Monday, January 08, 2007

 

Flames Penalty Killing

Click to enlarge this dubious data. TOI is in minutes; PKGA is simply PP goals scored by the opposition while that player is on the ice. The rates per 60 minutes (last two columns) are wrong: I excluded the seven 5-on-3 GA from the PKGA stats, but did not exclude the TOI spent at a 5-on-3 disadvantage, because that data is not readily available (and is near impossible for me to extract even "by hand"). What this table is useful for, then, is a thumbnail comparison of Flames penalty killers to each other. I think most of us knew that Dion Phaneuf is not a very good PK guy at this point in his career, but I'm surprised that Robyn Regehr continues to get rocked out there. (And per the above note, his rates flatter him, if anything: he's always out for 5-on-3's (on for six of the seven 5-on-3 GA), so if you subtracted that time off the TOI, his rates would look even worse).

Stats bleg: has anyone out there done any work, at all, analyzing 5-on-3 time or success? It would sure be interesting to know -- numerically -- just how big an advantage a 2-man advantage is. Virtually everything we know about the PP does not distinguish between 5-on-4, 5-on-3, and 4-on-3, the only exception being # of GF & GA.

As we've discussed a few times on this site, the Goal Differential between the PP team and the PK team is about 6 goals per 60 minutes. Roughly, every 60 minutes of PP time, the average team will score 4 additional goals and allow 2 fewer goals on the PP than they would at ES. But I don't have the faintest clue what effects 5-on-3 situations have on this figure. There was about 25,000 minutes of special teams play last season -- anyone know what % of that was 5-on-3? The Leafs have 12 5-on-3 goals already this season; the Predators have 2. How much of that is explained by success, and how much by opportunity? Just a question.

Comments:

Man, I gotta have some kids so I have more time to post stuff on the internets.

To answer your question Matt, check out the Hockey Numbers blog. He has a link to his own site of hockey stats, which has all sorts of stuff and even that, I think. I remember last year, after the Finals, he showed that the Oilers sucked at 5 on 3 for most of the year despite being decent at 5 on 4.
 


Nope, your spreadsheet looks pretty much like the one I got on the Ducks, also with the deficiencies of being able to differentiate 5-on-3 goals but not much use in doing so since I can't differentiate 5-on-3 ice time (though I get a sense of who plays those minutes by watching).

Only thing I've added is a game-segmenting function, so I can see how the numbers change if I just wanted to see the last month, or the last month on the road, or whatever segment of games tickles my fancy.
 


The Leafs have 12 5-on-3 goals on about 15-18 5-on-3 opportunities so I guess that is a bit of a mix between opportunity and success.

I second MC's auggestion.
 


Two questions: What's the team rate and TOI total?

Don't you think total PK icetime matters? I could care less if Regehr lives or dies, but he has played a lot more TOI than the other guys. The other thing to wonder is what happens when 5on3's expire and Regehr is left out on the ice - it's a 5on4 goal on the scoresheet but it's effectively a 5on3 goal.
 


I don't know why I'm making excuses for that dipshit, but Regehr's also failed to benefit from a single shorthander. They're pretty damn rare but CGY has a few of them and it's shocking he hasn't seen his PKGD benefit from at least a few of them scored by his forwards while he's out there. You'd be hard pressed to convince me that's his fault.
 


It's hard to figure out 5-3 stuff, I had some numbers crunched from last season (I'll try and find them). In general it works out to 16GA/hr or 1GA/2min. Some teams do really well (8 GA/hr) others really poorly (32GA/hr).
 


Ok I hope this isn't a double post...
here's last season's 5-3 results team by team:
http://www.sfu.ca/~cboersma/5-3.htm
 


Desjardins penned this little item that lists the scoring rates for each situation. Although he must have figured out the TOI for each, he didn't provide it.
 


Thanks, JvG, that's awesome.

I'm a little dubious about the Desjardins piece, Jeff, only because he says near the bottom that the PK team is "30% less likely" to score shorthanded 3/5 than 4/5. That percentage surely has to start with a '9' - there's been one 3/5 goal this entire season, and only one last season. (Why it's a good idea to Show Your Work, I guess.)
 

Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?