Jimmy the Groundhog predicted what we all feared Friday morning. The Sun Prairie groundhog saw his shadow at sunrise, indicating there will be six more weeks of winter.

In Pennsylvania, the more famous Punxsutawney Phil also saw his shadow. By these estimates, it looks like we might endure more cold weather.

But the tradition of Groundhog Day poses a worthwhile question: Are these underground animals worthy weather prognosticators for Janesville?

The answer depends on how you look at it.

The city of Sun Prairie boasts an 80 percent accuracy on behalf of Jimmy. Actually, Jimmy is always correct, according to the city of Sun Prairie, but the mayor “misinterprets” Jimmy’s prognostication 20 percent of the time, according to Sun Prairie.

Sun Prairie doesn’t keep a record of Jimmy’s previous predictions, so The Gazette couldn’t compare past weather data with his prognostications.

Instead, we looked at Punxsutawney Phil’s predictions since 1948 and compared them with The Gazette weather database to the find out how accurate the Pennsylvania groundhog is for Janesville.

According to our criteria, we found that Phil was right exactly 50 percent of the time. In the past 70 years, he accurately predicted Janesville remaining winter 35 times.

But some factors complicate our findings. For example, the definition of an “early spring” is muddy, particularly for Wisconsin.

Does an early spring mean higher-than-average temperatures in the six weeks after Feb. 2? Does it mean fewer inches of snowfall?

According to The Gazette’s weather database, the average Janesville snowfall between Feb. 2 and March 16 from 1948 through 2017 is exactly 10 inches. The average temperature is 26.9.

We decided that if Phil did not see his shadow, there was only one way for him to correctly predict an early spring for Janesville. In the six weeks after Feb. 2, the amount of snowfall in the region had to be less than 10 inches and the average temperature had to be higher than 26.9.

In 1983, for example, Janesville received 9.6 inches of snow in those six weeks, and the average temperature was 35.5. Phil did not see his shadow that year, predicting an early spring. So, according to our criteria, he was correct.

But in those same six weeks in 1988, 7.8 inches of snow fell to the ground, and Phil did not see his shadow. He was wrong, though, according to our criteria, because the average temperature was 23, almost four degrees lower than the 70-year average.

One unusual year in our examination is 1974. The six-week average temperature was almost 30, and Phil predicted six more weeks of winter. He ended up being correct, though, because over 14 inches of snow—4 inches more than average—fell in the following six weeks.

The results change slightly with different formulas. With the same overall averages, if snowfall is taken out of the equation and only temperatures are calculated, Phil would have been correct 34 out of 70 years; but if the equation only considers average snowfall as an indicator of spring, Phil would have correctly predicted Janesville weather 41 times.

So all of this begs the question: Is Phil right this year, and will we have six more weeks of winter to suffer through?

Just flip a coin to find out.