Playoffs and Fancy stats thus far

[Mubiki]

The 16 teams that made the playoffs this year were also the 16 top teams in 5 on 5 goals for / goals against ratio.

Last years the Kings in the playoffs ended up leading in that same statistic as they won the Cup, Rangers were leading in the playoffs until they ended up playing the Kings.

In the playoffs,

2012-13 Chicago and Boston co led and they were the finals.

2011-12 Kings and Devils were #1, #2 in that stat and they were the finals

2010-11 Boston led in that stats and they won the cup.

Being a good 5 on 5 team over enough playoff games (seems to be around 12-14 playoff games), and scoring more 5 on 5 than the other team seems to have some correlation going on to who does well in the playoffs and this year who made them. (Will be interesting to see if the trend holds up given more playoff games, so far Tampa is leading in that stat, and Detroit (special teams getting it done) is last, so who knows there, but the only team to advance so far while being at less than 1 ratio is Minnesota at .90 with a playoff leading 33% PP conversion rate)

Having more shots (attempted or on net) than the other team 5 on 5 also seems to correlate somewhat but unless those shots eventually turn into goals either for or against I'll continue to believe shot quality, the shooters and the goalies ability, along with match-ups, style of play, etc all have something to do with how it plays out more so than just using shots for and against as a barometer. (basically a high PDO is not just luck in all occasions)

Minnesota basically fits the bill. They would be positive if not for one REALLY bad game that heavily skewed the numbers.

 

[Sojourn]

They had a better ROW than some of the playoff teams if I'm not mistaken. Their horrid shootout performance is an anomaly in and of itself. But then again, advanced stats "predict" success in the game of hockey, not necessary earning points. Once it goes to a shootout you are playing for a point that's inherent anti advanced stat.

They had a better ROW than one of the playoff teams in the Western Conference. If you include the Eastern Conference, it's a total of two playoff teams.

And even if they had a better ROW than half of the playoff teams, it's still inaccurate to say they missed it exclusively due to the shootout. You aren't going to help any argument for shot differential statistic, by pointing out that they lead the league in the category, but failed to make the playoffs. If shot differential is such a strong predictor of success, it shouldn't have been a matter of earning three more points. And it is three points, because Calgary owns the tie breaker. Winnipeg was the team they had the ROW advantage over.

If you want to start picking out reasons they fell short, I wouldn't start with the shootout. I'd start with their record away from Staples Center. They had one of the best Home records in the West, and in fact the entire NHL. But they also had one of the worst records when they were the visiting team, being 3 games under .500. Only Edmonton and Arizona did worse in the Western Conference when they weren't playing at home. That wasn't because of the shootout, and had they made the playoffs they would have been the only team to make it, despite having a losing record as the visiting team.

 

[theButton]

Fooled by Randomness

 

[Mubiki]

They had a better ROW than one of the playoff teams in the Western Conference. If you include the Eastern Conference, it's a total of two playoff teams.

And even if they had a better ROW than half of the playoff teams, it's still inaccurate to say they missed it exclusively due to the shootout. You aren't going to help any argument for shot differential statistic, by pointing out that they lead the league in the category, but failed to make the playoffs. If shot differential is such a strong predictor of success, it shouldn't have been a matter of earning three more points. And it is three points, because Calgary owns the tie breaker. Winnipeg was the team they had the ROW advantage over.

If you want to start picking out reasons they fell short, I wouldn't start with the shootout. I'd start with their record away from Staples Center. They had one of the best Home records in the West, and in fact the entire NHL. But they also had one of the worst records when they were the visiting team, being 3 games under .500. Only Edmonton and Arizona did worse in the Western Conference when they weren't playing at home. That wasn't because of the shootout, and had they made the playoffs they would have been the only team to make it, despite having a losing record as the visiting team.

It is a strong predictor of success; not necessarily a strong predictor of earning points. But I'm not really trying to defend LA anyway. However, it's being disingenuous to discuss the effects of the stats and earning points and not point out that points can be earned in a skills competition. It seems having a poor year in the shootout can effect a team as much as a terrible PDO. It's not exactly surprising.

 

[Sojourn]

It is a strong predictor of success; not necessarily a strong predictor of earning points. But I'm not really trying to defend LA anyway. However, it's being disingenuous to discuss the effects of the stats and earning points and not point out that points can be earned in a skills competition. It seems having a poor year in the shootout can effect a team as much as a terrible PDO. It's not exactly surprising.

If success is winning hockey games, it should definitely be a predictor of earning points. You earn points, first and foremost, by winning.

I'm not trying to suggest the shootout doesn't have any influence, but it applies to every team equally. The Kings were not unfairly influenced by it.

Personally, I see shot differential more as a result of success than a predictor. I think a better team will continue to win games, but I don't think consistently winning games necessarily means you always have the edge in shot differential.

 

[Peen]

You need the corgi

i found this more funny than i should have.

 

[Mubiki]

If success is winning hockey games, it should definitely be a predictor of earning points. You earn points, first and foremost, by winning.

I'm not trying to suggest the shootout doesn't have any influence, but it applies to every team equally. The Kings were not unfairly influenced by it.

Personally, I see shot differential more as a result of success than a predictor. I think a better team will continue to win games, but I don't think consistently winning games necessarily means you always have the edge in shot differential.

I agree. I think people mistakenly phrase it incorrectly. They seem to think that good advanced stats produce success, when it reality it's teams that are seeing success often have good advanced stats. When phrased correctly it makes more sense why it's used as a predictor.

 

[Blues88]

I'll never understand the suspicion some carry when considering the merits of advanced stats. Are some people overly reliant on stats to gauge worth? Yes. Are some people overly reliant on their eyes to gauge worth? Yes.

Statistics provide context to how a team employs its talent and system(s). That's basically it. Humans assign significance to statistics. Controlling 5v5 shot attempts doesn't EVER ensure success, but can provide context into what success can look like, or what is more likely to contribute to success. It would stand to reason that more successful teams, in general, direct the puck towards the opposition net more often than their opponents. Doesn't mean an opportunistic team can't score a couple goals on 10 shots. After all, puck possession can be meaningless without shot attempts, attempts can be meaningless without shots, and shots can be significantly less impactful without scoring chances (which can be confusing and arbitrary).

Statistics merely identify trends. There are always exceptions.

 

[Sojourn]

I'll never understand the suspicion some carry when considering the merits of advanced stats. Are some people overly reliant on stats to gauge worth? Yes. Are some people overly reliant on their eyes to gauge worth? Yes.

Statistics provide context to how a team employs its talent and system(s). That's basically it. Humans assign significance to statistics. Controlling 5v5 shot attempts doesn't EVER ensure success, but can provide context into what success can look like, or what is more likely to contribute to success. It would stand to reason that more successful teams, in general, direct the puck towards the opposition net more often than their opponents. Doesn't mean an opportunistic team can't score a couple goals on 10 shots. After all, puck possession can be meaningless without shot attempts, attempts can be meaningless without shots, and shots can be significantly less impactful without scoring chances (which can be confusing and arbitrary).

Statistics merely identify trends. There are always exceptions.

For starters, shot differential is not an advanced stat. It's +/- for shots. Labeling it advanced is trying to give it more weight.

 

[Blues88]

For starters, shot differential is not an advanced stat. It's +/- for shots. Labeling it advanced is trying to give it more weight.

Well, before you start, realize that people who refer to stats as "advanced stats" generally include things like Corsi in with that mysterious group of numbers.

We're saying the same things here. Stats just function as numerical context. Nothing real provocative about them....

 

[Sojourn]

Well, before you start, realize that people who refer to stats as "advanced stats" generally include things like Corsi in with that mysterious group of numbers.

We're saying the same things here. Stats just function as numerical context. Nothing real provocative about them....

I understand that, I'm just saying it isn't an accurate label.

I should have been clear that I'm not disagreeing with you, just that it irks me when group those stats together and call them advanced. It just comes across as an attempt to give the statistic more significance by implying it's a more "intelligent" statistic. I don't see it as being that much different from goal differential, or +/-(which, obviously, is just a variation of goal differential).

The stats I'd love to see are the ones that teams are likely, themselves, using. I'm sure they have people actually tracking possession, not just assuming it, and tracking where possession is taking place on the ice. There is probably some variation of a scoring chance differential, where they have people tracking scoring chances for and against. Shot quality. And all that good stuff. Obviously, some teams have been investing more into this than others, and they have no reason to release this kind of information out when they feel it can give them an edge, but it would be pretty interesting stuff.

 

[Doctyl]

Well it looks like the finals will be 2 of the top 10 CF% teams again