Some Advanced stats

[Henkka]

Just found a new page for very interesting advanced stats, which Twitter-known Corey Sznajder has collected with his buddies.

Corey Sznajder is creating public database for hockey microstats. | Patreon

Tableau Public

There's those eye-collected Zone-Entry data, with carry-in, pass-in, dump-ins and total entry fails.

Also there's some shot contribution data and also data from Zone exits.

When checking our players, there was a lot of differencies between them. It was tough to get a big picture, who's overall best.

So I got an idea put all areas manually together, and see what I get. Compare them to league-best value and have a relative value. Then average from all areas. I also added some basic stats to have more complete package.

My package includes (so far with only forwards)

- Total Ice-time
- Points/min
- Goals/min
- Shot contribution data
- Total Possession entries
- Possession Entry success%
- Total possession Exits
- Posession Exit success%

This is a RANK I got:

1. Larkin - 0.689
2. Nyquist - 0.615
3. Athanasiou - 0.614
------------------- "1st line"
4. Mantha - 0.597
5. Vanek - 0.583
6. Zetterberg - 0.578
------------------- "2nd line"
7. Abdelkader - 0.555
8. Nielsen - 0.546
9. Helm - 0.516
----------------"3rd line"
10. Tatar - 0.514
11. Frk - 0.501
12. Sheahan - 0.500
-----------------
13. Glendening . 0.383
Svechnikov - not enough data
Bertuzzi - not enough data

Vanek-data is from Vancouver, and Sheahan-data is from Pittsburgh.

Biggest "found" in here was how bad Tatar was. He was especially worse on those possession stats, where he was a 4th liner at maximum, 2nd worst after Glendening, but his scoring raised him to be "borderline 3rd-4th liner".

When considerng Tatar's caphit and value, I think Holland sold him at perfect time, before it would have turned ugly. 5.25M 4th liner, my god.

Also, trading Sheahan seems to be nothing but a right move. This data was from Pittsburgh, and Sheahan looks weak also there. He was abysmal with us by eye-test, and looked better at Pens, but I'm quite sure, based on these stats, we did lose nothing.

Surprisingly Abdelkader was quite strong on possession data overall, he was 5th best, before adding scoring data, which did drop him to 7th. But that was interesting note. Same with Nielsen.

Vanek and Mantha raised clearly to 2nd line guys when scoring was added, they were mediocre borderline 2nd-3rd liners in all possession data.

Zetterberg was better on possession than in scoring, but on 2nd line level with both.

Larkin is the best in any way, and Athanasiou was second best in possession data. Nyquist raised on the same level than AA, when scoring was added.

Click to read more...

 

[Henkka]

I'll add those TOP values:

- Total Ice-time
League-high, Anze Kopitar 22:10
Detroit-high, Dylan Larkin, 19:51

- Points/60
League-high, Evgeny Malkin 3.97
Detroit-high, Dylan Larkin 2.32 (Tomas Vanek 2.89 in Vancouver)

- Goals/60
League-high, Patrik Laine 1.95
Detroit-high, Anthony Mantha 1.04 (Tomas Vanek 1.20 in Vancouver)

- Shot contribution data
League-high, Nikita Kucherov, 41.40
Detroit-high, Dylan Larkin, 30.29 (47th best)

- Total Possession entries/60
League-high, Connor McDavid, 25.11
Detroit-high, Dylan Larkin, 15.45 (23rd best)

- Possession Entry success%
Detroit-high, Frans Nielsen, 94%
League-high, unknown

- Total Possession Exits/60
League-high, Artemi Panarin, 15.85
Detroit-high, Dylan Larkin, 12.37

- Possession Exit Success%
League-high, Elias Lindholm, 64.4%
Detroit-high, Justin Abdelkader, 52.0% (Martin Frk 54.1%, but low sample size)

 

[Henkka]

Another big find was how bad Luke Glendening was. He is known about PK and his Defensive zone faceoffs, but hasn't been winning them a lot on last year. Something like 51%. Also even though he would win them, he can't get out from the D-zone. I looked those Jay Beagle stats, and he was 3 times better on his data. Superb guy, no wonder Capitals won the Cup.

Imo, we should trade Glendening as the next target. Also article in The Athletic showed, that Red Wings biggest weakness last season was their 4th line, where Glendening played most. These things just come together.

Another thing what I found from that Exit data, was that all Washington Capitals TOP3 centers were at TOP6 in Defensive zone Exits. Bäckström, Kuznetsov and Eller. Lars Eller was the best. And also Beagle had good numbers.

That was their key to Stanley Cup. Always safely out from the own zone, so the opposite team didn't have much O-zone time. Finnish hockey coaches talked about this on TV during the finals, and now I found a data which shows the same.

 

[Oddbob]

This is actually one of the reasons I hate advanced stats quite a bit. Nowhere at anytime last year, was Nyquist contributing more than Zetterberg, in fact, if he wasn't stapled to Z's hip for every shift, he would have been even lower in scoring.

 

[jkutswings]

This is actually one of the reasons I hate advanced stats quite a bit.

Sure, if they're used as the end-all, be-all. But if they're just one tool in the toolbox of evaluation, there's valuable information there. For example, if last year Tatar and Nyquist had both been on the block, at similar production, cap hits, and eyeball tests, this data seems like a convincing tiebreaker that Tatar was the right guy to deal.

Clinging to any one way ignores at least some of the total picture. But advanced stats still have their place.

 

[Pavels Dog]

This is a RANK I got:

1. Larkin - 0.689
2. Nyquist - 0.615
3. Athanasiou - 0.614
------------------- "1st line"
4. Mantha - 0.597
5. Vanek - 0.583
6. Zetterberg - 0.578
------------------- "2nd line"
7. Abdelkader - 0.555
8. Nielsen - 0.546
9. Helm - 0.516
----------------"3rd line"
10. Tatar - 0.514
11. Frk - 0.501
12. Sheahan - 0.500
-----------------
13. Glendening . 0.383
Svechnikov - not enough data
Bertuzzi - not enough data

Seems like a lot of work to get essentially the same result as if you'd just rank the players based on eye test. AA and Vanek probably end up too high (slide them both down below Z) but otherwise it looks about right.

 

[Henkka]

Sure, if they're used as the end-all, be-all. But if they're just one tool in the toolbox of evaluation, there's valuable information there. For example, if last year Tatar and Nyquist had both been on the block, at similar production, cap hits, and eyeball tests, this data seems like a convincing tiebreaker that Tatar was the right guy to deal.

Clinging to any one way ignores at least some of the total picture. But advanced stats still have their place.

For me I would use it just to remove the worst player. Who is really struggling and if the eye-test tell the same. Replace the worst player with one better and it could impact the team a lot.

Too bad we don't have Pavel's numers, it was always best to compare everybody in him. If you are even near, you are doing good.

 

[lomekian]

This does suggest a point me and a few others made last year. Abby's contract sucks, but he learnt enough playing alongside Z & D to improve his overall game, so he can still contribute despite his physicality starting to drop off - and at times he was very good last season. Skill wise he's a 4th liner, but he's smart enough to be a third liner who can do a job in the top 6 if there are injuries. There are a good number of goals last year that were made possible by small, unflashy plays by him in the build-up, and he's deffo one of our best on the boards. If this team was better, he'd be a much more popular 3rd liner.

 

[Frk It]

With the exception of Helm, it’s basically just a reflection/ranking of our best skaters at forward.

Seems like zone entries must matter a ton, because the top 3 guys are far and away the best at gaining the zone.

 

[raymond23]

This is actually one of the reasons I hate advanced stats quite a bit. Nowhere at anytime last year, was Nyquist contributing more than Zetterberg, in fact, if he wasn't stapled to Z's hip for every shift, he would have been even lower in scoring.

Gus last season played his best hockey in years in my opinion.

Didn't have the greatest output but I liked his game a lot more than in the past.

But yeah, would be interested to see his output without Z.