In his musings on the statistical challenges of evaluating Thoroughbred pedigrees, sales figures, and the statistics supposed to represent trends and behavior related to them, the great statistical curmudgeon cooked up a method to satisfy his hunger for a number that would express a measure of the median and average in one.

Sidebar One — the maverage

The maverage is a statistic invented by this researcher to supplant the median and supplement the average. The traditional way of evaluating groups of sales foals is to use both averages and medians. That has a number of difficulties, the main one being what do you do with the two numbers once you have derived them? You can combine the two numbers somehow, but not in any statistically satisfactory manner. It is better to use only one number to try to correlate sales prices with results.

Here is how the maverage is calculated. For prices up to $10,000, simply divide by 100. Ergo, $1,000 = 10, $2,000 = 20, etc. For prices between $10,000 and $1,000,000, take the square root. Ergo, $10,000 = 100, $40,000 = 200, $90,000 = 300, etc.

The square root of $1,000,000 is 1000. The square root of $4,000,000 is 2000. It was decided to minimize prices above $1,000,000 even further. Prices above $1,000,000 are more about human egos than the actual value of any given animal. Was The Green Monkey really worth $16,000,000? The prosecution rests. Ergo, it was decided that $1,000,000 = 1000, $2,000,000 = 1100, $3,000,000 = 1200, $4,000,000 = 13000 (not the 2000 square root), etc.

The maverage rationalizes prices so that they can be utilized as averages without the distortion of the highest prices.  It is calculated like an average (just sum up and divide) but behaves somewhat more like a median. The Price Index in Table 6 and 7 is simply the maverage divided by the overall maverage for the entire group (150 on the nose).

For a more detailed discussion of the maverage, see my series on the Rasmussen Factor on this website last August.