Question of the week: An option for a penalty shot?
by John Fontana on Jan 20, 2012 9:34 AM EST
in Question of the Week
Email
PrintSometimes, games are punctuated by too many penalties and too much time by either (or both) teams short-handed. I'm not trying to present this as a Tampa Bay Lightning issue, but as a general observation of the game of ice hockey.
Of course, certain teams are quite good with the man-advantage, others are not. Not nearly.
In a league that often employs a change for change sake, something that turns the league on its head... Well, I got a "What if?"
What if NHL teams had the option of choosing between a standard power-play or taking a penalty shot instead? Should this be an option?
This is derived in part for the fondness of the penalty shot (the rare happenstance in today's NHL where an infraction happens under specific circumstances and a player will be awarded a one-on-one shot on goal), and the mixed opinion on the post-game shoot-out...
Really, it's just a quirky idea that a coach could opt to have a player take a penalty shot instead of his team getting 2 minutes of man-advantage time on any given infraction. Some teams have largely ineffective power play units that can turn into defensive liabilities (yes, I do mean the Lightning here).
Really, what about it? Yeah, it's a gimmicky idea, as is the two-point conversion in the NFL... Just what it teams could opt to forgo man-advantage time and take a penalty shot instead?
Dani Toth - Staff Writer / Lightning Hockey Blog
No. But I do think that the reverse should be an option. If a penalty-shot is awarded, I would like to see the team receiving the penalty-shot being given an option to opt for a 2 minute power play instead.
Clare Austin - Staff Writer
It's a fantasy of mine that this team could decline a power play. They've given up 6 shorthanded goals against (T-28th). It's like opponents can foul the Bolts and then score on the penalty kill afterward. But in the end it's better that the team actually fix the power play.
Cassie McClellan - Managing Editor
Can't they just repeat the down? Oh. Wait. Wrong sport.
I like the option of taking a penalty shot instead of having a power play. That's probably something i wouldn't have come up with. Not sure that most teams would take a penalty shot over a power play, but some should seriously consider it now and then - if it were allowed, of course. I'm not naming any names, though they have a lightning bolt on their uniforms somewhere....
Matt Amos - Staff Writer / Don't Trade Vinny
It's not a terrible idea.. Do it. And allow the reverse, instead of the penalty shot, teams can opt for a power play.
And while we're at it - turning hockey into other sports - let's add these rules too:
- Any shot from outside the blueline with a goalie on the ice that goes in is worth two goals.
- Shorthanded goals are also worth two goals.
- Any goal scored by the goalie is worth three.
- Instead of an opposite-end face off, an icing results in the other team getting full possession of the puck at the blueline.
- And just for good measure, instill the Blades of Steel fighting rules: whoever wins the fight gains possession of the puck, and whoever loses the fight has to go to the box.
49 comments
|
0 recs |
Do you like this story?
Comments
On the basis of his suggestions, I nominate Matt for President.
“Blades of Steel” is one of my all-time favorite games. I need to find the opening (with the announcer saying “Blades of Steel!” followed by the sound of sword coming out of its scabbard) to get on my phone as a ringtone.
"Freedom is the freedom to say two plus two equals four. If that is granted all else will follow."
- Smith in Orwell's 1984
If you find that as ringtone...
Immediately let me know. And then I’ll pick you for my running mate.
Well, it looks like things are back to normal around here.
I'm not sure I understand the way this discussion is going, particularly bits like this:
Not sure that most teams would take a penalty shot over a power play
Why wouldn’t they? The best power play in the league right now (Vancouver) is converting at a 23.8% clip. When you subtract shorthanded goals (since you can’t exactly allow a goal by shooting a penalty shot), it drops to 22.7%.
Last year’s penalty shot conversion rate was 34.6%.
I can see situations where a team might prefer a power play to a penalty shot (they have a late lead and want to kill off two minutes with a man advantage), but the vast disparity between PP and PS success rates suggest that instituting this option would be akin to (mostly) eliminating power plays from the game.
I do agree with Dani’s reverse option though.
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 20, 2012 10:15 AM EST reply actions
Think of it in football terms for a sec, Jay…. I started doubting my own validity of the “what if?” when it crossed my mind:
Though a Power Play is seen as an offensive advantage, it also is a defensive maneuver. How so? Wearing down a short-handed opposition for two minutes at a time.Compare that to a sustained offensive drive that wears down an opposing D on the football field….
I mean, it’s only three-hundreths of a game, but if a team gets called for 6 penalties on a single night… That’s 20% of the game shorthanded and 20 percent of the game overworking members of the roster.
You pointed to another obvious defensive reason to opt for one (a PP) over the other (a Penalty shot) — the late lead. That amounts to ball-control, you’ll be better off with the opponent a man down
Typing is an adventure, and reading should be, too!
Sasquatch, Yeti, The Loch Ness Monster, Evan Oberg - legends that may or may not be real
Raw Charge.
by John Fontana on Jan 20, 2012 10:33 AM EST up reply actions
yeah, I get that there may be some reasons to take a power play
but the discussion seems to have gone along the lines of “well, most teams would take the power play, but it might be useful in some special cases to take a penalty shot instead.” And that doesn’t make sense to me. Given that the average team is TWICE as likely to score on a penalty shot than a power play (35% vs 17-18%), the vast majority would probably choose the penalty shot and the power play would be relegated to a special case strategy. It’s like the opposite of your two point conversion example because the gimmick becomes the standard and the old standard is reduced to a rarely used special strategy
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 20, 2012 11:22 AM EST up reply actions
I think teams in the lead would opt for a PP vice a PS, mainly to burn 2 minutes off the clock with the potential to score. Teams behind would probably opt for the PS over the PP for the reasons of higher scoring chance, and no time off the clock.
"Freedom is the freedom to say two plus two equals four. If that is granted all else will follow."
- Smith in Orwell's 1984
And, of course, it occurs to me
that most teams would take a look at the Lightning goalies and opt for the penalty shot.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
same goalies on the kill as on the penalty shot
although, as I mentioned earlier, the penalty shot success rates are so much higher that it’s not much of a decision in most cases
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 20, 2012 2:30 PM EST up reply actions
Dude. You’re going to argue that as a strategy you’d pick trying to score on 4 Bolts plus Roloson over trying to score on just Roloson? I mean if it were Pekka, I’d choose the power play, b/c at least there you have a chance to screen him, but …
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
what? didn't I argue the opposite of that? and say that taking the penalty shot was the obvious decision in almost all cases?
on the subject of Pekka, I’d be inclined to take the penalty shot, since I highly doubt his PS/SO save percentage is better than .830. That said, Pekka is Pekka, so if you have statistics proving me wrong, I’m open to reevaluating
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 20, 2012 5:16 PM EST up reply actions
Pekka’s shootout save percentage is .846 so far this season (11/13). Think he’s faced and stopped one penalty shot this season but can’t find anything on penalty shot save percentages. He was .794 last season (27/34).
Garon is .833 and Roloson is .667 this year.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
(sorry, I’m cooking right now, too)
Garon (5/6) .833, Roloson (6/9) .667; last season Roloson was at (18/24) .750; Garon (8/12) .667. And Garon’s considered a pretty good shootout goalie.
In other words, yes, there is a reason the Preds went to the shootout as many times as they did last year. It was strategic. Keep in mind that the 34 shootout shots Peks faced last season is about the same as one 60 minute game. Extremely small sample size, so one new event can change the numbers significantly.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
okay, so over the past two years
Rinne is at .813 against penalty shots or shootouts (plus any statistics from last year penalty shots that we’re not aware of)
Over the same span, the Preds are 83% on the kill. Which means that you have a better chance of scoring against Pekka on a penalty shot than on a power play, but the difference is very, very slight. So the statistics suggest that your assessment (better screen him than go on a penalty shot) is wrong, but barely
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 20, 2012 7:53 PM EST up reply actions
Possible, perhaps even likely, that the 2& difference between .813 and .830 is well within the margin of error. Not that it matters. I’d still choose PP against Pekka and PS against Garon or Rolosn (esp Roli).
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
it might be
I found an old article from last February saying Rinne was 79% for his career, so his overall numbers are probably around 80%. Still quite probably within error, at the very least, close enough that we don’t have a strong indication of one being better than the other. But I suspect your willingness to choose PP against Pekka is more personal preference than it being an objectively better choice.
The article also said that Rinne’s .790 save percentage was 4th all-time in the NHL. Which means, since the average penalty kill is more effective than 79%, there are at most three goalies in the last six years against whom you’re better off with a power play than a penalty shot. Which plays into my overall point that allowing a choice would make power plays quite a rare phenomenon.
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 20, 2012 8:54 PM EST up reply actions
All of Pekka’s numbers have gotten significantly better since the end of the 09-10 season. Still don’t know where they’re likely to settle, but he wasn’t even the starter in Nashville until late that season.
My problem with your analysis is that you are weighting all of these success rates the same. To me the penalty shots stats are not as reliable as penalty kill stats which aren’t as reliable as 5×5 stats which still have a fairly large margin of error. To me, they’re potentially suggestive, but they aren’t predictive of outcome. You still have to go with the eye test right now.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
not as reliable because the sample size is low?
sure, that’s all well and good. but if you look at league averages, the sample size gets much, much larger.
League-wide, the average penalty kill success rate is about 82%. The average shootout save percentage is 63.7%.
That difference is incredible. If a team has a tremendous shootout goalie and an average to bad kill, it may be advantageous to take a power play over a penalty shot. But even in this situation, I don’t think the advantage will ever be more than slight. I’m not sure what we’re talking about with standard error, so we’ll just take the 82% kill average.
This year, three goalies with more than 10 chances have a shootout save percentage over .820. Varlamov has .895 (19 shots), Pekka .846 (13 shots), and Lehtonen .824 (17 shots). Last year, there were a handful over .820, but none of those three were among them. Of those who faced 25 or more shots (still a very tiny sample size), only Lundqvist (.848), Hiller (.846), and Fluery (.842) managed. The year before that, not a single goalie with more than 15 attempts eclipsed the mark, while Lundqvist, Hiller, and Fluery combined for .711.
The point is, there may be goalies who are better at defending a penalty shot than an average penalty kill. But if there are, they are extremely, extremely rare. As you note, we don’t have a lot of data on Pekka since he took over full time. He may be one. He may be the only one. But even if he is, I don’t know if taking a power play against him is a better proposition than a penalty shot. Unless he can save 85% or more of the penalty shots he takes, it’s pretty close to an even call, or at least well within error either way.
So basically, on average, you’re 18 percentage points better off going with the penalty shot. With an elite goalie and an average power play, you MIGHT (assuming that .850 can be sustained over a longer sample, which I doubt), be 3 percentage points better off going with a power play. That tells me that picking and choosing based on goalies might help, but it won’t help much. And blindly picking a penalty shot every time is pretty damn safe. It can help a ridiculous amount, and the most it can hurt is quite minimal
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 20, 2012 10:36 PM EST up reply actions
Answer me this: how many shots league-wide in all shootouts this season? and how many shots league-wide on PKs all season? So I have some data to work with.
Because what I’m seeing is you saying that the save percentages over 25 shots is comparable to the save percentages over say 150 shots. It isn’t comparable in terms of predictiveness. That means that your 18 percentage points could be anywhere from 10 to 30 percentage points. Even a regular season game usually has more than 25 shots. You can’t say that Lundqvist has an 85% chance of stopping a penalty shot. You just don’t have that level of certainty.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
league-wide shots in shootouts was about 600 last season
and I found an article from last year saying that the league average over the past five years was 67% save percentage. So extrapolating, I imagine that’s with about 3,000 shots worth of data.
Penalty kill statistics are with roughly 4,500 opportunities league-wide this year, with an average success rate of 82-83%
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 20, 2012 11:41 PM EST up reply actions
4500 shots (1/2 of one season) > 3000 shots (5 seasons). The sample size is incredibly small, even at five years. You have no way of knowing if “you’re 18 percentage points better off going with the penalty shot.” You think you may be significantly better off with the PS, but you really can’t trust that trend in any given situation.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
okay, so here's the sticking point
the sample size of 3,000 shots is that small? Small enough to yield what kind of error? It seems to me like a fairly decent sized sample. Maybe not to provide an error smaller than 2 or 3 percentage points, but enough to provide an error much less than 8 or 10 percentage points
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 10:08 AM EST up reply actions
Yes it is that small. When evaluating a goalie for the purposes of predicting future save rates, I wouldn’t take less than 120 games (x30 shots per game=3600 shots) unless I knew a whole lot more about the goalie. And even then, I’d say the margin for error would be huMANgous big.
You can evaluate past performance based on 120 games with some degree of confidence, but not predict. Goalie performance is just that varied.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
hmmmm
I didn’t realize it was considered that small. So, adding what I tabulated from last year to the existing sample, we’re looking at about .667 save percentage over 3300 shots. Three questions:
1. What are we looking at the range of error. Is the league average somewhere bewteen .600 and .730, or is it between .640 and .690, or. . . ?
2. Can we say anything about upper and lower limits? Do the very best and very worst performances is that sample serve as a decent representation of “elite goalies won’t do better than this” and “bad goalies won’t do worse than this”? I would expect so, because variance is usually increased in small samples, so the edges should get closer together as the sample increases.
3. On a related note, what’s high variance from the norm in a shootout? Do we have any idea? With regular goalies, you don’t expect much more than a shift of 2-3 percentage points from the average. But regular shots are a lot different (read: waaaaaaay more varied) than shootout shots. Does that mean we should expect a wider distribution? Narrower? Does a more uniform sample mean we need fewer shots to have confidence about the results?
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 10:44 AM EST up reply actions
On those questions I don't really have an answer
I don’t know enough statistics to be able to figure out what the actual margin of error would be. I am sure it’s bigger than 3%. I’d expect something closer to 10%.
In fact, I’m not even going to go out on that limb. I just don’t know what the variance would be and without that knowledge, and knowing what I do know about the variability of goalie stats, I’m just going to have to leave it at “I don’t know, but I don’t think you do either.”
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
there are margin of error calculators on the internet
but you have to know the size of the actual population in order to figure out what the margin of error on your sample is. I went ahead and plugged in a sample size of 3,000 (five years of shootout stats) trying to predict a sample of 30,000 (roughly fifty years of shootout stats). The margin of error is 1.7 percentage points. I don’t remember too much of statistics, but 10% just sounded way off to me. Maybe we can get Steve to swoop in to the rescue here. But 1.7 sounds reasonable given the comparison you were making earlier. You said that 3,000 is a small sample size for a goalie. But would you really expect a goalie to be TEN percentage points worse than what he is over 3,000 shots? That is, a goalie that is at .915 for nearly a full season is really just a career .815 goalie? Doesn’t that just ring false? Whereas, with an error of 1.7 percentage points, a .915 goalie through one season should have a career mark between .898 and .932. Doesn’t that just sound about right to your goalie stat ears? It sounds right to me, but you’re the goalie nerd here.
On point #3, I agree that I certainly don’t know. Wondering if you had any insight. On point #2, on the other hand, just from what I know about sampling, I would be very, very surprised if the top and bottom career shootout goalies over a five year sample didn’t provided a rough upper and lower bound (within a couple percentage points of error) for what you could expect from the best/worst shootout goalies over a much longer span. Just because standard deviation is much, much higher over a small sample than over a high sample. Thus, I feel pretty comfortable saying elite shootout goalies won’t sustain a save percentage over .800 for their careers.
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 11:13 AM EST up reply actions
That is, a goalie that is at .915 for nearly a full season is really just a career .815 goalie? Doesn’t that just ring false?
No, I’m not saying that a guy’s career stats could be 10% lower or higher. I’m saying that in any given game/shootout, you could easily expect a goalie to be 10 percentage points off his career average.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
for that matter for any given season you could expect an average NHL goalie to be off his career average by far more than 1.7%. Roloson’s career SV% is .909. His high season save % is .933 (3%), his low full season is .884 (2.5%). That’s a 3% margin of error on his career stats, with a sample size of some 16000 shots.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
I'm not that concerned with 1.7 vs 2.5 or 3
it’s the difference between a PS shot being better by 13.3 to 16.7 percentage points or being better by 12 to 18 percentage points. Either way, it’s way, way, way better. Now if we’d started getting to 10 percentage points, that would’ve been an issue.
It’s quite possible that goalie stats are abnormal in some way and have a higher than usual error, but 3 makes a lot more sense than 10
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 11:40 AM EST up reply actions
16, 417 shots. Not 3000. Not 4500. 16,417 shots over a career gives you the confidence to think that Roloson will, over the course of some 1400 shots be within 3% of .909.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
you're doing it backwards
or we’re miscommunicating again. I’m taking the sample we have and trying to project over much larger samples to determine what to expect all-time from average, elite, and bad goalies. You’re taking larger samples and trying to project down to a given season.
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 11:48 AM EST up reply actions
according to sample size calculators
the margin of error when using 1400 shots to predict 16,417 is 2.51. So given one season’s save percentage for Roloson, you’re 95% confident that his career save percentage will be within .025. If not dead on, that seems very, very close
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 11:57 AM EST up reply actions
If I’m trying to decide whether to take a PS or a PP, why do I want to calculate a guy’s career numbers?
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
what you want
is as good a prediction as you can find as to whether he’ll succeed. I’m of the opinion that his track record in 100 or 200 or 300 (or 10,000, if you have them, which you don’t) attempts is a better indicator than his track record in the last 15, or guesses on whether he’s in the zone today, or a number of other things.
This is the observation on which the entire discussion rests. Playing the percentages is better than going with your gut based on who’s hot (provided that there’s a big enough gap in percentages, of course). And the bigger the sample, the more reliable the percentages.
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 3:28 PM EST up reply actions
But if you’re extrapolating from 300 to 3000 to whatever, you’re not getting more accurate data collection. You’re creating estimation errors. There isn’t any more data to get. Your sample size isn’t going to get any bigger than it is now. And you’re still stuck with the question of whether Pekka’s tracking the puck today or not.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
you're doing what you can though
if you can extrapolate with an error of only 3 percentage points, then if your number is more than 3 percentage points away from your kill number, what you have is plenty usable enough
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 8:18 PM EST up reply actions
oh
that’s what you were saying? then I think this entire argument might’ve been predicated on a HUGE misunderstanding.
When you have a sample size of 30 (or 3) in a population of 30,000, you’re obviously going to have massive error. Much higher than 10 percentage points, in fact.
I was making an argument that the smart thing to do is to play the percentages. Suppose your power play and penalty shot ability are equal (unlikely, but it simplifies things). Then you should choose a power play vs penalty shot by comparing the goalie’s SO/PS ability vs the team’s penalty kill ability. And because even elite SO/PS goalies are worse than mediocre penalty kills, you would choose the penalty shot the vast majority of the time. When the league averages are this far apart, playing the percentages are just a vastly better proposition than trying to guess whether this is going to be the same where the goalie is going to be better than his usual at stopping penalty shots.
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 11:37 AM EST up reply actions
As a coach making a decision aren’t you playing the variances in the stats? Aren’t you hoping for the time that the other team doesn’t live up to their stats? If a power play fails 80+% of the time, aren’t you always wanting to be in the less than 20% and not the 80%?
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
of course you want it to be
but you’re not banking on it. I’m not sure exactly what the statistics are, but say for every two minutes of 5 on 5, a given team has a 8% chance of scoring. In two minutes of 5 on 4, you have an 18% chance of scoring. In both cases, you’re hope to beat the odds. But wouldn’t you take the 5 on 4, if given a choice? It’s not even a very hard choice.
The difference between a PS and a PP is similar
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 11:45 AM EST up reply actions
also, I feel like there's some miscommunication going on
I never felt like performance over 25 shots generalized much at all. I cited statistics for individuals over 25 shots just because that’s the best we have for a given goalie in a given season. I then pointed out that none of the goalies with success in a (small) one season sample size sustained their success the next season, suggesting regression to the mean as the sample gets larger.
And I never said that Lundqvist has an 85% chance to stop a penalty shot. I said I highly doubt that anyone has that chance to stop a penalty shot. I said that Lundqvist maintained those numbers for a season but had been significantly worse the year before (roughly .700) and the year after (roughly .700), which again suggested regression to the mean and shed considerable doubt on the idea that a save percentage in the .850 range is sustainable
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 20, 2012 11:47 PM EST up reply actions
You are arguing that based on SO save percentage alone you can make a decision
I’m saying you can’t make a decision on SO save percentage alone. As a coach making this decision I know several things:
—my team’s SO/PS shooting %
—the opposing goalie’s SO SV%
—my team’s PP%
—the opposing team’s PK%
—who was on the ice when the penalty occurred
—that I get one shot on a PS
—that I get more than one chance on a PP
—that I can move or screen the goalie on a PP
—what kind of scouting I’ve got on this goalie
—how my team has looked in practice and in the game
—how the opposing team has looked in the game
—how the goalie has looked in the game
—whether there are any recent trends that would skew the numbers
I also know that I’m not facing the league. I’m facing this particular guy and his numbers and tendencies. I have to make a decision based on all of that information. The least reliable bit of information I have are the shootout percentages, both shooting and save. There’s a huge amount of give in those numbers. There’s a fair bit of give in the PP/PK numbers but that’s still far more solid than the SO numbers.
I have to decide and, to me, there are more situations in which I’m going to decide against the PS than you are allowing for. I’m not saying PS wouldn’t increase, and by quite a bit. I am saying that it’s not a simple as PS have X success rate and PP have Y success rate.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
I'm sorry, I wasn't trying to argue that the decision is based solely on SO save percentage
I was just trying to make the point that the penalty shot would be vastly preferred to the power play in most situations.
Obviously, your own team’s SO/PS shooting % and power play percentage is a huge factor, in addition to the opposing team’s save percentage and PK%. And recent trends as well. I remember last year when we were something like 23% on the power play and had lost several shootouts in a row. I could see us choosing power play then solely for the mental edge.
What I am arguing is irrelevant are things like your ability to move/screen a goalie on a power play or get more than one chance on a PP vs a penalty shot. Or perhaps not irrelevant, just already taken into account in the data. You’ve always been able to move or screen goalies and get multiple shots, and whether or not you’ve been successful doing so is factored into your level of power play success. And the ability of the other team to keep you from doing so is factored into their level of penalty kill success. So those sorts of factors have already been counted twice, there’s no need at all to triple count them.
Now I’m not sure what you’re saying when you say
The least reliable bit of information I have are the shootout percentages, both shooting and save.
Are you talking about the shootout percentages for individual people? If so, then you’re obviously correct, and we’ve agreed upon that for a good long while now. But if you’re talking about league-wide shootout/save percentages, I think they provided a good indicator that choosing the penalty shot is going to be preferable in the vast majority of situations. Can you overcome an average gap of 15 percentage points? Yeah. But you don’t do it often—it requires a combination of elite SO goalie, mediocre (or bad) kill by your opponent, excellent shootout ability* by your own team, and a good power play by your own team.
*note I say “ability” and not percentage. This is because of the earlier discussion of the unusably small sample size in individual team’s shooting percentages. But, that said, I expect the coach has a decent sense of the ability from watching practice
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 10:05 AM EST up reply actions
sorry, when I said excellent shootout ability there are the end, I meant poor shootout ability
got crossed up what I was talking about there for a second.
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 10:24 AM EST up reply actions
Yes I mean SH% for individuals. As the coach making the decision at this specific moment, I have to take into account the SO abilities of the 5 guys who were on the ice when the penalty occurred, because I have to choose one of them to take the shot. Unless they’re Parise/J. Jokinen/Datsyuk or someone like that, Most of my players will have made very few attempts and will thus have an incredibly misleading sh% in that situation.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
right
then I completely agree with that point
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 21, 2012 10:45 AM EST up reply actions
I’m saying that the specific goalie would make a difference to me if I were making a decision. The goalie isn’t just “the same goalie” as on the PP. It’s not equal. The problem with that league-wide PS success rate stat is that the sample size is so incredibly small that it’s predictive value is not that great. It’s even less predictive than a team’s PP or PK percentage.
I’d take a PP against certain goalies and a penalty shot against others. And if I were faced with these two goalies I’d take the PS every time.
R.I.P. Belak, Rypien, Boogaard, Lokomotiv.
"You don’t motivate people; you activate something in them that already exists." -Guy Boucher
yeah, that comment was just a little snarky
just saying that we have bad goalies on the kill too, not just on the penalty shot. But I agree with your overall assessment that it depends on the goalie and his particular strengths and weaknesses. But it would have to be an extraordinary goalie (like, better than Pekka) to make the power play worthwhile.
Although if Garon’s extra small sample size statistics extrapolated (which I’m sure they don’t), then it’d be better to choose a power play than a penalty shot against the Lightning when he’s in net. Roli not so much
If I cared more about my UNC side, I'd call myself "Tar Volon," and that'd be awesome.
Bolts, Canes, Preds (now in different conferences!). Canes mini-STH. Southern hockey solidarity
Rocky Top Talk
by Incipient_Senescence on Jan 20, 2012 7:55 PM EST up reply actions
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Recent FanPosts
Pages
Raw Charge on the Web
In Association With
Latest Puck Daddy - NHL from Yahoo! Sports
- Why the Devils and the Kings are evidence the Capitals shouldn’t trade Alex Ovechkin
- Eulogy: Remembering the 2011-12 New York Rangers
- Stanley Cup Final: 5 keys for New Jersey Devils and Los Angeles Kings
- Friday’s Three Stars: Devils head to Final; Brodeur, Kovalchuk the difference in Game 6
- Devils eliminate Rangers in Game 6, ending fear of the past, setting course for the future
SBNation.com Recent Stories
