Warning: Math (and the CGC Census) 0

[valiantman]

Question: What does the CGC Census have in common with Y = 0.00000000505*X^4-0.0000585*X^3+0.250095*X^2+5.88495*X?

Answer:  They're the same.

X = number of days since January 1, 2000

Y = predicted number of CGC slabs on the CGC Census

If the 20+ year formula holds up, the CGC Census will have these totals in the future:

The obvious problem is that the next 4 years would need to have the same number of graded comics as the prior 21 years combined.

Is CGC ready for that kind of growth?

Edited July 16, 2021 by valiantman

[SeanYork]

Can you use this to predict the length of the Turn Around Times that far out?

[valiantman]

3 minutes ago, SeanYork said:

Can you use this to predict the length of the Turn Around Times that far out?

Nope, but we'll clearly be able to see whether the service can keep up with the pace.

[bc]

 Thanks for this - Very cool! 

So using the data for July 31, 2021 as an example, would this be true?

 

Number of Books Slabbed per Day = (CGC Census Actual) / (X) Days

Number of Books Slabbed per Day = 6,377,537 / 7863

Number of Books Slabbed per Day ~ 811

 

If that equation is viable, CGC slabbed an average of +9% more books every year from 2001 to July 31, 2021.
 

The projection for the next 4 years suggest that will increase to +14% more books per year with a pace needed of about 1428 books a day. That's a 43% increase over the production output at July 31, 2021.

That would probably require significant capital outlays for both equipment and human resources 

 

Hello Blackstone!

-bc

[GeeksAreMyPeeps]

I think the rate of growth in the number of books slabbed probably corresponds somewhat to the amount of money entering the hobby. (I don't have data on this, but I think common sense would indicate that the growth of the number of books that it's financially worth slabbing would clearly be affected by this.) Since I don't expect the amount of money entering the hobby to continue that growth rate, I don't think that the projections will hold. And actually where reality breaks with the projections (if it does) may be a good point to reference, if the hobby experiences a downturn, to see if those two points are simultaneous.

[valiantman]

50 minutes ago, bc said:

 Thanks for this - Very cool! 

So using the data for July 31, 2021 as an example, would this be true?

 

Number of Books Slabbed per Day = (CGC Census Actual) / (X) Days

Number of Books Slabbed per Day = 6,377,537 / 7863

Number of Books Slabbed per Day ~ 811

 

If that equation is viable, CGC slabbed an average of +9% more books every year from 2001 to July 31, 2021.
 

The projection for the next 4 years suggest that will increase to +14% more books per year with a pace needed of about 1428 books a day. That's a 43% increase over the production output at July 31, 2021.

That would probably require significant capital outlays for both equipment and human resources 

 

Hello Blackstone!

-bc

Yes, the 811 slabs per day is an average, but that ~811 number was probably only true around the start of 2015.

You can calculate Y/X all along the timeframes, and if you isolate the difference from one year to the next, you'll see that CGC started around 161 slabs per day and they're currently at 3,000 slabs per day.

[valiantman]

25 minutes ago, GeeksAreMyPeeps said:

I think the rate of growth in the number of books slabbed probably corresponds somewhat to the amount of money entering the hobby. (I don't have data on this, but I think common sense would indicate that the growth of the number of books that it's financially worth slabbing would clearly be affected by this.) Since I don't expect the amount of money entering the hobby to continue that growth rate, I don't think that the projections will hold. And actually where reality breaks with the projections (if it does) may be a good point to reference, if the hobby experiences a downturn, to see if those two points are simultaneous.

The numbers so far are fascinating, since they match 95%+ to the actuals from the start of CGC (when people were very skeptical) through the change in leadership (2008) and the ups and downs of the hobby, all the way to the explosion of slabbing (doubling the total CGC Census in the past 5.5 years).  I'm not sure whether the downturn you're suggesting over the next couple of years will be drastic enough to skew too far away from the projections since the numbers have already worked for 21+ years.

[Cat]

Does it take into account days when they have a pizza party? 

[valiantman]

Just now, Mecha_Fantastic said:

Does it take into account days when they have a pizza party? 

Yep, that's why accuracy is at 99.93% instead of 100%. 

[Cat]

1 minute ago, valiantman said:

Yep, that's why accuracy is at 99.93% instead of 100%. 

 

Astounding. 

[DWL]

That is very interesting!  Thank you for looking at the numbers.  I think your later post suggesting you look at the numbers from a year to year perspective might provide more insight.  For example, if those numbers are correct, they would suggest the number of books on the census increased by 928,217 between 2019 and 2020 (5,799,679 - 4,871,462 = 928,217).  If you look at the period 2010 to 2011, the actual number of books on the census increased by 162,071 (1,701,423 - 1,539,352 = 162,071). 

This would suggest an increase in the number of books on the census of 19.1% between 2019 and 2020 but a 10.5% increase between 2011 and 2011.

The really interesting thing is that the number of days does not change, at least in theory.  I guess people could work weekends but if we assume approximately 270 working days each year, that would suggest CGC slabbed approximately 600 books per day during the period 2010 to 2011 but  slabbed 3,437 per day (!!) during 2019 - 2020.  If that number is correct, perhaps this also explains why there seems to be so many quality control issues people are highlighting lately.  3,400 books per day does not seem realistic to me but I guess it might be possible.

Anyway, just my   Thanks for this, take care.

Edited July 16, 2021 by DWL
Minor change in text.

[BitterOldMan]

Before I retired, I worked as the statistician for a Fortune 500 company.   When I testified as an expert witness, the plaintiffs always hired the Chairman of the Statistics Department at Stanford.   
 

What we have is fitting a polynomial to nonsensical year, which provides a good fit, but provides zero predictive ability.

Knock yourselves out and play statistician, just don’t take the predictive power seriously.  You could fit a spline model and obtain a perfect fit.

[DWL]

4 minutes ago, BitterOldMan said:

Before I retired, I worked as the statistician for a Fortune 500 company.   When I testified as an expert witness, the plaintiffs always hired the Chairman of the Statistics Department at Stanford.   
 

What we have is fitting a polynomial to nonsensical year, which provides a good fit, but provides zero predictive ability.

Knock yourselves out and play statistician, just don’t take the predictive power seriously.  You could fit a spline model and obtain a perfect fit.

Yes indeed.  This reminds me of that old joke about econometric models finding a non existent black cat in a dark room. 

But, if you look at the actual change and percentage change from year to year, there is something here.  The growth in the number of slabs CGC appears to be slabbing per day seems, well, intriguing. 

[valiantman]

8 minutes ago, BitterOldMan said:

Before I retired, I worked as the statistician for a Fortune 500 company.   When I testified as an expert witness, the plaintiffs always hired the Chairman of the Statistics Department at Stanford.   
 

What we have is fitting a polynomial to nonsensical year, which provides a good fit, but provides zero predictive ability.

Knock yourselves out and play statistician, just don’t take the predictive power seriously.  You could fit a spline model and obtain a perfect fit.

Exactly, we're looking at the polynomial that approximates 21.5 years of CGC grading totals.  Its predictive ability is greater than zero, but approaches zero at time increases.

Zero predictive ability would mean that tomorrow the CGC Census total could be 7 or 7 billion, even though the formula would predict closer to 6.4 million.

We can be quite confident that we're better than zero on predictive ability, and we can be quite confident that your selected user name is appropriate. 

Edited July 16, 2021 by valiantman

[Math Teacher]

2 hours ago, valiantman said:

Question: What does the CGC Census have in common with Y = 0.00000000505*X^4-0.0000585*X^3+0.250095*X^2+5.88495*X?

Now you're talking my language. Very impressive statistical analysis.

[Cliff R.]

Doesn't a 4th degree polynomial have 4 solutions?  

 

[Math Teacher]

15 minutes ago, BitterOldMan said:

Before I retired, I worked as the statistician for a Fortune 500 company.   When I testified as an expert witness, the plaintiffs always hired the Chairman of the Statistics Department at Stanford.   
 

What we have is fitting a polynomial to nonsensical year, which provides a good fit, but provides zero predictive ability.

Knock yourselves out and play statistician, just don’t take the predictive power seriously.  You could fit a spline model and obtain a perfect fit.

May I make a suggestion? How about we compare the predicted amount with the actual number at the end of 2021? Then we can discuss the predictive ability.

[ADAMANTIUM]

2 minutes ago, Cliff R. said:

Doesn't a 4th degree polynomial have 4 solutions?  

 

After that, dont you pick the one closest that's not derivative?

[ADAMANTIUM]

Or is it asymmetric

I forget been too long

And I think I only took intro to statistics haha

Edited July 16, 2021 by ADAMANTIUM

[valiantman]

9 minutes ago, Cliff R. said:

Doesn't a 4th degree polynomial have 4 solutions?  

It has four solutions for Y = 0.

That would be when the CGC Census total was zero... and I'll admit, the formula should probably be useless when the CGC Census total was zero, except that it also works then, too.

When X = 0 days, Y = 0 census.

It also results in a zero census when X = -23.402 days, 5,803 + 4,014i, and 5,803 - 4,014i.

Besides 0 days, it works with negative 23 days, 15.9 years + 11 imaginary years, and 15.9 years - 11 imaginary years.

Math is fun.