VHL S68 Advanced Metrics

[DMaximus 1,045]

VHL S68 Advanced Metrics

 

One of the frustrating things with STHS is that it doesn’t really provide much information to create advanced statistics that are all the rage in sports today. I took matters into my own hand and created a script that counts all of the events in the full play by play reports of games. That information allows me to then tabulate some advanced metrics. I’ve been running the script for the team I’m on for the last couple of seasons. Now, I’ve finally been able to expand it to run on all games.

I’m going to focus on 3 advanced metrics in this article, Corsi, Fenwick, and PDO.

Corsi counts the number of shot attempts for compared to the number of shot attempts against (including shots that miss the net and shots that are blocked). Note that this is shot attempts, not just shots on goal. This stat is useful as it has been proven to be a better predictor for success than goal differential or shots on goal. Teams with positive Corsi numbers generally have more possession of the puck, which can ultimately lead to more wins.

Fenwick is the same thing as Corsi, but it does not count shots blocked. The thought process being that shot blocking can be a skill that teams should receive credit for. If your team blocks a lot of shots it will lower your Fenwick Against number and increase your total Fenwick %.

For both Corsi and Fenwick if a team is over 55% they are considered elite.

To get the PDO value, you add your team’s shooting percentage and your team’s save percentage together. In general, this is considered a measure of luck. The norm is to be at 100, if your PDO is over 100, you are considered to be “lucky” (Specifically 102 or higher is seen as being lucky more than good), if it’s under (98 or lower), you’re “unlucky”. I use quotes here because this number can be influenced by good goalkeeping, which means a higher save percentage and a higher PDO.

If you wish to read more about these metrics, there’s a ton of information out there, including https://www.si.com/nhl/2014/09/28/fancy-stats-primer-advanced-analytics-corsi-fenwick-pdo-qualcomp

 

So without further ado, here’s the stats so far for the VHL S68. When I ran this, all teams have played 15 games, roughly 1/5 of the season.

 

Corsi

CorsiFor is the shot attempts taken by your team

CorsiAgainst is the shot attempts allowed by your team

Corsi is CorsiFor minus CorsiAgainst. A positive number means you’ve taken more shot attempts than allowed. Negative means you’ve allowed more than taken.

CorsiForPct takes your team’s CorsiFor number and uses it to make a percentage of the total shot attempts in the game. A CorsiFor% of over 55% is elite.

Team Name	Average of CorsiFor	Average of CorsiAgainst	Average of Corsi	Average of CorsiForPct
Calgary Wranglers	72.313	52.750	19.563	57.559
D.C. Dragons	41.600	94.400	-52.800	30.089
HC Davos Dynamo	66.563	58.188	8.375	53.273
Helsinki Titans	71.500	55.571	15.929	55.806
Malmo Nighthawks	52.533	59.600	-7.067	47.002
Moscow Menace	72.800	68.867	3.933	51.075
New York Americans	66.200	68.067	-1.867	49.353
Prague Phantoms	59.333	85.667	-26.333	41.141
Riga Reign	65.071	57.857	7.214	52.973
Seattle Bears	68.625	64.938	3.688	51.074
Toronto Legion	70.000	63.400	6.600	52.415
Vancouver Wolves	76.929	53.143	23.786	58.604
League Average	65.256	65.256	0.000	50.000

 

Fenwick

The exact same as Corsi, just that shots that are blocked are not counted as an attempt.

Team Name	Average of FenwickFor	Average of FenwickAgainst	Average of Fenwick	Average of FenwickForPct
Calgary Wranglers	59.938	45.000	14.938	56.825
D.C. Dragons	35.800	81.733	-45.933	30.015
HC Davos Dynamo	56.438	49.813	6.625	52.956
Helsinki Titans	62.429	47.500	14.929	56.318
Malmo Nighthawks	45.733	51.133	-5.400	47.351
Moscow Menace	62.200	59.000	3.200	51.123
New York Americans	56.267	57.200	-0.933	49.657
Prague Phantoms	50.600	72.933	-22.333	41.141
Riga Reign	55.071	50.214	4.857	52.328
Seattle Bears	58.813	54.688	4.125	51.264
Toronto Legion	58.533	53.933	4.600	51.843
Vancouver Wolves	66.643	44.214	22.429	59.664
League Average	55.656	55.656	0.000	50.000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PDO

Row Labels	Average of PDO
Calgary Wranglers	100.329
D.C. Dragons	99.372
HC Davos Dynamo	102.777
Helsinki Titans	98.711
Malmo Nighthawks	101.191
Moscow Menace	101.064
New York Americans	97.583
Prague Phantoms	97.156
Riga Reign	100.294
Seattle Bears	101.042
Toronto Legion	100.463
Vancouver Wolves	99.651
Grand Total	100.000

 

 

Since we’re only 15 games in, all of this is with a very small sample set and could be heavily skewed by strength of schedule. Because of that, I’m not going to analyze this data.

 

Some other things I’d like to call out. Many times Corsi, Fenwick, PDO are only counted during 5 v. 5 play because power play and penalty kills can skew the numbers. I am not doing that. I am counting every shot attempt regardless of the on-ice situation. Hopefully future enhancements to my script will be able to just count 5 on 5 play. For now this is what I have.

Enjoy!

Edited September 21, 2019 by DMaximus

[Cxsquared 455]

Nice write up. I've actually also been working on a script to scrape Corsi stats together. I'd be really interested to see how you are doing it. In my current iteration I only take into account even strength attempts but I have to manually input the penalty times. My ultimate goal is to create a sim visualizer with the full game play by play. Can't wait to see what the stats look like towards the end of the season. Send me a DM if you want to collaborate.  

[Philliefan 69]

This is awesome! One gripe, though...

 

Can you recalculate them, not counting the games when we play D.C.? I'd like to get a good idea of everyone against actual teams

Edited September 24, 2019 by Philliefan

[Enorama 1,980]

3 hours ago, Philliefan said:

This is awesome! One gripe, though...

 

Can you recalculate them, not counting the games when we play D.C.? I'd like to get a good idea of everyone against actual teams

 

 

 

we have wins somehow

[Motzaburger 1,590]

Dammit @DMaximus I was gonna do Corsi on my next stats article lol

[DMaximus 1,045]

4 minutes ago, Motzaburger said:

Dammit @DMaximus I was gonna do Corsi on my next stats article lol

 

Still do it! All I did was regurgitate some numbers. I'm sure you can add some value to what I did.