The Faulk-JBo pairing has the 36% GF%, the Faulk-Petro pairing has an even 50% GF%
To me, the underlying numbers show that Faulk plays fairly solid for most of his TOI, it's just his mistakes are big ones that have generally ended up in the back of the net, and that ends up sticking on our minds more than a mistake that doesn't result in a goal against. If even a few of those big mistakes don't end up in our net, he'd probably have a much better perception on here in my opinion. He seems like he's caught in between not being an ideal fit and having a bit of bad luck in terms of bounces both for and against.
Good catch, thank you!
I think you are dead on with the bolded.
Excellent post! Thanks for all the work you put in. An old geezer like me who is new to advanced stats really appreciates
the comprehensive breakdown.
This really helps!
Haha happy to help Simon. "Advanced" stats may seem unapproachable at first, but it just takes a little time to understand exactly that the stat actually measures.
For example, a player's CF% (corsi for percentage) is pretty simple: it compares the amount of shots on the opponent's net vs the amount of shots on his own net while a said player is on the ice. If a player's CF% is high (above 50%), it means that when he is on the ice more shots are hitting the opponents goalie/net than his own. A player's GF% (goals for percetage) just compares how many goals his team scores vs how goals are scored against his team while he is on the ice. Pretty straight forward so far.
Now let's compare a shot from the point vs a tap in goal from back door pass - obviously the back door tap in shot is more dangerous/likely to score but corsi considers them both just a shot on net. If the shot from the point is an apple and back door pass/tap in is an orange, then corsi often compares apples to oranges.
Stats like HDCF% (high danger corsi for percentage) and xGF% (expected goal for percentage) essentially try to improve on corsi by taking into account how dangerous a shot is and allow us to compare apples to apples. HDCF% works the same way as corsi (comparing number of shots on opponents net vs own net when a player/pairing is on the ice), but only counts shots taken from within a certain area around the net. This means that instead of comparing the back door tap in with the shot from the point, the tap in is compared to other shots that come from within a couple feet of the net. xGF% also takes into account how far away from the net the a shot comes from, but is a bit more complicated in that it tries to take into account additional factors (ex. is the shot off the rush, is the shot off a rebound, ect).
None of these stats are perfect, but once we understand what the letters and numbers actually mean, then we can compare them different stats and use them as a tool to better evaluate what we are seeing. For instance, we would generally expect a pairing that has a high HDCF% (more shots on opponents goal vs shots on their goal from within a couple feet) to also have a high GF% (more goals scored than given up). Looking back at the Faulk - Pietrangelo pairing however, they have a high HDCF% at 57.29 but their GF% is fairly low at exactly 50%...this is why I think they may have had some back luck! The eyetest still matters, but we can try to use these stats to help find blind spots.
If you have an Athletic subscription, this article is a very good read and does a great job breaking down what a lot of "advanced" stats actually measure -
An advanced stat primer: Understanding basic hockey metrics
Great write up and presentation of stats.
Bouw-Parayko have been drastically outproducing their expected GF% going back well into last season. At this point, we probably have a 100+ game sample size of their actual GF% being double digits better than their expected GF%. At a certain point, I have to believe that their play style isn't properly analyzed by expected GF%. The eye test tells me that they are a great shutdown pairing and we have 100+ games of our goaltenders having a noticeably higher SV% with them on the ice. They are both such big bodies and clog passing lanes so well. My theory is that they do a wonderful job of preventing cross ice passes and clearing rebounds, so a big percentage of the high danger chances against are "less" dangerous than they usually are because they aren't being taken immediately after a pass or broken play. A goalie square to a shot from 10 feet out has a much higher chance of stopping it than a goalie who is sliding over for the same shot. But the current state of analytics counts those as the same value chance. Bouw-Parayko limiting the amount of times our goalies have to move pre-shot could very well explain how they consistently outperform their expected numbers by such a large margin. The sample size is just too big for me to believe it is luck when my eyes tell me they do a fantastic job defensively.
Yeah, Bouw - Parayko's xGF% is always going to be skewed due to their usage. It's just not really possible to have a high xGF% when you start so many shifts in your own zone against other team's top lines. It goes back to what I was saying to Simon, none of these stats are perfect.
The bolded definitely matches my personal eye test. How many times did we see Parayko out muscle a guy and clear a rebound trickling towards the goal line last playoffs alone? Bouw and Parayko do a great job clearing the crease, tying up sticks, knocking away rebounds and blocking shots. They generally seem to play a highly structured game which limits the "surprises" for the goalie and makes their life a lot easier.
Still, the difference between Parayko and Bouw's GF +/- vs xGF +/- is really remarkable. Faulk's is as well, tho in the opposite direction. Here are their individual 5v5 stats from Corsica:
| GF | GA | GF +/- | xGF | xGA | xGF +/- |
Parayko | 33 | 29 | 4 | 30.39 | 35.31 | -4.92 |
Bouw | 33 | 32 | 1 | 30.25 | 37.22 | -6.97 |
AP | 51 | 44 | 7 | 39.15 | 35.52 | 3.63 |
Faulk | 35 | 42 | -1 | 34.82 | 32.38 | 2.44 |
[TBODY]
[/TBODY]
On a semi related note, this discussion lead to me finding out that Corsica is finally back up! FWIW Parayko and Bouw face toughest 5v5TOI%QoC on the Blues (Pietrangelo close 3rd).