With the final bids for the painting and sketch now in hand, we can finally announce a winner for the Mother of Runes Auction Guessing Game.
First, let’s take a look at the results of the auctions:
Definitely a strong result for these auctions.
While it appears my $15,500 guess was a bit off (okay, way off), I’m not too concerned as I knew it was an aggressive and slightly overly hopeful one.
To say that 683 entries for the Guessing Game was a bit more than I expected is an understatement. You all blew me away with your enthusiasm!
With that settled, it’s time to announce the winner!
and the winner is...
Matt Nicholson with a guess of $7,098
Congratulations to the winner of the Guessing Game!
Arrangements are being made to send out the prize package from Terese, so check your email if you haven’t already.
In case you missed it, take a look at what the winner will be receiving:
Terese put a lot of effort into these prizes and I’m certain that they will be appreciated by the winner.
Sadly, no one guessed the exact closing bid of either of the auctions to win the secondary prizes. Don’t fret, the Artist Proofs from the OMA Collection will be saved for the next Guessing Game.
To get a better idea of the guesses sent in for this contest, I handed over the guessing data to resident OMA statistician, Ryan Sittler, to crunch the numbers and see what he could uncover.
OMA Statistician Report
OK, I ran the data through a program called SPSS, basically the most hardcore tool around for data analysis.
The first thing it generated was an overview of the data. Some of the raw numbers I found interesting as well as what can be inferred from the data.
I don’t know how much you know about stats so I’ll explain some of what this means. I realize you may well know all of this already; I promise I’m not being pedantic.
The mean is the arithmetic “average” that we usually hear people talking about.
Median is the value that falls dead center among all entries (so 50% of the actual guesses are above it and 50% of the actual guesses are below it).
Mode is the value that appears most frequently. In the case of the composite guesses, three values had the same number of guesses – that’s why it’s called trimodal.
Standard deviation is essentially how many numbers you would have to jump up or drop down from one value to the next, in order to see things evenly spaced out (that’s a bad way of describing it and not entirely accurate… but it’s the easiest way to wrap your brain around it).
When you look at mean, median, and mode together you can start to see skewness and kurtosis of the distribution. Ideally, we always want data to fall into a perfect Bell Curve. It rarely happens. Skewness tells you how off-center the data are (to the right or left of the curve) and kurtosis tells you how “peaked” the data are.
The normal person on the street really won’t care about anything other than Low, High, and Mean values. The other stuff is for stats freaks… or people that want to make inferences.
I like to make inferences, so I went through and did what’s called a Z-score analysis to see where the outliers were at in the data. The person that guessed $213,017 for the combined total, for example, was either completely out to lunch or just trolling. It throws off the data.
Z-scores can show me where the natural breaks are in the data… and what can be cut. I didn’t do this perfectly, but for our purposes, it should be fine. I was fairly conservative in what I cut out. It would have been completely OK to cut more deeply, but I didn’t need to.
After outliers were removed, we have the following results:
This time the SD dropped a lot (though it’s still pretty large). More importantly, the skewness and kurtosis are almost non-existent compared to what they were. This is a much more normal distribution and a better look at what the participants socially constructed as a guess for each value.
The painting, for example, had a mean value of $8,317.07 before the outliers were removed (conservatively). Now the mean value is $7,898.61 – much closer to the actual selling point. If I was more liberal with what I cut, I suspect the mean would be closer to $6500-$7000. Not far at all from where it actually settled.
So, what does all of this mean? That’s always up for interpretation, but here’s what I can infer from the data:
- People, in general, overvalued with their prediction. I can say this confidently because you had so many entries. If it was like 20 people… I wouldn’t be so confident. But with 600+ – especially when there was an incentive to be correct
- It’s clear to me: the average person is not good at guessing art values under these circumstances
I’m sure I’ve given you more information than you can use. But, hopefully there is something here that will pique your interest.
Thanks to everyone that made an entry to the contest, put in a bid on the auctions or spread the word about either.
Keep reading OMA to get ready for the next Auction Guessing Game!
Did you enjoy this Guessing Game? Is there anything I can do with future games to make them better?
Please fill out this quick poll to give me some valuable feedback. (You can select more than one option)