|
Who needs Assists? Not “Superman” Rick Nash! After all, the Blue Jackets’ winger scored an unassisted hat trick against Detroit on Saturday. The feat helped to round out his evening of outrunning a speeding bullet and jumping over a couple of tall buildings in single bounds. Note to Detroit: You might want to think about reining it in a bit on any Friday nights activities – You’ve been plastered 16-2 the last two Saturdays and 21-4 the last three Saturdays.
All kidding aside, the unassisted hat trick is one of the rarest accomplishments in sports. In fact, eleven perfect games have been thrown in Major League Baseball since Montreal’s Maurice “Rocket” Richard last recorded an unassisted hat trick in 1958. On topic for this week – As unassisted Goals themselves are uncommon to begin with, we know there are plenty of Assists to go around.
Last week, in our ongoing quest for optimal fantasy performance, we looked at the best measures for predicting future performance in Goals. This week we will try to do the same for Assists. Once again, the question we are looking to answer is: What number of prior fantasy weeks’ performance gives the best indication of the coming week's performance?
Before we dive into what this week’s research uncovered to answer that question, let’s mull over the nature and value of the Assist for a moment. An Assist is given to the last two players on the Goal-scoring team to touch the puck before the Goal scorer. If, after possession of the puck is gained, less than two players touch the puck before the Goal is scored, less than two Assists are given.
The value of an Assist towards a team’s success is often debated. In mainstream hockey statistics, Points are the total of Goals plus Assists, inferring that Goals and Assists are of equal value. This has never seemed quite right to me. In contrast, Tom Awad’s GVT system considers a Goal to be 1.5 times the value of an Assist, a more reasonable assumption. In our TUFNEL fantasy league, we give Goals 3 times the value of Assists, keeping in mind that two Assists are most often tallied per Goal and that we make deductions for Shots On Goal. For your fantasy league, you will need to put those philosophical questions aside and value Assists per your league’s point system. Now, let’s figure out how to best predict how many Assists your players will record during the next fantasy week.
For Goals, we learned that the last 8-10 weeks’ performance was the best indicator of future performance. Settling on the last 9 weeks, or the last 27 games, we then compared perceptions of actual players in contrast to their last 3-6 games, to see where we could find better value than what may be perceived by other fantasy owners.
Coefficient of Variation – Goals and Assists: Past x weeks as a predictor for next week
Number of previous weeks of data
Top 35 1 2 3 4 5 6 7 8 9 10 YTD
Goals, CV 77% 62% 58% 56% 55% 54% 53% 50% 51% 50% 57%
Assists, CV 62% 52% 48% 46% 45% 43% 43% 43% 43% 43% 47%
The table above shows last week’s data averaging the top 35 players in Goals year to date along with new data averaging the top 35 players in Assists year to date. Once again, the measure used is the Coefficient of Variation (CV), which is the standard deviation divided by the average, a normalized statistic that allows equivalent comparisons between all players, regardless of scoring level.
What is immediately apparent is that there is less variability in predicting Assists based on past performance than Goals. For instance, when looking at the last 1 week’s performance, the CV for Goals is 77%, while the CV for Assists is 62%. Therefore, using last week’s Assist total as a predictor for next week’s Assist total is more likely to be accurate than using last week’s Goal total as a predictor for next week’s Goal total.
The conclusions regarding the best predictors for Assists, in number of previous weeks, are nearly identical to those for Goals. As with Goals, the last 1 week's performance had by far the least correlation to next week's number of Assists (62% CV). Likewise, the last 2 weeks’ performance was not as accurate as the other sorts (52% CV). Predictions based on the last 3 weeks, last 4 weeks and Year To Date were essentially equal (48% CV, 46% CV, 47% CV), while the best predictors for Assists were the last 6-10 weeks’ performance (all at 43% CV). This is good news for the fantasy owner, as the last 4 weeks data (46% CV) that you likely have available at your league’s website is within spitting distance of the marginally better last 6 weeks’ data (43% CV).
Last time, we looked at Sidney Crosby, Evgeni Malkin and Alex Ovechkin to see how their Goal scoring variation contrasted with that of the other top Goal scorers. Let’s look at the three superstars again to see how they vary in CV for both Goals and Assists:
Coefficient of Variation - Goals
Number of previous weeks of data
Player 1 2 3 4 5 6 7 8 9 10 YTD
Crosby 115% 87% 86% 87% 82% 84% 70% 54% 42% 43% 80%
Malkin 82% 67% 58% 51% 52% 53% 51% 53% 49% 46% 56%
Ovechkin 63% 45% 46% 47% 44% 41% 37% 38% 38% 39% 45%
Top 35 77% 62% 58% 56% 55% 54% 53% 50% 51% 50% 57%
Coefficient of Variation - Assists
Number of previous weeks of data
Player 1 2 3 4 5 6 7 8 9 10 YTD
Crosby 48% 40% 44% 41% 41% 37% 37% 33% 35% 34% 39%
Malkin 46% 34% 31% 32% 32% 29% 31% 31% 30% 32% 37%
Ovechkin 52% 51% 47% 46% 44% 40% 40% 40% 38% 36% 49%
Top 35 62% 52% 48% 46% 45% 43% 43% 43% 43% 43% 47%
An immediate observation that can be made is Crosby’s wild variability as far as Goal scoring, but his steady consistency with Assists. Sid is more consistent with his weekly output of Assists than the average player, but significantly less consistent regarding Goals. On the other hand, while Ovechkin is slightly more variable in his Assist performance than Crosby and Malkin, his predictability in Goals and Assists are unusually similar to each other. What does that mean to you as a fantasy hockey player? Not much, as the good news is that when you look at Goals by the last 9 weeks and Assists by the last 4 or 6 weeks, you will get a fairly accurate estimate of how Crosby and Ovechkin, and Malkin and most other players, will perform in the next fantasy week.
Now it’s time to look at which players may currently be undervalued or overvalued with respect to Assists, by comparing the performances of the last 3-6 games against the much better measuring stick of the last 12 games. This is a sort you will likely have available at the fantasy website of your choice.
Undervalued – Better performance in previous 12 games than in previous 3-6 games
Assists Assists/week
Previous x games Previous x games
Player Team Pos 3 6 12 3 6 12
Mark Savard BOS C 2 6 10 2.0 3.0 2.5
Todd White ATL C 2 5 10 2.0 2.5 2.5
Craig Conroy CAL C 1 5 10 1.0 2.5 2.5
Mark Recchi BOS RW 0 7 8 0.0 3.5 2.0
Martin Havlat CHI LW 1 4 11 1.0 2.0 2.8
Steve Sullivan NAS RW 1 5 6 1.0 2.5 1.5
Pavel Kubina TOR D 1 1 7 1.0 0.5 1.8
Cory Sarich CAL D 0 2 6 0.0 1.0 1.5
-Over the last month, Mark Savard has been a mix of good and bad. While the Bruins’ All-Star shockingly has no Goals since February 7th, he continues to tally Assists at an excellent clip. His last 12 game pace of 2.5 Assists/week is in line with what you can continue to expect from him.
-Atlanta’s Todd White may be a free agent in leagues that do not value Assists highly. The former Golden Knight is a very steady performer as far as Assists are concerned and is worth picking up if you are thin at Center.
-Fellow Clarkson alum Craig Conroy may be useful in deeper leagues. His recent 2.5 Assists/game is excellent, but keep your eye on any change in his role in light of the Flames’ roster shuffle of Olli Jokinen for Matthew Lombardi.
-Mark Recchi has 537 Goals and 891 Assists in his two decades long NHL career. Recchi is a useful pickup in many leagues, having posted 2.0 Assists/week over his last 12 games, with the potential for improved production now that he is on one of the top lines in Boston.
-Here’s what I had prepared for Martin Havlat: “Keep Martin Havlat active if he is on your roster. The 7 year veteran is finally healthy –knock on wood– this year and on pace for career highs of 73 Points and 81 games. The Blackhawks’ winger has remade himself this year, with 64% of his Points coming from Assists as opposed to 56% for the rest of his career.” He’s injured now, so disregard that, except for the “knock on wood” part. Keep an eye on his status.
-Coming back from serious injury, Steve Sullivan premiered this season on January 10th, with consistent playing time coming since February. Sullivan should be good for at least the 1.5 Assists/week that he has achieved over the last 4 weeks, with the potential for more with good health and the Predators’ improved play as of late.
-Pavel Kubina was listed as overvalued list week in regards to Goals, but he is similarly undervalued regarding Assists, with a superb 1.8 Assist/week clip recently. Picking up the Maple Leafs’ Defenseman makes most sense in leagues that minimize plus/minus.
-Calgary’s Cory Sarich is another Defenseman worth taking a look at. Sarich has averaged 1.5 Assists/week over the last 4 weeks, regardless of having tailed off over the past 2 weeks.
Overvalued – Better performance in previous 3-6 games than in previous 12 games
Assists Assists/week
Previous x games Previous x games
Player Team Pos 3 6 12 3 6 12
Steve Ott DAL LW 3 6 7 3.0 3.0 1.8
Matt Cullen CAR C 5 5 5 5.0 2.5 1.3
Erik Cole CAR LW 5 5 5 5.0 2.5 1.3
Dan Hamhuis NAS D 1 4 4 1.0 2.0 1.0
-Deeper leagues and leagues that reward heavily for PIM can take a look at Steve Ott. The Stars’ agitator has been playing well since January and is currently putting up 1.8 Assists/week.
-Matt Cullen is useful in some leagues, but don’t expect 2.5 Assists/week from him, as the Hurricanes recent roll is not sustainable.
-Erik Cole is back in Carolina, where he was a member of the 2002 and 2006 Stanley Cup Finals teams, after his brief stay in Edmonton. Though overvalued by his last 1-2 weeks’ performance, Cole should be able to improve upon his anemic 1.3 Assists/week after his return to familiar settings.
-Defenseman Dan Hamhuis last appeared on my fantasy roster a couple of years ago. Occasionally, he will look like a useful pickup to the untrained eye when on a brief hot streak.
Having deciphered the mystery of predicting both Goals and Assists now, you should have significantly improved tools to optimize your free agent pickups and your weekly lineup choices. We will look to fill in some other key pieces of information in the coming weeks, as the fantasy playoffs are here for most leagues.
As for myself, I’m heading to the fantasy semifinals after a tough matchup against “The Lockouts” this past week. Thanks, Rick Nash!
If you’re still alive in your league as well – Good luck!
Timo Seppa is an author of Hockey Prospectus.
You can contact Timo by clicking here or click here to see Timo's other articles.
| 
0 comments have been left for this article. 
Conn Smythe Watch: Lundqvist Leads