Like several have commented, it is quite possible to get a burst of noise that causes an error in the years part of the data. Of course, there are workarounds that could be added to the code in the clock, but that adds cost. So possibly a better antenna to hve a stronger signal, or something like that. But the year is seldom my concern, it is almost always the minutes thyat I need to be correct about.
Yours is a WWVB clock (60 KHz) and uses a ferrite loop antenna, usually parallel to the display face. The loop has two nulls in the reception pattern. Orientation of the loop is more important than the absolute position of the clock. Find the vector from your location to Fort Collins, Colorado. Assuming the clock is flat like a wall clock, place it so that a line normal to the display face (for a circular clock like a wall clock it would be the axis of the circle) points toward the Colorado transmitter. You can be off by 20 or 30 degrees without problems. However, if a line parallel to the face of the display points to the transmitter, you are in the null and moving the clock without rotating it won't help much. For instance, if you are due east or west of the transmitter, placing the clock flat against a wall with north-south orientation will work, but placement on an east-west wall will not.
All the clocks in my house are radio controlled (7, I think). They are in all sorts of positions, next to all sorts of electronics (two in my electronics lab), and some on the inside of 24" thick stone walls. They all tell the same time and all change with Summer and Winter changes. Some are connected to the UK time standatrd, some to the German one.
I have gone through all that to point out that yes, THAT clock was designed my a monkey, but not ALL radio clocks are. Mine are all different makes, by the way - some digital, some analog.
Maybe the American atoms are different to the European ones ... :))
Most of the clock like this that I've seen work like "normal" digital clocks when they don't see an adequate signal. Holding the current time makes a lot more sense than randomly setting the time. I've had experiences where I'll move a clock from one room to another and find out that it does not have enough signal to reset. In fact, I have a friend that can't get decent reception on one of these clocks so she gives it to me, it sets properly at my house and I bring it back to her.
Now that that Atomic-Digital clock of yours was really designed by monkeys; glad you pointed that out yourself. There are simply too many unpredictable things that can happen during the day and most of these unpredictables will render the clock unreliable and, at times, grossly inaccurate. You might just as well peek outside the window and guess the time.
It isn't just the "reception." Decades ago, we used to use a Heathkit WWV clock. It used to jump all around in time. Sometimes it would be off by an hour. Sometimes it would be off by several minutes.
The problem wasn't just the clock. They were receiving a subcarrier from WWV/WWVH (NOT WWVB). The subcarrier would transmit either short or long pulses each second in a BCD code to indicate the time. However, selective fading would often shorten a pulse here or there. The detection circuits, even though they had a strong signal, would decode these pulses incorrectly.
Our solution was to have the vendor add a sanity check to the circuitry to reject time changes of more than one minute.
In retrospect, a real WWVB clock would have been a better idea. You don't get selective fading at 60 kHz.
Perhaps NIST would be willing to reconsider the 100 Hz subcarrier format and use something more reliable on WWV and WWVH, such as a very slow fsk signal with error detection codes.
There was a cheap chipset that was released for WWV reception a number of years ago. Some of the white goods guys wanted to put it into appliances. The reception was terrible and the chipset manufacturer calmly stated that you had to find the right positioning for good reception, perhaps by a window. I laughed and told him that nobody was going to run around their kitchen with a double oven and try to find the best reception for the clock.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.