Does Defense Win Championships?

The Right Stats Prove the Importance of Big D

By TSC Analyst/Contributor DANIEL URBAN

LOMBARD  You don't have to be around the sporting world too long before hearing "offense wins games, defense wins championships."   And on its face, it seems plausible.  If you have a great defense, your team will be in every game and have a chance to win every game. 

But does the adage "defense wins championships" hold up to statistical analysis? 

To find out I pulled statistics from the NCAA softball statistics page.  I started with a list of the top 50 defensive teams based on fielding percentage from 2016.  Then I pulled a list of the top 50 winningest teams based on win percentage from 2016.  Of the top 50 fielding teams, 21 were also on the top winning percentage list. 

To measure how closely related fielding percentage is to win percentage, I used a mathematical formula to determine correlation.  Correlation measures the amount to which one number set increases as the other set increases, and decreases as the other set decreases. A perfect correlation is measured as a "1."  The smaller the correlation, the lower the number with a "0" equating to no correlation.

I expected the correlation between fielding percentage and win percentage to be a perfect "1" or close to it.  But I was surprised by what I saw. 

The final correlation between fielding % and win %?  0.411.

Based on this data set, there is only a mediocre correlation between fielding percentage and win percentage.  There is a positive correlation, but it isn't very strong.  If defense wins championships I would expect the correlation to be very strong.  Plus, I would expect that comparing a list of the top 50 fielding and top 50 win percentages would be a very high correlation.

Perhaps if I could expand the data for all teams rather than just the 50 best in fielding percentage and win percentage I might find a higher correlation.  Or perhaps fielding percentage is not a good way of measuring defense.  Perhaps the best way is by runs against or ERA. 

But based on this data, the answer as to whether defense really wins championships is: "eh, sort of."

However, there may be some flaws in my method.  Everyone knows teams with the best pitchers usually win the championships in NCAA softball, but fielding focuses on the overall defense much more than the pitcher, and a great pitcher can win games despite a mediocre defense because great pitchers strike batters out and prevent the ball from ever getting to the defense.

And fielding percentage can be very subjective—it relies on an umpire making a decision on whether a fielder should have made a play or not.  And that is not always easy to determine.

So how do I really prove defense wins championships?  I decided to correlate "total runs against" versus "win percentage." 

When using "total runs against" rather than "win percentage," the correlation jumps up to -0.55.  Don't be confused by the negative—it is negative because, as earned runs go down, win percentage goes up.  This is called a negative correlation.  A -0.55 correlation is much stronger than 0.41.  In fact, it is 25% stronger.  That is a huge jump!  And after reviewing my two attempts at quantifying defense, I am confident that "total runs against" is a much better method of measuring total defense because it incorporates everything a pitcher does rather than just the fielders, and it is a much more objective measure than fielding percentage. 

I am satisfied that I have a numerical value that evaluates how important defense is to winning.  But to be able to say that defense, rather than offense, wins championships, I need to run a similar test for offense to see if the correlation is stronger or weaker than runs against.

To evaluate the benefits of offense I compared "runs for" with "winning percentage" for the top teams. 

The correlation between offensive "runs for" and "winning percentage" is only 0.36, a full 35% less than the correlation between "runs against" and "win percentage" and even a few points lower than the correlation between "fielding percentage" and "win percentage."  In NCAA softball at least, the better your pitching and defense is, the more likely you are to win.

All you college programs out there, in case you didn't know it already, load up on pitching and defense and you are more likely to succeed!  It's better to offer a scholarship to a star pitcher rather than a star hitter.  Now go out there and find more defense!!! Read more.