[Home]


If you review a particular human bidder's recorded bids during a single month, you will see that they do not place bids every day that month, most likely. Also, if you break a day into 24 separate one hour slots (i.e. 0:00-1:00am, 1:00-2:00am, ..., 15:00-16:00pm, ... etc), there will be specific one hour slots that that human bidder never bids in that month, most likely.

An Activity Matrix is a small rectangular picture which quickly displays this behavior. It is composed of 720 colored squares. Each April Activity Matrix has 30 columns and 24 rows of little squares: blue, white, or yellow. Each column represents a day in April and each row represents a one hour block of a 24 hour day. A blue square indicates that the user bid, at least once, on that particular day within that particular 1 hour block. A white or yellow square means they did not bid at that time.

Below, inside the red outline, is Ldbaro's April Activity Matrix. Notice that Ldbaro didn't bid on 5 specific days and didn't bid during 10 specific 1 hour slots. In particular, Ldbaro did not bid on Easter, April 8, nor 4 other April days. Also, Ldbaro never bid from 9pm to 7am PDT time zone during April.



For each user, over a 1 month period, you can define a number called a Rest Coefficient by adding the number of days they didn't bid, that month, together with the number of 1 hour blocks they never bid in, that month. For example, Ldbaro's Rest Coefficient for April equals 15 because he didn't bid on 5 days and he didn't bid during 10 one hour slots.

In December, humans averaged a Rest Coefficient of 40.5 with a Standard Deviation of 14.1 while robots averaged a Rest Coefficient of 2.3 with a Standard Deviation of 1.5. Among December's 4017 users, there is almost a perfect correlation between Rest Coefficient (how often a user takes time off from bidding) and whether they are a human or robot. Excluding a few exceptions, only robots have December Rest Coefficients less than 10 and only humans have December Rest Coefficients of 10 or more. This makes sense. Only a robot bids every hour and every day for an entire month.

In April, there is once again two groups of users where each group exhibits very different Rest Coefficients (and the dichotomy is more pronounced than it was in December). There are 883 April users with a Rest Coefficient of zero! These users never take a break from bidding and, if they were humans, they must sleep, eat, and work at a different time every day! 62 more users have a Rest Coefficient of 1 and 3 users have a Rest Coefficient of 2. In total, there is a group of 949 super bidders with an average Rest Coefficient of 0.07 and a Standard Deviation of 0.3. These 949 never-resting users also share the same unnatural Bid Histogram when users should have unique Bid Histograms, [explained here]. These bidders are robots. In contrast, April's second group of bidders, humans, have an average Rest Coefficient of 37.4 with a Standard Deviation of 14.4 and unique Bid Histograms.

In conclusion, using April Activity Matrices, all users with a Rest Coefficient of 2 or less are robots while all users with a Rest Coefficient greater than 2 are humans. Click [here] to see all April Users' Activity Matrices and Bid Histograms.