2011-2012 NHL Leaders & Losers in A1 vs. A2

[Pengwin7]

When looking at player performances, I always take a look at several things that could fluctuate:

1. Shooting percentage (SH%). Did a player shoot much higher or lower than their career average? If so, adjust predictions accordingly for a more normal shooting percentage.

2. Powerplay points (PPP). Did a player's team struggle on the PP? Many times PPs just click or fizzle in a given year and then the addition or subtraction of one player (say, Jagr?) can completely re-work the entire PP efficiency.

A good 3rd one that doesn't show up in common statistics is assists: 1st or 2nd assists.
Did a player get a lot of 2nd assists? (I like to call these "token" assists)
Or is the player a bonafide 1st assist kind of guy.

Here is last year's top players listing a RATIO of A1:A2 (both in percentage)
The list includes the 464 players that averaged more the 0.2assist1 & 0.2assist2 per 60minutes. (In other words, guys that averaged at least 0.4 assists per game).

BEST (A1:A2)
84:16 Perry, Corey (ANA) [*19 first assists & 4 second asssists for Perry last year... indicates his assist total SHOULD increase with more A2.]
84:16 Stafford, Drew (BUF) [*25 first assists & 5 second asssists for Stafford last year... indicates his assist total SHOULD increase with more A2.]
83:17 Kennedy, Tyler (PIT)
81:19 Versteeg, Kris (FLA)
80:20 Tanguay, Alex (CGY)
80:20 Malkin, Evgeni (PIT) [*As an example, this means that of Malkin's 59 assists last year, 47 were 1st assists and ONLY 12 were 2nd assists. Wow!]
79:21 Selanne, Teemu (ANA)
78:22 Burrows, Alex (VAN) [*Burrows was also one of the most Assist2 heavy guys in 2010-2011... so it shows this statistic is unreliable and CAN fluctuate.]
78:22 Couture, Logan (SJ)
75:25 Nash, Rick (CBJ)
75:25 Staal, Jordan (PIT)
75:25 Cullen, Matt (MIN)
75:25 Cammalleri, Mike (CGY)
75:25 Whitney, Ray (PHX)
74:26 Kane, Pat (CHI)
74:26 Bergeron, Patrice (BOS)
74:26 Alfredsson, Daniel (OTT)
74:26 Pavelski, Joe (SJ)
73:27 Hejduk, Milan (COL)
Others above 70:30 (Craig Smith, Grabovski, Taylor Hall, Tavares, E.Staal, Antropov, Oshie, Stepan, Gerbe, Spezza, A.Steen, Stamkos, Zetterberg)

******
About 80% of the NHL will fall between 70:30 & 40:60 ratio split
******

40:60 Ehrhoff, C. (BUF)
40:60 Landeskog, G (COL)
40:60 Carle, Matt (PHI)
40:60 Read, Matt (PHI)
38:62 Campbell, Brian (FLA)
37:63 Hamhuis, Dan (VAN)
35:65 Bogosian, Zach (WPG)
35:65 Seabrook, Brent (CHI)
35:65 Benn, Jamie (DAL)
34:66 Morrow, Brenden (DAL)
33:67 Elliott, Stefan (COL)
33:67 Connolly, Brett (TB)
33:67 Nikitin, Nikita (CBJ)
32:68 Bieksa, Kevin (VAN)
28:72 Wolski, Wojtek (FLA)
24:76 Boyle, Dan (SJ) [*For Boyle, this works out to his 39 assists as: 9 first assists and 30 second assists.]
17:83 Foligno, Marcus (BUF) [*Really, it's a small sample size. Foligno only had 7 assists and this is the ratio for having one 1st assist & six 2nd assists.]
WORST A1:A2


Getting a ton of 2nd assists isn't necessarily a bad thing. It's the pass that sets up the pass.
Often, the sign of a lot of 2nd assists can be a complimentary player on a good powerplay (Hamhuis & Bieksa: VAN, Read & Carle: PHI, Boyle: SJ).

But there are some curious players on there where you have to raise an eyebrow.
Marcus Foligno & Nikita Nikitin were a few players I saw listed as great sleeper picks in their drafts... but how lucky were those assists?
Were they just luckily riding the coattails of their teammates?

Also - I now have some concern on a few players that I wouldn't have thought otherwise about: Jamie Benn & Gabriel Landeskog.
We always point to a good young player and expect their point totals to rise... but if they had a boatload of token 2nd assists... will they rise for sure?

We can also think... hmm, if those guys were picking up all the 2nd assists... who was picking him up the 1st assists in COL & DAL:

Answers:
in COLORADO
73:27 Hejduk, Milan
68:32 O'Reilly, Ryan
67:33 Stastny, Paul

in DALLAS
69:31 Dvorak, Radek
68:32 Ribeiro, Mike
59:41 Eriksson, Loui


If you are curious about any other players Assist1:Assist2 ratio just ask. Maximum of five players.

I get these stats from behindthenet.ca & move them into a spreadsheet and then ratio by A/(A+B).

[Skin Blues]

I went back 3 years and found the average standard deviation of all players that played at least 60 games in each of the past 3 seasons. The average standard deviation for A1s was 0.16, and for A2s was 0.13. Not a huge difference, but it suggests that guys who put up big A1 totals are more likely to decrease their assist totals than those with high A2 totals. Counter intuitive, but it's what the numbers show. Even looking at the 25 best assist men over that 3 year span, the difference in standard deviation is even more pronounced (0.23 to 0.16).

Guys like Daniel Sedin, David Krejci and Alex Ovechkin have huge variations from season to season. Daniel and Ovi in particular have had a huge decline in primary assists while their secondary assists have not seen the same plunge.

This is of course highly dependent on line mates, so a guy that goes from playing with superstars to scrubs, or vice versa, will not follow the patterns of the average player. I would imagine that A1s correlate to talent and drop with age, while A2s correlate to the players you're surrounded by and are less affected by age.

[kevinsrangers]

nice work. rep for ya both.

[Pengwin7]

I went back 3 years and found the average standard deviation of all players that played at least 60 games in each of the past 3 seasons. The average standard deviation for A1s was 0.16, and for A2s was 0.13. Not a huge difference, but it suggests that guys who put up big A1 totals are more likely to decrease their assist totals than those with high A2 totals. Counter intuitive, but it's what the numbers show. Even looking at the 25 best assist men over that 3 year span, the difference in standard deviation is even more pronounced (0.23 to 0.16).

Guys like Daniel Sedin, David Krejci and Alex Ovechkin have huge variations from season to season. Daniel and Ovi in particular have had a huge decline in primary assists while their secondary assists have not seen the same plunge.

This is of course highly dependent on line mates, so a guy that goes from playing with superstars to scrubs, or vice versa, will not follow the patterns of the average player. I would imagine that A1s correlate to talent and drop with age, while A2s correlate to the players you're surrounded by and are less affected by age.

It's great to have somebody else to track these.
In general, A1 & A2 are likely "self-correcting statistics".

But I think we can extract some value from examing A2 counts.
My intuition is indeed that I should be:
*pessimistic about players with unusually high A2 counts
*optimistic about players with unusually low A2 counts

Two small studies:

1. Let's look at the incredible drop in Corey Perry's points.
2010-2011: Perry had 98pts. 50G, 48A (26 1sts, 22 2nds), 17.2% SH on 290 SOG, 31 PPP
2011-2012: Perry had 60pts. 37G, 23A (19 1sts, 4 2nds), 13.4% SH on 277 SOG, 23 PPP

There was a 38pt drop there from the Hart year to last year.
Obviously, the goals represent a 13pt decrease... but there's also the 25pt assist decrease.

Q: Should we be worried about that?
A: Probably not. Perry had an "unlucky" year in only getting FOUR (just FOUR) 2nd assists. A big portion of that is probably due to overall flow of the ANA PP with Lubo injured. (and Perry's PPA count decreased from 17A to 9A, even though his PPG count maintained at 14 PPG.)

Perry's 37G is pretty on-target for his current situation & skill level.
But I think we can expect Perry to put up about 20-22 1st assists & about 13-15 2nd assists.

Forecast for Perry: 70-74pts (37G & 33-37A)... and this is pretty consistent with Perry's numbers in 2007-2008 & 2008-2009.

2. Now, how about a player who we do NOT have a lot of data on:
2nd year player Gabriel Landeskog
Landeskog had a really nice year last year with 52pts and +20.
It's hard not to like the kid.

He got 270 SOG and shooting 8.1% resulted in 22G. (6 PPG). [An important aside: Without a 2nd NHL season from GL, I'm not sure if 8.1% is relatively low or high for his sniping skill level. ?]
Of his 30 assists, what worries me is that there were 12 first assists & 18 second assists. Sometimes it's not so much a worry, there were probably several Avalance (O'Reilly, Landeskog) scoring lines.

But we all see Landeskog with a great 52pt season and many are probably expecting him to make a typical 20%+ sophmore jump to around 62-65pts. Heck, Seguin (#2 overall pick) jumped from 22pts to 67pts.

While I like Landeskog, I would probably bet that he doesn't hit 60pts in his next full-season... simply because he is unlikely to get 18 second assists.
(As a reference, Evgeni Malkin only had 12 second assists last year of his 59A).
I'm not bashing Landeskog, I love him, and I'd expect him to raise his goals from perhaps 22 to 27. I'd also expect his first assists to raise from 12 to 20. So that's 47pts on goals & first assists. I would expect the second assists, however, to be more around 10-12... and this would make Landeskog a 57-59pt player.



Of course, (in honesty) - these are picking a few examples that "work".
Alex Burrows had an unusually high A2 total in 2010-2011 riding 3rd-wheel with the Sedins. Of Burrows 22 assists, only FIVE were 1st assists and the other 17 were 2nd assists... so I would've pegged him for a drop last year. Instead, Burrows put up 24 assists, with 19 first assists and only FIVE second assists.


Anyways - all of this is mostly just discussion (and mostly because its JUST FREAKIN SLOW HERE ON THE BOARDS... DAMN LOCKOUT).
It seems intuitive that there will be cases where we should be looking at high 2nd assist totals and predicting a player for a relative-decrease (Landeskog)... and that we should be looking at low 2nd assist totals (Perry) and predicting a player for an increase.

I'm not sold myself yet... it feels a little bit like the PDO stat in the Dobber Guide, which - sorry to say, I don't feel is necessarily a telling or helpful statistic.
Sometimes the research just had to be discussed, evaluated, and then (unfortunately) thrown out as a good-try and not-of-significant-value.

[27Blue]

i think, like shooting percentage, you will want to look at a player's career as opposed to just last season... not convinced there is much to it, like PDO, but i think a bigger sample for each player would be a necessary starting point...

[Pengwin7]

but i think a bigger sample for each player would be a necessary starting point...

true.
but the case-in-point is that people make decisions on players based on what they did in the previous season.
so if a player had an usually high-or-low a2 season, we may be able to correct that number to a more reasonable level.


I'm thinking of setting up some benchmarks of expectations for types of players:

Example:
BEST CASE: Player likely for Assist Increase? (Reason: He had an unlucky, low A2 season last year)

0%-20% A2: Regardless of player type, player has too few second assists and should be flagged as VERY LIKELY for overall assist increase. (ex. Corey Perry, Drew Stafford)
20%-30% A2: Passing should be player's top offensive skill. (If true, increase player's forecasted assist total significantly from original estimate)
30%-40% A2: Player should be primarily a passer. (If true, increase player's forecasted assist total mildly from original estimate)

40%-50% A2: Common ratio of second assists

50%-60% A2: Player should be primarily a scorer. (If true, decrease player's forecasted assist total mildly from original estimate)
60%-70% A2: Scoring should be player's top offensive skill. (If true, decrease player's forecasted assist total significantly from original estimate)
70%-100% A2: Regardless of player type, player has too many second assists and should be flagged as risk for overall assist decrease. (ex. Dan Boyle)

WORST CASE: Player likely for assist decrease? (Reason: He had a lucky, high A2 season last year)

[Skin Blues]

I think for Landeskog it's fair to assume that as he enters his prime his A1s will increase. It seems to be a trend, but I have no mathematical proof because I simply don't know how to do that. Maybe I could chart A1/A2 ratio vs Age. I'd assume there's a peak at around the 3rd/4th year in the league, or maybe the age 25/26 season.

As for Corey Perry, he's also had a drastic drop in A1s. 0.95 to 0.66 to 0.47 (per 60 mins) over the past 3 seasons. That's very concerning.

Then you have old guys like Ray Whitney and Kimmo Timonen just skyrocketing their A1 totals. I don't really know what to think about all this...

[Pengwin7]

As for Corey Perry, he's also had a drastic drop in A1s. 0.95 to 0.66 to 0.47 (per 60 mins) over the past 3 seasons. That's very concerning.

Then you have old guys like Ray Whitney and Kimmo Timonen just skyrocketing their A1 totals. I don't really know what to think about all this...

I like your dig on Perry.
I do believe there is some value in that.

I mean, we all know that Perry is a goal-scorer... so perhaps it's not so much a "decline" as it is that the 0.95 was WAY too high, and maybe the 0.66 way a bit too high or his average.
Often times a goal-scorer gets his rebound banged in = A1. (LUCK assist)
Or the defenses key on him so hard that he can slip in easy gimme tap-in to a teammates on the other side of the net = A1. (SKILL assist, from being overmarked.)

We all have a tendency to use words like "increasing" or "decreasing"... but I think it's also key to look at a player and just point out when a season was lucky or everything-clicked (like Perry's HART season). Goals are easiest to adjust. Playing with a great passer (Oates, for Hull) or just having a great shooting percentage season are very obvious things to adjust, statistically. But assist totals have always been the black-box of point-scoring-analysis & predictions.

The find on Whitney & Timonen may also have value.
I mean... these older guys may get slower... but their hands are as good as ever and their vision/timing/hockey-sense only improves with age.
This is part of the reason I don't downgrade old-guys too much... the old guys with "hockey-sense" are very valuable on the PP.

So, it's not necessarily a surprise to see their A1 stats/ratio improve.
Essentially - a passer can control his A1 output, more than he can control his A2 output.

A guy like Ray Whitney had played a few years with Doan & Vrbata (I believe those were his linemates) so every year he plays with them, the value in his hockey-sense allows him to anticipate how to make the perfect pass for those players.

Same goes for Timonen... another year with Giroux & a chance to play with a deft PP-guy like Jagr allows his hockey-sense to get maximize value by threading nice passes - hence, nice A1 count.

In any small sample size (one season), there will be a lot of skewed stats... but I do believe there is some value there - and I think you are uncovering some of it.!!

[The Guvnah]

80:20 Malkin, Evgeni (PIT)[/B] [*As an example, this means that of Malkin's 59 assists last year, 47 were 1st assists and ONLY 12 were 2nd assists. Wow!]

This is very impressive IMO. I'm wondering about some of the other assist leaders -- Sedin, Giroux, Karlsson, Thornton...

[27Blue]

This is very impressive IMO. I'm wondering about some of the other assist leaders -- Sedin, Giroux, Karlsson, Thornton...

2 years ago giroux had 31 assists and only 2 of them were secondary...

that's the highest ratio that i was able to find.

[Pengwin7]

Henrik Sedin 60/40 split: 40 first assists, 27 second assists.
Claude Giroux 59/41 split: 38 first assists, 27 second assists
Erik Karlsson 54/46 split: 32 first assists, 27 second assists
Joe Thornton 50/50 split: 30 first assists, 29 second assists

Joe Thornton is probably the only interesting case as we'd expect him to be first-assist heavy. San Jose's Joe Pavelski (74:26) and Logan Couture (78:22) were both incredibly first-assist heavy last year. That bodes well (IMO) for their futures. Jumbo Joe & Dan Boyle racked up several second assists - possibly on the PP... not sure.

San Jose would be an interesting team to look at first & second assist distribution. I think Frozen Pools can do this (but I don't have a subscription).

[Pengwin7]

2 years ago giroux had 31 assists and only 2 of them were secondary...

that's the highest ratio that i was able to find.

Mmhmm... it was a 95/6 ratio.
That was the 2009-2010 season, where he scored 47pts in 82 games (29 first assists, 2 second assists).

Most of us had Giroux pegged for only 50-60pts for 2010-2011... including some of DobberHockey's smartest minds. I had him pegged for 62pts (end of thread) - so even I missed way low. Little higher than some others though!

If we had all been crunching these A1 vs. A2 numbers after that season, we would have ALL bought Giroux prior to his 76pt break-out season in 2010-2011. I mean, getting 29 first assists in 16min of ice-time is pretty sensational. Something that maybe should've been dug on him or noted. I certainly would increase my predictions on a player if I knew they got 29 first assists to make up their 47pts. I mean... 29 first assists, that's indeed a 70pt player most of the time.

So... just another reason why this little study might hold some value!!!

[Maaaasquito]

Sorry Pengwin,

I read through this a few times and I still don't get what you're trying to drive at?

Are you saying that first assists is the key to success and identifying future trends/production?

[rempel1]

This is very interesting stuff. Let me see if I understand this correctly.

I have Logan Couture pegged to have another season around 67 - 70 points. Every year in his career his numbers have always been assist heavy accept the last 2 years. I'm curious if 2010/2011 was also a year where he didn't have many 2nd assists.

Last year he had 34 assists(78:22). That works out to 26 A1 and 8 A2. If I take 45% of the assists from last year, I get 15 A2 giving me 41 assists(7 more than last year).

With this info, should I be placing him to have a 74 - 77 point year?

[Pengwin7]

Sorry Pengwin,

I read through this a few times and I still don't get what you're trying to drive at?

Are you saying that first assists is the key to success and identifying future trends/production?

With points, we know that we look at things like ice-time (Even-Strength & PP), linemates, and maturation.

With goals, we all know that we can look at a combination of SOG & shooting percentage, and considerations to upcoming year playmaking-linemate quality.

With assists - we never really get into examining the statistics.
Maybe we should?

If a guy has 30 assists... what does it tell us?
Most of the time, I think we all just take assists and use the same number for future seasons.

But if we know a guy had 28 first assists & 2 second assists... I believe we can say "Wow, this guy must be an incredible passer to set-up goals - and perhaps he was very unlucky that he only got 2 second assists."

I think first assists is a fairly reliable stat.
It's something we should "count" on with a little more security.

I think second assists is a less reliable stat.
If a player had a high number of second assists, perhaps we should adjust his future assist totals to be a bit lower (with regard to second assists).
If a player had a low number of second assists, perhaps we should adjust his future assist totals to be a bit higher (with regard to second assists).


Claude Giroux's 2009-2010 vs. 2010-2011 breakout season seems to be a glowing example of the value in assist-analysis.
In that season, Giroux had 47pts, averaging 16:37 in ice time.
16 goals, 31 assists.
Most of us pegged him for 50-60pts in 2010-2011 because there were a lot of star players in Philly eating ice-time (M.Richards, J.Carter, Briere, JVR).

Giroux went BOOM and put up 76pts in 2010-2011, in just 19:24.

Q: Should we have seen that coming?
A: No (would be the common answer), but maybe the answer should have been yes.

1. Let's start by making an assuming that Giroux was going to get 19min/game. That would have extrapolated his point totals from 47 to 54pts.

2. Let's say that we give him an extra +5% bump just for maturation. That brings him up to 57pts. That's in the 50-60pt range that most of us pegged him at.

3. But here's the thing... Giroux's 31 assists break down to 29 first assists & 2 second assists. If first assists are the more reliable, then we could expect that Giroux would put up that total again + a few more (based on the ice-time increase and maturation). Perhaps 33 first assists would be fair. What about those second assists? Well, two (for any full-time player) is far too low. Historically, I'd guess that 2 of every 3 goals scored have a second assist... which would mean Giroux should be inline for about 20 second assists (not two). That's 18 additional points!!!

4. What does this equal?
We bump his 16G up to 20G. [He actually scored 25G]
We bump his 29 first assists up to 33A. [He actually scored 34 first assists on a 66:34 A1:A2 ratio]
We bump his 2 second assists up to 20A. [He actually scored 17 second assists]
That extrapolation prediction amounts to 73points.
He actually scored 76points in that break-out year. (a few extra goals on above-average shooting percentage)


So... while we all pegged Giroux for 50-60pts... perhaps there was enough information (the strong A1:A2 ratio) to have pegged Giroux for significantly more.

Back in that thread (which I'm sure you reviewed ), nobody mentioned the ratio of first assists or second assists... so nobody predicting 70+pts for Giroux really could use this basis to justify their high predictions.

Anyways - this is a good case to examine... and perhaps (going forward) we should spend some time every off-season looking at player's A1:A2 splits.



ps. You know I love your stuff... but I have a hard time believing you really didn't understand the direction of this thread.