logo

Why we love #fancystats

Patrick Johnston
12 years ago
alt
You sure you know the purpose of attack?
(Photo: Ben Nelms / Reuters)
The former coach of the French national rugby team, Pierre Villepreux, once said the purpose of attack is not to score, rather to disrupt the defence. Continue to disrupt the defence, you will eventually score. 
It’s a hard philosophy to embrace, but it’s an absolutely valid statement. You mess the other team up and you should win. It’s the reason why we pay attention to advanced analytics here at Canucks Army.

Why do we use numbers?

The most important thing to understand about statistical analysis in any sport, not just hockey, is what the numbers track. This blog loves to talk about scoring chances. Scoring chances are a reflection of the number of times you break down the defence in a threatening way. Simple logic tells you that if you do that enough times, you will score. Do it more than the other team, and you will win more often than you lose. 
On Wednesday, Cam Charron and I had an excellent exchange on twitter about how the Canucks recent run wasn’t truly reflective of how they’d been playing and similar this was to early in the season.
In both cases their results were mismatched with the numbers. You many remember, the Canucks were said to have sputtered out of the gate, ‘playing’ at best .500 hockey despite outchancing their opponents. When we looked at ‘the process’, we saw that they were playing very well but were immensely unlucky.
The Canucks have been unbelievably lucky in their past eleven games (and even longer, really). Their record in the days before shootouts would be 4-1-6, as opposed to the 8-1-2 that they are currently credited with. Further, the Canucks have been pretty lucky in the last month, with an attack that’s in neutral but being buttressed by lights-out goaltending. 4 ‘real’ wins in 11 games? Not what you need from a championship-calibre team.
The Canucks simply haven’t been breaking down opposing defences enough to win games in regulation.

How teams win games

Villepreux’s notion of how games are won is especially relevant to the the Canucks. Playing a puck-possession game like they do, the Canucks seek to probe the other team’s weaknesses. Each time in the offensive zone is a new chance to not just score, but to destabilize the defence. When we talk about ‘systems’, what do we really mean? Instinctually we think it means disciplining players to play a certain way, and only a certain way. What it really means is giving the players a framework to understand where their teammates are going to be in a particular situation.
The Sedins are probably the best reflection of this principle. They’ve played together for so long that they understand innately where each other is going to be, stringing together passes that seem totally unanticipated to anyone but themselves. Lately, of course, they’ve been sputtering, for reasons that aren’t totally clear. They haven’t been dominating the chances game like they normally do. This tells us that they aren’t controlling the puck in the offensive zone, which is normally their forte. The fact the Canucks are doing as ‘well’ as they are is masking this somewhat.
Systems are created to promote a particular pattern of play and give teammates a certain understanding of each other’s tendencies. They are also designed to exploit an opponent’s weakness and disrupt their ability to adjust. The more chaos you create in the other team’s defensive alignment, the more chances you get to score goals.
It’s hard to focus on the process, but it really is the most predictive way to assess a team’s performance. Advanced stats tell us what a team is doing ‘in-game’ and suggest where things will go next. Sometimes it’s easy – the amount of shots the team has been giving away lately is backed up by their poor form in the scoring chance count – but other times it’s hard, like last fall when their results stunk.
The simple equation? They were losing the close ones in October, now they’re winning them. But please, do ask us ‘why?’
In my next piece coming a bit later today, I’ll look at a specific example of how breaking down a defence leads to a scoring chance (and a goal). 

Check out these posts...