Scoring in playoff elimination games

[Hockey Outsider]

Does anyone know if there's a site that shows players' stats in playoff games where their team can eliminate their opponent and/or be eliminated?

I think I've figured out how to do this in Excel but I wanted to validate that the end results are correct. Also it's challenging to keep track of the situation in a series when a player misses games (maybe someone else has already done the work).

Click to read more...

 

[Hockey Outsider]

I'll post what I have for Ovechkin in the hopes that someone can tell me if these numbers look right (or wrong). Note that I've excluded the 3 play-in games from 2020:

Ovechkin - production by game number

Game #	Games	Goals	Assists	Points	PPG
1	22	9	9	18	0.82
2	22	17	12	29	1.32
3	22	12	14	26	1.18
4	22	11	10	21	0.95
5	21	10	7	17	0.81
6	17	8	8	16	0.94
7	12	4	4	8	0.67

[TBODY] [/TBODY]

Ovechkin - production by game type

Situation	Games	Goals	Assists	Points	PPG
Can eliminate opponent	28	11	8	19	0.68
Can be eliminated	27	13	11	24	0.89
Mutual elimination	12	4	4	8	0.67

[TBODY] [/TBODY]

Let's not turn this into a discussion about what these numbers mean (ie "how good is Ovechkin in the playoffs"). I'm just trying to validate the data, and I picked him because he hasn't missed any playoff games in his career (it's a manual process to identify and adjust for these).

 

[Mickey Marner]

I started a thread about Patrick Kane in elimination games.

Patrick Kane in elimination games.

@pnep provided tables with an enormous amount of data, but the tables don't seem to have survived.

I'd try asking him about it.

 

[pnep]

im don't saved it
sorry

 

[morehockeystats]

I can produce that data. I haven't the bandwidth to do it for free though... If you need only goals and assists, I think I can generate a CSV for $50.

 

[filinski77]

I’ve tried looking once, but there doesn’t seem to be an easy way to pull the data (unless you pay for it, or are much more tech savvy and have the free time). I did make a post last summer comparing Kane/Ovi/Crosby (not sure I remember the exact context of why), but it only looked at games where your own team was facing elimination.

In games where your team is facing elimination (includes game-7):

	GP	G	P	G/GP	P/GP
Kane	20	7	25	0.35	1.25
Ovechkin	27	13	24	0.48	0.89
Crosby	22	4	16	0.18	0.73

[TBODY] [/TBODY]

 

[Hockey Outsider]

I’ve tried looking once, but there doesn’t seem to be an easy way to pull the data (unless you pay for it, or are much more tech savvy and have the free time). I did make a post last summer comparing Kane/Ovi/Crosby (not sure I remember the exact context of why), but it only looked at games where your own team was facing elimination.

In games where your team is facing elimination (includes game-7):

	GP	G	P	G/GP	P/GP
Kane	20	7	25	0.35	1.25
Ovechkin	27	13	24	0.48	0.89
Crosby	22	4	16	0.18	0.73

[TBODY] [/TBODY]

Thanks for sharing. Our numbers are identical for Ovechkin. I've set it up so that it's easy to do the analysis when a player hasn't missed any of his team's games. If he has (which is the case for Crosby, Kane - and most stars), it's a pain to set up.

 

[pnep]

I started a thread about Patrick Kane in elimination games.

Patrick Kane in elimination games.

@pnep provided tables with an enormous amount of data, but the tables don't seem to have survived.

I'd try asking him about it.

Patrick Kane in elimination games.

Im impressed just by the fact that he has seen 33 elimination games already. Is that true? Damn. im not sure. this is from a USA today recap 3 days ago. No group of players knows more about what it takes to win in the postseason than these guys. Since 2009, the Blackhawks have won 75...

hfboards.mandatory.com

 

[pnep]

* - min. 10 Elim. Games (PO Seasons 1988-2015)​

 

[Hockey Outsider]

* - min. 10 Elim. Games (PO Seasons 1988-2015)​

@pnep - first, with everything going on in Russia, hopefully you're okay.

This is one of the very few stats I've ever seen that makes Forsberg not look great. He ranks second (behind only Lemieux) in points per game in non-elimination games (1.29), but he apparently only scored 0.71 PPG in elimination games (on par with Derek Roy, Alexandre Burrows, Andrew Shaw, Mark Lamb, and Derek Plante). No star forward is anywhere close to that 0.58 PPG dropoff (the next closet HOF forward, from what I can tell, is Cam Neely with a 0.40 PPG dropoff). Definitely surprising.

Patrick Kane looks really good here. He scored 0.82 PPG in non-elimination games (tied for 36th out of 322 players - good but not great) but he jumps up to 1.42 PPG in elimination games (behind only Lemieux and Messier). Other HOF forwards with big jumps include Denis Savard (surprising), Messier, Kurri, Toews, and Zetterberg.

Out of curiosity, what definition is being used for "elimination games"? The table shows Ovechkin played in 24 elimination games but based on my count, he's either played in 28 (when the Capitals can eliminate their opponent) or 27 (when the Capitals face elimination). (EDIT - it looks like your data is from 2015, that explains it - I also have Ovechkin at 24 elimination games at that point in time).

 

[authentic]

@pnep - first, with everything going on in Russia, hopefully you're okay.

This is one of the very few stats I've ever seen that makes Forsberg not look great. He ranks second (behind only Lemieux) in points per game in non-elimination games (1.29), but he apparently only scored 0.71 PPG in elimination games (on par with Derek Roy, Alexandre Burrows, Andrew Shaw, Mark Lamb, and Derek Plante). No star forward is anywhere close to that 0.58 PPG dropoff (the next closet HOF forward, from what I can tell, is Cam Neely with a 0.40 PPG dropoff). Definitely surprising.

Patrick Kane looks really good here. He scored 0.82 PPG in non-elimination games (tied for 36th out of 322 players - good but not great) but he jumps up to 1.42 PPG in elimination games (behind only Lemieux and Messier). Other HOF forwards with big jumps include Denis Savard (surprising), Messier, Kurri, Toews, and Zetterberg.

Out of curiosity, what definition is being used for "elimination games"? The table shows Ovechkin played in 24 elimination games but based on my count, he's either played in 28 (when the Capitals can eliminate their opponent) or 27 (when the Capitals face elimination). (EDIT - it looks like your data is from 2015, that explains it - I also have Ovechkin at 24 elimination games at that point in time).

I've seen those stats posted about Forsberg before. Definitely surprising. The same thing is true for MacKinnon today, roughly the same difference he drops in elimination games. What are Gretzky's stats for these I wonder?

 

[authentic]

I’ve tried looking once, but there doesn’t seem to be an easy way to pull the data (unless you pay for it, or are much more tech savvy and have the free time). I did make a post last summer comparing Kane/Ovi/Crosby (not sure I remember the exact context of why), but it only looked at games where your own team was facing elimination.

In games where your team is facing elimination (includes game-7):

	GP	G	P	G/GP	P/GP
Kane	20	7	25	0.35	1.25
Ovechkin	27	13	24	0.48	0.89
Crosby	22	4	16	0.18	0.73

[TBODY] [/TBODY]

I wonder what Malkin's are?

 

[Michael Farkas]

I wonder what Malkin's are?

Malkin:
Can eliminate: 36 GP: 11 G, 24 A, 35 pts [0.97 pts/gm]
Can be eliminated: 20 GP: 4 G, 12 A, 16 pts [0.8 pts/gm]

I can't total these because I counted game 7s as both. But, with the below...
Game 7s: 7 GP: 0 G, 6 A, 6 pts [0.86 pts/gm]

You can work it out.

Note: I do not consider the 2020 play-in round as playoff games. If one did, the Pens had one elimination game vs Montreal and Malkin did not have a point.

 

[Hockey Outsider]

I did a deeper dive because I was really surprised by Forsberg's numbers. As expected though, @pnep is right - Forsberg played in 41 elimination games during his career and only had 29 points (12 goals, 17 assists).

He had 21 points in 20 wins, and just 8 points in 21 losses. Or looking at it another way - Colorado went 5-14 when Forsberg was scoreless in elimination games, 10-6 when he had one point, and 5-1 when he scored two or three points.

This wasn't "skewed" by his post-lockout peak (28 pts in 38 games before the 2005 lockout, 1 pt in 3 games after).

He had some huge games of course (the biggest moment was probably in 2002, when he scored the final two goals of the series against San Jose to help the Avalanche come back from a 3-2 series deficit) - but there were way more scoreless games than I remember (including one stretch with three assists in nine consecutive elimination games - actually ten consecutive elimination games, one of which he missed). That's not to place all the blame on Forsberg of course (they ran into some very good goalies including Joseph in 1998, Belfour in 1999, and Nabokov in 2004, and Sakic had his share of scoreless elimination games, but appeared to be more consistent).

One more point - of the 38 elimination games he played in before the lockout, Colorado had the series lead 22 times, their opponents only 8 times, and there were 8 game sevens. You can probably argue that, if not for Forsberg's scoring heroics in all of the other games, Colorado wouldn't have been in such a favourable position so often.

 

[OrrNumber4]

What are Jumbo Joe's numbers?

 

[The Macho King]

I did a deeper dive because I was really surprised by Forsberg's numbers. As expected though, @pnep is right - Forsberg played in 41 elimination games during his career and only had 29 points (12 goals, 17 assists).

He had 21 points in 20 wins, and just 8 points in 21 losses. Or looking at it another way - Colorado went 5-14 when Forsberg was scoreless in elimination games, 10-6 when he had one point, and 5-1 when he scored two or three points.

This wasn't "skewed" by his post-lockout peak (28 pts in 38 games before the 2005 lockout, 1 pt in 3 games after).

He had some huge games of course (the biggest moment was probably in 2002, when he scored the final two goals of the series against San Jose to help the Avalanche come back from a 3-2 series deficit) - but there were way more scoreless games than I remember (including one stretch with three assists in nine consecutive elimination games - actually ten consecutive elimination games, one of which he missed). That's not to place all the blame on Forsberg of course (they ran into some very good goalies including Joseph in 1998, Belfour in 1999, and Nabokov in 2004, and Sakic had his share of scoreless elimination games, but appeared to be more consistent).

One more point - of the 38 elimination games he played in before the lockout, Colorado had the series lead 22 times, their opponents only 8 times, and there were 8 game sevens. You can probably argue that, if not for Forsberg's scoring heroics in all of the other games, Colorado wouldn't have been in such a favourable position so often.

I kind of wonder if there's a mental aspect to game 7s where teams just... completely lock it down, and play a bit more worried to give up a goal versus scoring a goal.

Just thinking about Tampa - their last two game 7s were 1-0 wins, but watching them the ending wasn't really in doubt. Just the team as a whole clamped down in a way that meant there wasn't a whole lot of scoring chances either way.

 

[CupInSIX]

The problem with tracking elimination game stats is that it feels like a lot of them are blowouts by the 3rd period. Points contributing to GWG, GTG & non-EN insurance goals would tell more of the story.

 

[Steerpike]

I feel like this line of inquiry is silly. The playoffs are already short. Scoring in the playoffs is going to be hard regardless of whether it's an elimination game, overtime, ect.

Shrinking your sample size down to just the "deciding" moments and goals does injustice to the effort which got the team to that deciding moment. It also creates these pointless definitions of "clutch". A random player scores one important goal and now they are forever clutch because there are extremely few moments in which you can demonstrate clutchness.

You might as well just read through a list of Stanley Cup Clinching goals.


Now, If you really insist on thinking the game is fundamentally different during elimination games, you need to broaden out the set of events the impress you in order to get the sample size back into meaningful territory. Look at Corsi or Fenwick during those games to get a sense of who is playing well.

If you want to make a statement like "Mattews/Marner/Thornton/Ovechkin aren't good when it really matters" you need to show that the metric you're judging them by is actually predictive. Otherwise "so and so can't get it done" is really "so and so has gotten vaguely unlucky and my list of good players is limited to Maxime Talbot because he scored a Stanley Cup winning goal"

 

[Hockey Outsider]

I feel like this line of inquiry is silly. The playoffs are already short. Scoring in the playoffs is going to be hard regardless of whether it's an elimination game, overtime, ect.

Shrinking your sample size down to just the "deciding" moments and goals does injustice to the effort which got the team to that deciding moment. It also creates these pointless definitions of "clutch". A random player scores one important goal and now they are forever clutch because there are extremely few moments in which you can demonstrate clutchness.

You might as well just read through a list of Stanley Cup Clinching goals.


Now, If you really insist on thinking the game is fundamentally different during elimination games, you need to broaden out the set of events the impress you in order to get the sample size back into meaningful territory. Look at Corsi or Fenwick during those games to get a sense of who is playing well.

If you want to make a statement like "Mattews/Marner/Thornton/Ovechkin aren't good when it really matters" you need to show that the metric you're judging them by is actually predictive. Otherwise "so and so can't get it done" is really "so and so has gotten vaguely unlucky and my list of good players is limited to Maxime Talbot because he scored a Stanley Cup winning goal"

I agree with most of your comments. We're dealing with small sample sizes, and a few good or bad plays can have a huge impact on a player's reputation. (Steve Smith played more than 800 regular season games, but most people remember him for a single bad play).

I think looking back on past events can be descriptive (ie accurate in describing what actually happen). For example, and I don't know if this is true, but hypothetically, if Joe Thornton only scored 0.5 points per game in games where his team could be eliminated, I think it would be factually correct to say that his offense disappeared, and this contributed to the Bruins/Sharks/etc not meeting expectations. But you can't point to McDavid (or Matthews, or whoever) mid-career and say that they're inherently not a clutch player and never will be.

The other point is we need to be careful and avoid over-analyzing the data. Patrick Roy is generally (and rightfully) regarded as one of the greatest playoff performers in NHL history. But he has a losing record in game 7's. The reality is 15 teams will be eliminated each spring, so a goalie with a long playoff career will (eventually) lose a lot of elimination games. Roy's record in game 7's is disappointing, but had he been a weaker goalie, his teams would have been eliminated in 4, 5 and 6 games more often. Placing undue focus on his losing record in game 7's could distract from the fact that he was really good in the rest of most of his series.

 

[Bear of Bad News]

I don't really see anyone ascribing abilities here - I mean, you were even proactively careful when you said "Let's not turn this into a discussion about what these numbers mean".

 

[daver]

These numbers could be used to differentiate players with similar production but not to elevate someone up a level. Especially if we are talking career numbers.

If someone wants to put a bit more weight on a series clinching goal in Game 7 vs. a gamewinning goal in Game 1 that seems reasonable. As are points against a seemingly more difficult opponent or in the later rounds of the playoffs.

A player like Kane has earned a rep for his clutchness but this, IMO, doesn't mean you take him over Crosby who simply will give you more offense (among other things). Helping your team win in 6 games with more points > scoring less points but scoring the Game 7 OT winner as unless it is the SCF, it gives your team a chance to rest and play one less playoff game.

 

[pnep]

Scoring and W/L % in playoff elimination games (PO Seasons 1918-2021)

updated stats

 

[pnep]

ELIM_PO.xlsx

 

[Michael Farkas]

This has been very distracting to me at work today...what a treasure this is...

One thing (among many) very interesting....
Worst. Joe Thornton -31 (57 GP)
2W. Henrik Sedin -20 (33 GP)
3W. Dino Ciccarelli -20 (45 GP)
4W. Mike Allison -18 (26 GP)
5W. Troy Murray -18 (31 GP)
6W. Joe Juneau -18 (31 GP)
7W. Daniel Sedin -18 (33 GP)
8W. Bobby Smith -18 (55 GP)

 

[sr edler]

I feel like this line of inquiry is silly. The playoffs are already short. Scoring in the playoffs is going to be hard regardless of whether it's an elimination game, overtime, ect.

Shrinking your sample size down to just the "deciding" moments and goals does injustice to the effort which got the team to that deciding moment. It also creates these pointless definitions of "clutch". A random player scores one important goal and now they are forever clutch because there are extremely few moments in which you can demonstrate clutchness.

You might as well just read through a list of Stanley Cup Clinching goals.


Now, If you really insist on thinking the game is fundamentally different during elimination games, you need to broaden out the set of events the impress you in order to get the sample size back into meaningful territory. Look at Corsi or Fenwick during those games to get a sense of who is playing well.

If you want to make a statement like "Mattews/Marner/Thornton/Ovechkin aren't good when it really matters" you need to show that the metric you're judging them by is actually predictive. Otherwise "so and so can't get it done" is really "so and so has gotten vaguely unlucky and my list of good players is limited to Maxime Talbot because he scored a Stanley Cup winning goal"

But who is claiming guys like Talbot, Rupp, Krupp, Kaberle, Moen, Eller and Colton are these greatly historically clutch players in the sport just because they scored a Cup clinching goal? I don't think I've seen this anywhere.

When people talk about clutch playoff performers they usually try to map out a larger sample size, and you don't even necessarily have to win a Cup or gold medals to be considered a clutch player. Trevor Linden for instance never won a Stanley Cup but his elimination games stats are great over an extended sample size, and it's also verified by any reasonable eye test. Actually his clutch game extended into the Olympics where he sent Canada into OT against the Czechs in 1998 with a late game-tying goal, a game his team still lost in the shootouts.

Saku Koivu is another guy I would consider clutch in the playoffs, because the eye test and his stats supports this, despite his not-so-very-good-at-all team never moving further than the 2nd round of the Stanley Cup playoffs. And Koivu also always brought it in the Olympics where he medalled 4 times as a key player with an underdog Finnish team.

Of course you always have to consider nuance and circumstances and context, but to claim there's no such thing as clutch players removes the human element from the picture, IMO.