For some time now, a number of people have been making the argument that it is a waste of resources for NHL teams to draft goaltenders, and to some extent defensemen, early in the draft. Instead, teams should draft forwards, because elite forwards are found only through the draft. There is a growing trend to propose extending this drafting strategy further. Some now believe that NHL teams should take very few goaltenders at all in the draft.
There are obvious parallels to hockey. Traditional philosophy is “build from the net out”, just like in baseball, “pitching wins pennants.”
— 67Sound (@67sound) August 12, 2014
However, there’s growing data defencemen and goalie prospects are far harder to project than forwards. Just like pitchers vs. hitters.
— 67Sound (@67sound) August 12, 2014
A smart team might consider focussing on drafting almost exclusively Fs. Then get D by grabbing young talent with shitty PDO but good CF%.
— 67Sound (@67sound) August 12, 2014
@coreypronman I wouldn’t be exclusively F either but wondering if the scale needs to be tipped a bit more.
— 67Sound (@67sound) August 12, 2014
The contention is that the opportunity cost is too high and goaltender evaluation and projection is too bad to make drafting goalies worthwhile. Doing so means that a team would forgo a forward they’ve (theoretically) got a better chance to turn into a real NHLer in favor of a goaltender they have a worse chance to get return on. Instead, the argument goes, teams can stock their goaltenders through free agency and trades.
I’ve done some work on goaltender drafting and, in recent years anyway, it’s pretty clear that most drafted goalies don’t really become NHL goalies. Among goalies drafted between 2000 and 2010, only 35% had played any NHL games at all by the end of the 2012-13 season. A few more have made the jump since then, mostly younger guys who are just breaking in, but not enough to move that needle significantly.
Let’s be generous and say that, given enough time for things to shake out, 40% of goaltenders drafted by NHL teams will have at least one NHL game. This is the rate for draft classes between 2000 and 2006, and it’s likely to end up in that general range for more recent years given enough time. It’s got to be hard for anyone to look at that and consider drafting goaltenders a high-value enterprise for any team.
But there are problems with the logic that leads from such evidence to ending or severely curtailing drafting of goaltenders.
Essentially what proponents of this strategy are arguing for is specialization. The assumption is that NHL teams are so much better at identifying and developing young forwards that they should concentrate on that and let someone else develop the goaltenders.
Those guys you want your team to sign in free agency or trade for have to be developed by someone. Someone has to find them at the amateur level, train them, coach them, get them into the right playing situations, and provide them with resources. That process is not foregone if a goalie isn’t drafted. In fact, it’s harder. Nor would teams be able to avoid scouting and evaluating goalies. They would simply be doing that at a different point, directing it at a different spot.
Specialization only makes sense if NHL teams are much, much better at identifying and developing forward talent than at identifying and developing goalie talent. However, they’re not. Drafting for forwards actually has similar rates of failure as drafting for goalies and the way we examine the data can exaggerate success rates for forwards.
Often those who have concluded that forward drafting is of much higher value than goalie drafting look at the data in a particular way that distorts the difference between them. Most of the time, they construct their argument by comparing how many “elite” players (variously defined) at each position were drafted in various rounds. The formula is X percent of elite players were drafted (or drafted before Y position), ergo drafting is more successful. This is missing some important context, however.
Far more forwards are drafted than either defensemen or goalies and there are far more roster spots for forwards than for defensemen or goalies. A larger population with a greater opportunity to play will naturally lead to a higher number of players being elite, even if there’s no difference at all in success rates.
Hypothetical:
Assume that 50% of each group make an NHL roster, 12% of each group is a “success” and 4% of each group are elite.
|
Forwards |
Defensemen |
Goalies |
|
|
Drafted |
500 |
250 |
125 |
|
Make Roster |
250 |
125 |
63 |
|
Success |
60 |
30 |
15 |
|
Elite |
20 |
10 |
5 |
|
Percent of successes/elite |
57% |
29% |
14% |
By looking at the data this way–looking only at the last line of the table–it appears that forwards are not only scouted and projected better, they’re scouted and projected profoundly better. It’s easy to go from this formulation to the assumption that forwards are successful far more frequently than their colleagues in other positions. But we’ve already stipulated that, in fact, the success rates are the same. There’s just a bigger population to choose from and more opportunities to play. In fact, because of how this example is set up, their success is happening in direct proportion to their representation in the draft.
To find the real value of forward drafting vs goalie drafting, you have to look at overall drafting. What percentage of draftees at each position meet some reasonable benchmark of success? For instance, number of NHL games played.
I did that for the draft years 2006-2008:
|
No drafted |
> 200 GP |
>100 GP |
>10 GP |
0 GP |
|
|
Forwards |
362 |
55 (15%) |
84 (23%) |
140 (39%) |
199 (55%) |
|
Defense |
204 |
25 (12%) |
43 (21%) |
71 (34%) |
110 (54%) |
|
Goalies |
69 |
2 (3%) |
6 (9%) |
17 (24%) |
42 (61%) |
So, basically, only about 15% of forwards drafted become even full-time NHL players in the 5 to 7 years after their draft (the bar here is 28 to 40 games played per year for 7 or 5 years, respectively). Three-quarters–77%–of drafted forwards don’t even qualify as marginal NHL players (100 GP in 5+ years). I have a hard time considering that a huge success rate, but it is better than defensemen by a little and goalies by a lot. More on this later.
That’s all happening at the top end. At the bottom end, the rates of draft failure are pretty similar between the three positions. In fact they’re proportional to each position’s representation in the whole population.
|
% of all drafted (n=635) |
% of > 200 GP (n=82) |
% of >100 GP (n=133) |
>10 GP (n=228) |
0 GP (n=351) |
|
|
Forwards |
57% |
67% |
63% |
61% |
57% |
|
Defense |
32% |
30% |
32% |
31% |
31% |
|
Goalies |
11% |
2% |
5% |
7% |
12% |
The biggest gains that forwards have over goalies is in getting above the 200 game mark. Goalies are getting into the NHL at only a slightly lower rate than their colleagues. They have a hard time racking up the games, though. When you see discrepancies of that nature, it’s time to ask why.
A big reason for this is that goalies simply don’t play as many games as forwards do. A forward can easily play 60 to 70, even 80, games a year, even in their first or second NHL season. Only a handful of goalies ever get up above 60 games year in and year out, and those are not guys in their first few years. NHL teams with entrenched starters and solid backups are going to give their young goalie prospects few games in their first years in the league. Even a backup goaltender gets only 20-25 games a year. Thus it takes goaltenders much longer to rack up games (or GVT or any other measure of goalie value.)
Think of it this way: 25 is still pretty young for an NHL goalie. Few goalies are going to have 200 or more games by the time they’re 25. In fact, only 14 goalies have done so in the past 14 years. More than 300 forwards have.
The point here is that how we measure success matters. It’s very easy for the benchmarks we use to skew our conclusions about overall drafting.
To what extent is the apparent difference in success rates for forwards and goaltenders an artifact of how the positions differ in practice and how we measure success? If drafting fails at the more or less the same rate for all positions, how much of that failure can be attributed to a real difference in our ability to scout or project? Looking at the data this way, it appears that there isn’t any substantial difference in scouting or projection, merely a difference in usage and numbers. This question certainly needs greater examination.
But it brings us to the second problem I have with the argument against drafting goalies. It’s logically untenable to assume that if a player doesn’t become an NHL success the mistake was made in initial evaluation and choice at the draft. The only way to determine if a team drafts well is to look at the end product of another process altogether: a players’ progress through amateur and minor hockey and into the NHL. It is literally impossible to analytically disentangle drafting from development.
The same is, of course, true of evaluating development programs. Is there a flaw in the development process or are the people tasked with development simply being given players who aren’t very good? It’s impossible to tell. We cannot unmarry the two processes analytically. And both have to work well.
There are reasons that NHL GMs will continue to draft goalies–the reward when you hit it right is very, very high, for one. For another, GMs are competing with each other for talent, and the draft is intended to enforce parity. It gives controlled access to the same pool of prospects and it prevents bidding wars from breaking out. Imagine if 30 teams could have bid on the services of a 23-year-old Henrik Lundqvist. While this is something that the players might like, that’s a no-go for ownership.
There are other reasons drafting won’t be going away any time soon. Drafted goalies are assets just like any other drafted (or undrafted) player, and it’s tough to argue that GMs should hand that advantage over to another team or institution. Not to mention the fact that if you acquire goalies in trade or free agency, you are still giving things up for them, including the opportunity cost of cashing in on their early years. GMs need extra value in the early years of a player’s career to make up for the cost of their later years.
Still, there is much room for improvement in goalie drafting and development. That can be done, however, without resorting to extreme measures like kind of specialization being proposed. Many of these are already being done quietly by NHL teams, in fact.
Improving evaluation by encouraging the innovation being done in integrating video and data-based analysis.
Make scouting increasingly comprehensive. At this moment, only somewhere around 16% of professional goalies who played their amateur hockey in Russia are drafted into the NHL. Compare that to about 60% of pro goalies who played amateur hockey in North America and 40% who played in Finland or Sweden. That’s an imbalance that is at least partly due to lack of exposure, though there are other factors at play as well. This is improving all the time.
Use position-specific scouting for goalies. Send goalie scouts, at some point in the process, to scout goalies. Again, this is something that almost all NHL teams most likely do to some extent already. It could be done earlier and cover more prospects.
Invest in goaltender development. This position is the most technically intensive position in hockey. You cannot develop teenage prospects into NHL goalies without consistent, competent, creative goalie coaches at the minor league level. Right now, many young goalies have to do this on their own during the offseason. Only about half of the league have hired full-time goalie coaches for their AHL affiliates, and many of those also cover other minor leagues as well.
Integrate drafting with your overall vision for goaltender development and practices. Draft the kind of goalies you have the resources, personnel, and philosophy to develop right. The process should be integrated from top to bottom, with a clear endgame in mind.
The blueprint here is the way that Finland and (more recently) Sweden deliberately focused on developing goaltending as a competitive advantage. They invested in coaching and they integrated and improved training at all levels of the sport. This is often done very poorly in North America at this point, in part because of the structure of North American hockey leagues, although this is slowly changing.
In the meantime, GMs who haven’t already should adjust drafting strategies to realities. I fully agree with those who argue against drafting goalies high. Only in exceptional cases (such as Andrei Vasilevskiy) does it make sense to draft a goalie before the late second round at the earliest. And don’t be hesitant to explore older goalies who have passed draft age.
However, the gap between goalie drafting and forward drafting isn’t nearly as stark as it’s been made out to be. It’s much more worthwhile to make drafting and development at all positions better than to attempt to specialize in elite forwards to the exclusion of other positions.
