2019 CPL Championship Leg 1 -- Defensive Pressure

In my last post I concluded that Cavalry had an abysmal offensive game, specifically in their passing. I wondered if that was in any way because of Forge's defensive pressure. Watching the match, my eyes didn't think that Forge applied a lot of up-field pressure. When Cavalry did have the ball near midfield, Forge closed them down very quickly and effectively.

When I wrote the post I had only been researching and thinking about defensive metrics for a couple of weeks. I hadn't built out any models or analyzed data, and because of that I was hesitant to comment on how much defensive pressure played in the match.

Passes Per Defensive Action

Well, I had a few hours after writing that post and curiosity got the best of me. Now I do have one defensive model to look at. This one is known as PPDA, or Passes Per Defensive Action. Orignally created by Colin Trainor and published on the Statsbomb blog in October of 2013, and updated again in July of 2014, PPDA has become a commonly accepted measure of how much pressure a team applies to their opponents.

Colin's posts give a much better, and more detailed, explaination of the metric than I can/will here. For those that don't want to read his content, the TL;DR is that PPDA measure how frequently a defensive team pressures the ball and/or the ball carrying player in the possessing team's defensive 60% of the field.

Pressures on the ball includes tackles (successful or not), interceptions, fouls committed, and duels won. The number of passes that a defensive team allows (passes allowed) is divided by the sum of the ball pressure actions that the defensive team takes.

PPDA = (Passes Allowed)/(Tackles + Interceptions + Fouls + Duels)

A team with a low PPDA interrupts their opponent's passing more frequently with a defensive action. This is a sign of a team that applies more pressure to the ball carrier and ball up the field.

The larger the number gets, the less pressure a team applies up the field. This could be the sign of a team that likes to sit back and absorb pressure before launching counter attacks.

CPL for the season

We can't use PPDA by itself to tell if Forge's defensive pressure contributed to Cavalry's passing woes. What we need to look at is if Forge applied an above normal amount of pressure to Cavalry. To do that we need to first establish /“normal/“.

Looking at all the games that CPL teams played in the league and CanChamp (excluding Leg 1 of the Championship), I was able to establish the following baseline numbers.

Cavalry FC 10.60
York9 FC 11.67
Forge FC 11.97
FC Edmonton 12.83
Pacific FC 12.98
Valour FC 13.14
HFX Wanderers FC 14.02

Through the 2019 season, Cavalry were the highest pressing team in the league. Forge, on the other hand pressed higher than the mean for the league (12.46), but not by a lot.

On the other end of the spectrum, I fully expected to find HFX at the bottom. Even with Skublak's heroic pursuit of the ball in the opponent's defensive end, the team philosophy seemed to be one where they sat back and absorbed a lot of pressure before lashing out on the counter.

2019 Championship Leg 1

I put the Leg 1 data into the same metric to see what the pressure rates were for this game.

Forge FC 7.83
Cavalry FC 14.33

And we can immediately see two things. Forge way over performed their season average (-4.14) and Cavalry underperformed (+3.73). Interestingly they both moved away from their season average by about the same amount, although I don't think anything can be read into this.

But Forge didn't seem to apply that much more pressure in this game compared to any of the other games in the league this season. So a bit more digging into the numbers was warranted. And the reason for their spectacular PPDA performance was, in part, due to Cavalry's horrible passing performance.

Passes Allowed Tackles Interceptions Fouls Duels Won PPDA
Forge FC 274 3 2 21 9 7.83
Cavalry FC 573 10 2 19 9 14.33

Cavalry only attempted 274 passes during the game. Forge had 35 defensive actions. Because the numerator (passes allowed) in the PPDA equation was so low, it didn't take that big of a denominator (defensive action) to output a low PPDA.

Cavalry on the other hand would have needed to have had double their defensive actions (40) to get the roughly an equivalent PPDA because Forge made over twice the number of passes. Why did Cavalry not get double the defensive actions?

Part of the formula for PPDA is that it only looks at defensive actions taken in the opponents defensive 60% of the field. That means that it excludes the actions taken in the defending team's defensive 40%. The reason it does this is because PPDA measure the high pressure that a team exerts on it's opponent and expects that all teams will apply significant defensive pressure in their defensive 40% of the field. More can be found on this reasoning in Colin Trainor's July 2014 article.

The impact of high pressure

Yes, high pressure can impact a game. A shakey backline can be made to turn over the ball in convenient locations. High pressure can harrass a team to the point where it has to drop back and spend more energy trying to leave their end than they spend on attack.

I don't think that was the case in Leg 1 though. I think Forge played their normal defensive game and that their PPDA numbers were made more impressive by Cavalry's inability to do anything with the ball.

Extra credits

On an interesting side note, the PPDA table that I generated for all the teams involved in the CPL league + CanChamp games had non-CPL teams bookending it. The Whitecaps (9.44) and Fury (10.0) had the best PPDAs with Vaughan (10.69) sitting behind Cavalry in fourth. On the other end, AS Blainville had the least high pressure (15.59) just beaten out by both the Impact (14.89) and Toronto FC (14.10).

How do these PPDA numbers compare to other leagues? According to OptaPro, in 2015-2016 the top 10 teams from the top European leagues had PPDA values ranging from 7.5 to 8.82. So, yah…the CPL doesn't apply nearly the pressure that other leagues do.