Does Defense really win Championships?

There is a truism that Attack wins games, defense wins titles, but is it really true?

Embedded tweets still don’t work on substack, but this is from Orbinho but I have seen this from many places and I am not posting this just to signal him out for anything

This is a very common truism and when I hear things like this it triggers something in me to question if it is actually true or not.

My natural tendency is that this is not actually true. Thinking through things as a thought experiment, a very unbalanced team with a great defense and just an okay attack feels like a team that is limited with their points total. The reason for this is that even with a perfect defense that never allows a goal, if the attack cannot muster a goal you will end up with a fair amount of 0-0 draws along the way and maybe even some of those fluke 20-yard screamers finding a corner from time to time.

Teams in red ended the season as Champions

A team with a modest defense but with an amazing attack will concede goals but if you are scoring 3+ goals per match you will not drop a lot of points along the way.

Getting into a back-and-forth scoring match will be fun for fans but it also is a situation where variance works for you given that you are not up against a limit like you are on defense where you can’t allow negative goals but you can score 5 or 6 in a match.

Overall teams that end up winning the League have both a strong attack and a strong defense, so it does seem that there is a positive feedback look between simply being a good attack and a good defense that comes with having better players.

To test this I pulled the points totals and goals scored and goals allowed for the Premier League going back to 2000/01 season and ran a regression. What I got here was the following equation for points.

Points = 50.2 + 0.65 x Goals Scored - 0.61 x Goals Allowed

If we take the team with the best-ever defense of Chelsea 04/05 (15 goals allowed) and combine that with an average attack (51 goals scored) we would expect that team to win 74 points over a season. That’s pretty good but that is a borderline top-four team rather than one that wins a Championship.

Now if we take the best-ever attack of Manchester City (108 goals scored) and combine that with an average defense for goals allowed (51 goals allowed) we would expect that team to win 89 points over a season. That’s not guaranteed to win the Championship but it does get a lot closer to the title than the team with the historically best defense.

So long story short, what really matters for point accumulation is scoring more than your opponent. Both ways work but on the margin you would rather score a few more goals than allow a few less goals.