<
>
Insider

Stats insights, warnings for draft forwards

8/5/2013

Analytics at the prospect level has been an interesting area the last few years, and there seems to be growing interest. In preparation for the 2013 NHL draft, several NHL teams hired stat guys and consulting groups to help them with their prospect analysis, and some teams are even having their video guys track micro stats at the junior level.

For most of the lower levels of hockey, many studies have shown there is notable predictive value in stat-based analysis for prospects. The main issues facing analytics for junior, collegiate and other levels is adjusting for context and trying to separate the relationship between production and the typical uncertainty/miss rate of projecting prospects. After all, stats can't tell you everything about prospects -- they need to be adjusted for a variety of factors, they can't replace scouts and they are subject to statistical noise. However, if used correctly for the NHL draft, stats can help teams find some overlooked sleepers and may prevent them from making a glaring mistake.

Here are some notable examples of forwards whose stats told a much different tale than their draft position in the not-too-distant past, and the team to which they were drafted.

Benoit Pouliot, LW, Minnesota Wild

When it comes to analyzing draft prospects -- and especially CHLers -- there is one golden rule: Look at the birth date. Pouliot, the fourth overall pick in 2005, was a late birth date, born in November 1986, and thus was not drafted until the next season with the 1987 class. When comparing future output by players in similar circumstances in the CHL -- and with one having a late November birth date and the other a January birth date -- the January-born player traditionally produces at a 15 percent better rate in the NHL. That extra season of hockey experience, particularly in a lesser league (relative to professional ones) like the CHL, makes a significant difference. It can give the player an edge over his competition and make him look better than he really is. Such was the case with Pouliot.

Now, in Pouliot's only CHL season before his draft selection he was the OHL Rookie of the Year. Among the OHL's "87s" -- players eligible to be drafted in 2005 -- Pouliot's 65 points in 67 games ranked fifth behind Bobby Ryan, Dan Ryder, Steve Downie and Bryan Little. But in terms of age (and physical development) Pouliot was really closer to players born in 1986 and drafted a year earlier. If you add in 1986 birth dates, Pouliot ranked 13th in OHL scoring, with players like Rob Schremp, Wojtek Wolski and Dave Bolland also besting him.

In their NHL careers, Bolland, Downie, Little and Wolski have all produced higher points-per-game averages than Pouliot, with most being around 15 percent higher. Pouliot's size coupled with high-end speed and skill is a tantalizing set of attributes, but a player with his late birth date failing to produce better numbers in the CHL should have set off red flags, especially as a top-five selection.

Mike Cammalleri, LW, Los Angeles Kings

For the most part, especially when drafting forwards, NHL scouts have proved to be an efficient group who identify and take the top forwards early. The one most notable inefficiency from the last two decades has been with small forwards, and Cammalleri is a prime example of that. As an 18-year-old forward playing college hockey, his 61 points in 42 games gave him a per-game pace on par with Dany Heatley (who was a year older) that season. Among under-19 players in college, his points per game were 0.30 PPG higher than anybody. Per Hockey Prospectus' Robert Vollman, the CCHA -- where Cammalleri played -- is a 20 percent tougher league than junior hockey, so adjusting his scoring pace upwards from where junior players were gives you a good idea of how well he produced. Stephen Weiss, a top-five pick that season, scored around a 1.30 PPG pace in the OHL. Given Cammalleri's high skill level, great shot and tremendous production, there was a strong argument he should have been a first-round pick and not a mid-second-round selection, even with his small stature.

Marc Savard, C, New York Rangers

Continuing off the Cammalleri thought process, Savard is another -- if not the -- prime example of skilled, small players being undervalued in the draft, especially during the low-scoring era. One didn't need to make a lot of fancy tricks or calculations to see Savard's skills relative to his peers. For example, as a 17-year-old he led the OHL in scoring. Among under-18 players, the next best player scored nearly 60 fewer points and almost a point per game worse than Savard. As an under-17 player, he led the field by 20 points. If you combined the 1995 OHL scoring of Bryan Berard and Chad Kilger, who were top five picks in Savard's draft, it was only 30 more points than Savard recorded that season. Concerns about Savard's size and all-around game were reasonable, but with his skill and hockey sense it's hard to make a good argument why he should have lasted until the fourth round.

Milan Hejduk, RW, Colorado Avalanche

In the 1993-94 season, Hejduk played in the Czech Extraliga and notched 11 goals and 15 points in 32 games. The Czech youth program and their top league has lessened in quality in recent years, but in the 1990s the top Czech league was arguably the fourth-best league in the world behind the NHL, as well as Russia and Sweden's top pro leagues. Between 1990 and 1998, all players who went to the NHL after playing in the Extraliga retained 50 percent of their scoring production, even after the step up in competition. However, the average age in that sample was 22, whereas Hejduk was only 17 that season. Over that same time period, players retained merely 25 percent of their scoring from the CHL. Hejduk was a 0.5 point-per-game player in the Czech league, but he arguably could have been over a point per game in the CHL using league equivalencies adjusted for age. If you believe his skill and goal-scoring could translate to the North American ice, he arguably should have gone a couple of rounds higher at the 1994 NHL draft, where he was eventually picked 87th overall.

Colton Gillies, LW, Minnesota Wild

There is a legitimate argument for exactly how much merit stats can have in junior player analysis, especially when players are very clustered in good scoring numbers. However, there is not a good recent historical record of CHL players drafted at age 17 with 42 career points in nearly 130 games. Barring some major injury issues -- or a team so deep that high-end prospects can't get ice time -- those scoring numbers are a giant red flag. When Gillies was drafted 16th overall in 2007, it was obvious what he was: a big, physical forward who kills penalties and skates well, with little offense and a lot of time in the penalty box. Those are dangerous guys on whom to spend high picks, because of their limited ceiling. Of course Gillies kept the physical attributes, but the lack of hockey skills limited him in the NHL.