mbld?

"single for fmc/333bld/444bld/555bld/mbld, and average/mean for the rest" is true.

I basically wanted to use "average" if possible.
Although, the number of average data in fmc/333bf is too small to use for statistics.

 

[okayama]

"single for fmc/333bld/444bld/555bld/mbld, and average/mean for the rest" is true.

I basically wanted to use "average" if possible.
Although, the number of average data in fmc/333bf is too small to use for statistics.

I see, thanks! It would be nice if that is explained in the page (e.g. in "Notes").

 

[kotarot]

I see, thanks! It would be nice if that is explained in the page (e.g. in "Notes").

Thanks for your advice and I added it in the notes.

 

[cmhardw]

Wow, if I understand correctly then I am really bad at 3x3 for my level of skill at BLD.

Inputting my WCAID I am:
62.5% faster than the predicted 3BLD
63.9% faster than the predicted 4BLD
47.5% faster than the predicted 5BLD

If I were 63% faster on the 5BLD (for consistency's sake) that would be a time of 8:58.84 on 5BLD. That seems a little fast to me, but not out of my reach. This gives me a goal to strive for on 5BLD

Thanks for making this site!

 

[ottozing]

According to this, my skewb times should be over 10 seconds slower.

Dank.

 

[Kit Clement]

As a statistician, using simple linear regression for (in most cases) data with increasing variation is really making me uncomfortable. Results would be much better with a GLM or even just log transformations on the variables.

 

[Laura O]

Well, "correlation does not imply causation", so I don't like that the results are called "objectives".

But I'm quite confident that I can meet most of the objectives: being 20 seconds slower in 4x4, 30 in 5x5, 1 in 2x2, 2 minutes in 3BLD, 7 seconds in OH, 21 moves in FMC, 5 seconds in Megaminx, 6 seconds in Pyraminx, 30 in Square-1, 18 in Clock, 30 in 6x6 and 13 in 7x7 won't be that hard.