By Max Breitenbach - www.silverdollarco.net for CoinWeek ……….
For a long time, I’ve been wondering what the best times for buying and selling coins on eBay are. There have to be certain days of the week and times of the day where bidding activity is higher—obviously there won’t be the same number of people bidding on coins at 2:00 in the morning as there will be at 5:00 in the evening. But setting aside those common sense rules, when exactly the best time for buying and selling coin ? A couple months ago, I decided to start collecting data from eBay’s completed listings in order to find the most optimal days and times, based on auctions’ final prices. In this article, I’m going to share the results of my analysis.
In order to find the most optimal times, I first had to select the types of coins that I wanted to study. I couldn’t just pick any coin, I had to pick a date and grade with a relatively standardized price. Otherwise, price variations would throw off my data. I couldn’t pick coins that derive a majority of their value from bullion (double eagles, common-date circulated silver, etc) as fluctuations in the bullion price would skew the data. I also needed to find coins that were frequently sold on eBay—tracking the sell prices of MS-65 1955 Doubled Die cents might be interesting, but it’d be a period of years before I could collect enough data points to make any meaningful conclusions.
In the end, I settled on four coins. Each of them have fairly stable selling prices, allowing me to calculate for what I’m interested in—price variations due to auction end-times. The coins I studied were 1893 Columbian Exposition half dollars in typical circulated condition, 1881-S MS-64 Morgan dollars graded by either PCGS or NGC, 1945-S PCGS/NGC MS-65 Mercury dimes, and, for the sake of a little international variety, 1890-1900 British crowns in typical circulated condition.
For my data collection, I quickly eyeballed all the sold listings and avoided coins with exceptional qualities (i.e. “+”, “*”, or CAC slab designations, rainbow toning, etc) as those coins would tend to sell for much more than the typical price. I also constrained myself to true auction listings (as opposed to Buy It Now) with more than 1 bid. My reasoning was that BIN listings and 1-bid auctions represent the valuations of a single person, as opposed to multiple bidders. If one clueless person bids on a coin that’s priced 50% over book value, then recording that value would skew my data away from the true average price. After collecting the data, I discarded all outlier prices* so that unusually fierce bidding wars or massive bargains wouldn’t inflate or deflate my numbers.
The data isn’t quite as conclusive as I would have hoped, but there are still some interesting trends. Thursday was the highest sell price day for 3 of the 4 coins, while Wednesday and Sunday had the lowest sell prices on average. From this rudimentary analysis, it appears that Thursday might be a good day for sellers to set their auction end-dates, while buyers should look out for bargains on Wednesdays and Sundays.
The Time vs Price graph (all times Pacific Standard Time) contains the data points from all four coins, normalized by the mean. The line through the graph is a line of best-fit, with half the data points above it and half below it. The best-fit line provides an idea of the general trend of prices—as you can see, it gradually increases from early morning up into the evening.
It appears that the vast majority of eBay auctions end at around 6:00 pm PST (9 Eastern). There aren’t enough data points from listings later at night or early in the morning to really make any definitive conclusions but it appears the best time for sellers to set their auctions to end is somewhere around 5 to 7pm PST. The best time to look for coins to buy is, as you might expect, earlier in the day (probably because most people are either sleeping or at work).
Obviously there are some limitations to this analysis, the largest of which is a simple paucity of data. The prices I collected represent only a few months worth of data, on a very small number of coins. I personally plan to continue to collect data for these coins and to start collecting data on several more as well. If anybody else is interesting in this sort of thing, I would highly encourage you to begin collecting data on specific coins you’re interested in buying or selling. Who knows what interesting trends might appear?
*For outliers, I discarded all values more than 2.5 standard deviations above or below the mean.