Wasn't sure where else to post this, but I thought it was super interesting. Article in The Athletic today interviewing Steve Valiquette about goaltending trends from the advanced data his company collects. Anyone with a subscription I would encourage you to read it.
- Eye tracking for goalies, the older they get the harder it is to track the puck, so the data shows that younger goalies do better on shots with clear line of sight vs. older goalies who I'm guessing perform about the same when it comes to scrambly plays that rely more on positioning. I wonder if this has affected Bob, lots of times he lets in clear shots or seems to not track the puck well.
- More goals being scored from low slot cross ice passes, so that would be goals that go cross ice below the hash marks. I find we don't have much possession time in the O-zone to get set up and do these kind of plays, usually more of a rush type offense, which is probably to our detriment. I liked it more when we had a cycle type game.
- Unblocked shots from the slot, the shooting percentage has gone down over the years from the teens to 7.5% to 5.3% in the playoffs this year. The thinking here is that maybe because everyone is always told to shoot in this area based on the old advanced stats data that show this is a good area to shoot from goalies are more prepared for the shot. So now it's actually a low percentage chance, but a lot of advanced stats data will still show this as a high percentage shot.
The last two points are especially relevant for advanced stats discussions using the publicly available data from sites like naturalstatrick. I'm wondering if the scoring chances data is out of date given that cross-ice, low slot chances from the outside should still be considered high chance, but currently are considered low chance. Same goes for shots from the slot which gets the highest danger score, but many of those shots are actually below average in scoring % based on the actual data. That also means stats like expected goals or expected goals saved, which is based on older data, could be telling us something that is no longer applicable until the models are updated. This feeds into my skepticism of publicly available advanced stats from sites like naturalstattrick which is useful, but the data can be flawed.
This isn't me being anti advanced stats, but it does show that the stats need to be viewed carefully and the models constantly updated. If the Panthers decide to incorporate more advanced stats into their decision making, hopefully they do it right...
