Tampa wanted to pick up Risto at the 2019 trade deadline. They may have also wanted to pick him up at the most recent deadline. It was to play in the top-4. There are maybe 2 or 3 teams in the NHL that wouldn't have Risto in their top-4. When the rumour spread that Risto wanted out of Buffalo in the fall there were tons of scouts at the games. Buffalo wanted more for Risto than teams wanted to pay. But insiders have said that return Bufalo wanted for Risto was very high. We also know that Dubas engaged in trade talks with Buffalo for Risto. Dubas has far better analytics available to him than anyone on here. And Dubas is orders of magnitude better at understanding those analytics than anyone on here.
The argument against Ristolainen is built entirely on publicly available “advanced stats”. The problem with evaluating him based on those advanced stats is not because of Buffalo, although that does play some role. The problem with evaluating him with those advanced stats is that they are completely terrible for evaluating players. All teams know this which is why they have their own analytics departments in house. There is little doubt that the current obsession with bad advanced stats will go down as the stupidest period of time in hockey history. They rot people’s brains.
Recently I was told, based on advanced stats, that Hughes carried Tanev, and was better defensively as well. Advanced stats made that clear.
Let’s look at the first game of the season. Hughes was clearly better. Especially defensively. Tanev’s relative statistics were terrible and Hughes were great. When apart Tanev was allowing more than 150 CA/60. Hughes 0 CA/60. Tanev must be the worst defensive player in the history of the game. Hughes the best. At least according to the advanced stats:
Player | TOI | CF | CA | CF/60 | CA/60 | SCF | SCA | SCF/60 | SCA/60 |
Hughes | 18.55 | 19 | 10 | 61.5 | 32.3 | 10 | 3 | 32.3 | 9.7 |
Tanev | 17.75 | 17 | 17 | 57.5 | 57.5 | 8 | 6 | 27.0 | 20.3 |
Together | 14.97 | 15 | 10 | 60.1 | 40.1 | 7 | 3 | 28.1 | 12.0 |
QH w/o CT | 3.58 | 4 | 0 | 67.0 | 0.0 | 3 | 0 | 50.3 | 0.0 |
CT w/o QH | 2.78 | 2 | 7 | 43.2 | 151.1 | 1 | 3 | 21.6 | 64.7 |
[TBODY]
[/TBODY]
Except the advanced stats don’t tell you which player was better, defensively or otherwise. They tell you which player was used in defensive situations and which player was used in offensive situations when they were apart. The advanced stats then tell you the player who was used in defensive situations is terrible defensively and the player who was not used in defensive situations was great defensively. It could not possibly get any stupider than that, but that is what advanced stats do, and people evaluate players based on those completely misleading numbers and are then convinced that coaches are stupid and they are brilliant. They further believe that dinosaur coaches lack critical thinking, but that they have lots even though every advanced stats devotee believes the exact same things and has never thought critically about any number they come across that makes zero sense.
What happens when Tanev and Hughes are together if they are in their own zone and recover the puck is that while the Canucks are heading up the ice, Hughes will go with them and Tanev will often head off, with Myers hopping on. When the Canucks then lose possession Hughes will head off. This is a period of time with Hughes on the ice where racking up positive advanced stats numbers are easy and there is almost no risk of any negative advanced stats numbers (Da advanced stats sayz dat Hughes n Myers r betr dan Hughes n Tanev). Tanev, on the other hand, is out without Hughes for almost exclusively the most challenging minutes he plays. Heavily defensive minutes. When his team does not have possession. When high negative advanced stats are inevitable, and any positive advanced stats are unlikely.
Of course, the complaint is always “you can’t use a small sample size” as if a large sample of misleading statistics suddenly becomes valuable. And in fact it is absolutely necessary for anyone who cares about what these stats are telling you to look at them in a small representative sample to see how usage is affecting them. This difference in usage is something that the Canucks coaching staff used throughout the season because they are intelligent enough to know that having Hughes on in offensive situations is valuable to the team and having Tanev on in defensive situations is valuable to the team. The advanced stats don’t tell you that, because the advanced stats almost never tell you anything of value - especially for defensemen. The very different usage when apart make it impossible for the large sample size statistics to be anything except extremely misleading.
But maybe other things are more valuable. For instance, one of the reasons that we have been told that Risto is terrible defensively is that he is poor at possession exits. And it would stand to reason that two players with great possession exit percentages when playing together would be great defensively, but in reality that doesn’t seem to work out. Ristolainen was only at 27% for possession exits (Larsson was at 25% and Savard 24%). By comparison Sandin was at 55%. The highest in the NHL (although the sample size was small) and as much as I love him he is not good defensively at all.
Now if you look at the 5 players who Sandin played with the most three of them had very good possession exit rates (Barrie 43%, Holl 46% and Liljegren – not listed, but it was very high in a small sample size) and two had poor possession exit rates (Marincin 23% and Ceci 29%). But the reality is that Sandin looked good when playing with Ceci and Marincin and looked terrible playing with Barrie, Holl and Liljegren. In fact with the former two he had xGF of 58.4% despite only 15.7% OZ starts while with the latter three he had an xGF% of 41.9% with more than 74% OZ starts. The possession exit numbers are completely useless because they don’t take into consideration the situations in which the player has his puck touches in the defensive zone. Very often, especially in certain systems, D with high possession exit percentages work much better with partners with low possession exit percentages than they do with partners with high possession exit percentages.
Sandin doesn’t win battles in the corners or in front of the net. When he gets the puck in his own zone it is because he is the first back on a dump in or someone passes it to him or something similar. Yes, he is very good at getting the puck out in those situations, but he is not good at getting back possession in the first place. The same goes for the other three he played with that good possession exit numbers. The two he played with that have poor possession exit numbers are D who are winning battles in the corners or in front of the net far more often. But those wins leave you with very little time to get the puck out with possession, in fact if it gets out with possession it is generally because you managed to get it to a support person (ie Sandin) who managed to do that. You get no credit for that. Your support partner does. You only get the negatives.
All people making decisions in hockey know that the publicly available data is complete trash. That is why they spend significant amounts of money to obtain their own far better analytics.
Despite their being more than 7 billion people in the world there is essentially no chance that any of them would be dense enough on their own to believe the following three things: that
Lawrence Pilut has been one of the best D defensively in the NHL, despite his reputation being the opposite and his defensively minded coach only feeling that he was worthy of 13 games this season - on the third pairing; and that
Joshua Ho-Sang is one of the best defensive players in the NHL, and worth more than 7M a year, despite having not played an NHL game in a year and a half and no team even being willing to pick him up off waivers for free, and that
Drew Doughty is terrible. But if you are devotee of the current terrible publicly available advanced stats you should believe all three.
Every coach Buffalo has had played Risto far more than any other D, played Risto against the top competition, played Risto during the hardest minutes. The publicly available analytics do nothing buy mislead about him. He is not Giordano or Josi. I highly doubt that his still available, but he would give Toronto a third legit top-3 D, who is physical, can play big minutes and play in all situations, while having 2 years left on his contract and being years younger than Larsson or Savard. That is the reality.