Monday, 1 February 2016

From MilesR: Week 6 Statistical Update - Buckle Up baby, It's Stats Time

I know, I know you're all saying not more Stats from Miles, please make it it stop! I decided to humour him because he has the power to affect both my credit rating and my Challenge points Total. Plus he sets my heart a flutter when he uses the word statistics. So going with the pull the bandaid off quickly school of thought, here's Miles.....

We're at the midway point with 6 weeks in the books and 6 weeks to go with the Analogue Hobbies Painting Challenge VI or "Seis Grande" as it is known on the "streets".

I must warn you to spread out plastic sheets and don protective clothing as you are all about to have your minds blown - no really it makes a big mess and I don't want to hear any complaints of be responsible for dry cleaning bills.

So what's the big development on Challenge hump day?  Participation rates eked up a bit, with 6 intrepid souls joining the ranks of honour after submitting their first item and raising the participation level from 61 to 67 or 76%   I think we can get to 100% but it will require just a little bit more effort out there.

Perhaps the most surprising thing is that we didn't see a drop off in production during weeks 5 and 6 - a total of 128 submissions for 11,924 points were submitted during the last two weeks bringing our total points level to 30,252 points.  Just for perspective, during weeks 3-4, we had 133 unique submissions for 11,936 points - I'd say that was a pretty level amount of painting.

In terms of projected points, we saw a few participants raise their points targets (brave and worthy souls they are) and we now have a total points target level of 70,624 from 69,290.  We again wheeled out the Acme Point-Projectometer to estimate what the ending results might be.  Kids - projecting points is a very difficult and dangerous amalgamation of science, magic and financial services industry ethics to derive it's outcome.  Do not try this at home.

The projectometer was fed in the results through week six, along with a few kittens (we ran out of bats due to Curt's interesting way of expressing frustration) and after some serious gear grinding out popped the result of 70,264 points projected for Seis Grande.  It's actually very close to the week 4 projection of 69,290.  Of course all projections have a error factor of up to +/- 100%.

In terms of what's being painted, 28mm still holds the dominate position, accounting for 63% of the total points through week six.  That's down from 67% at week 4 as we saw 15MM stage a mini rally by more than doubling it's point total to go from 3% to 7% of total points.

So with out further ado, I present to you, dear reader, the irrepressible, the irreplaceable, the unintelligible, the astounding Week 6 Statistical table

Weeks Completed246
Weeks Remaining1086
Participants w/ a Submission416167
Total Submissions98231359
Total Points6,39218,32830,252
Participation Rate47%69%76%
Average per Submitter156300452
Avg per Submission657984
Submitter % of Target Completed16%34%51%
Projected Points
Remaining Bouns Rounds12,30010,2507,750
Non Submitters @ 25%9,1594,0562,370
Projected Total Points59,81169,29070,624
Figures Painted
Cav / Art34140326
Cav / Art434115
Cav / Art106180234
Score Distribution
Bonus Round "Bonus Points"0%11%11%
Other Scales8%13%15%

To assist in your perusal of the statistical beauty of the chart above I have included the following musical interlude to set the right tone.  I think you will be "Thunderstruck"

My god Miles, I must be running a high fever indeed.  That almost made sense.  perhaps you'd like to move on to explain other mysteries of the universe to us?  Like why Donald Trump leads the polls?  Or why Iowa Democrats choose their presidential nominee by standing in opposite corners of a room and shouting at each other?    And AC/DC?  On a banjo?  Have you no sense of decency?  Do you not believe in the Geneva Convention?  As a side note, I used that infographic in one of my classes, and had one team of my students work out the survival function for la Grande Armee (the other did it for Anson's Circumnavigation).
Anyway I think I'll wander down to the local water hole for the amateur axe throwing competition.  That outta dull the pain in my brain one way or another.


  1. Interesting stuff Miles and thanks for preparing all the statistical and arithmentalic gobbledigook for us.

    However, being one of these non-statistical analyst types, I believe I have spotted a small fault in the programming of the "black box of wonders". Surely the Challenge is 13 weeks, not 12? Does this mis-programing of the data mean that we will suffer an irreversible meltdown and bring on the nuclear winter of discounted tents? ;)

    1. Details, details
      Never let the facts get in the way of statistical significance

  2. Next up multivariant analysis. I work a lot with huge geochemical data sets that are not normally distributed - think ski jump ramp not bell shaped hill. The cool and simple way to handle this is to use Median Absolute Deviation, aka MAD. Sums up both stats and the challenge

  3. I'm not in favour of anything I can't adjust with a hammer, stats included.

  4. Excellent work here and I think that should allow us to identify the key growth drivers and push the challenge down to the rain makers who can then run with the ball whilst pushing the envelope to get the ducks in a line and singing from the same hymn sheet.

  5. To paraphrase "Glenngarry Glenross"
    - Coffee is for AHP posters