3x3x3 WCA avg. times for everybody with <= 150s.

[JohanAR]

I made a histogram of all registered average times from World Cube Association as of 2013-04-14. There are probably better ways to represent this data but I've never used matplotlib before and it was late. Sharing in case anyone else finds this interesting. If anyone desperately wants a graph of some other cube size or single times I might be able to make it.

Spoiler

 

[moralsh]

So when you start cubing and you improve you're just becoming very common until you hit sub-19, from there your improvement makes you special

Liked it

 

[JO5561]

This is awesome, great job

 

[antoineccantin]

Interesting.

 

[Akiro]

What do the graduations on the left represent?

 

[uniacto]

What do the graduations on the left represent?

I think it's the number of people.

Nice graph by the way. It sort of motivates me to get faster so I won't be in the majority haha

 

[mark49152]

Nice graph, interesting! Thanks!

 

[Bestsimple]

Nice graph by the way. It sort of motivates me to get faster so I won't be in the majority haha[/QUOTE]

Or get slower...

 

[KrisM]

Awesome graph If it's not too much work, you should do 4x4 and 5x5 too!

 

[Iggy]

Wow nice work!

 

[PianoCube]

So, with an official average of 19.40, Im very normal
Then it's time to get sub 15 soon to get out of that group

 

[mark49152]

It would be interesting to see the same graph from 5 years ago to see how things have changed.

 

[JohanAR]

I think it's the number of people.

Correct. However, I think the width of each bar is quite unscientific.. IIRC I just told matplotlib to divide the data into 1000 groups (not all visible since I cut the xaxis at 150s). If I figure out how, I might change it so that each bar is 1s wide.

Awesome graph If it's not too much work, you should do 4x4 and 5x5 too!

I'll update the Python script so that you don't have to manually cut out the table, then I'll post those graphs and the script too if anyone wants to play around with it. Just don't expect too much user friendliness

It would be interesting to see the same graph from 5 years ago to see how things have changed.

Since that information is available on WCA's page it would be quite trivial to graph it. I'll see if I can draw both that one and the current times in the same plot.

 

[Ninja Storm]

This is of average times though. People who started when they averaged a minute and have just now gotten to the sub-20 or sub-15 marker will have their speed be ill-represented.

 

[JohanAR]

This is of average times though. People who started when they averaged a minute and have just now gotten to the sub-20 or sub-15 marker will have their speed be ill-represented.

It shows people's best average from any WCA competition, so it's not an average over a person's career.

Here are the requested plots:

3x3 average 2013 vs 2008

Spoiler

4x4 average

Spoiler

5x5 average

Spoiler

The last two look a bit wonky because of larger time interval and fewer registered times.

Here's the short script I've used to generate them. You have to look at the records page source and delete everything the lines with table data. Expect zero user-friendliness and no support

Spoiler

f = open('events.htm')
ls = f.readlines()
f.close()

def tim(st):
    b=st.find('class=\'R2\'>')+11
    e=st.find('<',b)
    c=st[b:e].find(':')
    if c >= 0:
        return 60*int(st[b:b+c]) + float(st[b+c+1:e])
    else:
        return float(st[b:e])

ts = map(tim, ls)

import matplotlib
matplotlib.use('Agg')
from matplotlib import pyplot
pyplot.figure(figsize=(12, 8), dpi=100)

xmax = 200

pyplot.hist(ts, bins=xmax, range=(0, xmax))
pyplot.xticks(range(0,xmax,10))
pyplot.yticks(range(0,800,100))
pyplot.savefig('avg_times.png')

 

[~Adam~]

It's just so difficult to get off the top of that graph for me =(

 

[Mike Hughey]

2013 vs. 2008 might be more interesting for 4x4x4 and 5x5x5 than they are for 3x3x3, due to the improvement in puzzle hardware.

 

[JohanAR]

I'll see what I can do.. I merged those graphs with GIMP so it's a little bit more work, and I might need to adjust some parameters to get something better for low number of samples. Doing 2x2 and OH too.

 

[Petro Leum]

imo it would also be interesting to merge the 2013 and 2008 graphs by standardising one of them so it matches the highest point of the other. do you know what i mean? by that it would be easier to see how much the community has improved.

 

[Lapinsavant]

Can you make a simulation of what would happen in a few years? Very interesting thought