When looking at draft prospects, it's rare that a scouting report will ever have a projection that describes a player's exact future. Usually a report will consist of something along the lines of:
"Likely projects as an average top-six winger. If the skating takes a jump, he could be a below-average first liner, but if the physical game does not translate to the pros well, he could end up a journeyman."
However, when it comes down to ranking prospects and choosing prospect A over prospect B, if they both have very different projections how can you compare them? Well, the first step is to find out what the value is of those different projections. To find the value of players at different production points, I took GVT and Goals Versus Salary data from 2009-10 for all players under the age of 28. I chose under 28 as all players become UFA's after their age-27 season and get paid market price which makes them lose most of their value unless their production point is scarce. I also filtered out all players who came into the league young and got their big contract after playing seven years or in the rare case of an elite player simply skipping his second contract. The GVS data for forwards and defenders were different enough that I felt it was warranted to do two different graphs which was likely caused by the fact young defenders tend to produce less than forwards in their early 20's. The games played limit is 30.
Here is for forwards:
And for defense:
Then you take the GVT and put it next to its respective counterpart on the fitted GVS line:
# GVT GVS GVT GVS
1 24.6 18.0 22.4 13.7
2 22.0 15.4 18.9 11.4
3 21.3 13.9 14.6 10.1
4 21.2 12.8 12.0 9.2
5 19.1 12.0 11.8 8.4
6 17.1 11.3 11.4 7.8
7 15.3 10.7 11.1 7.3
8 15.3 10.2 10.6 6.9
9 15.0 9.8 10.4 6.5
10 14.3 9.4 9.8 6.2
11 14.3 9.0 9.2 5.9
12 13.0 8.7 9.2 5.6
13 12.7 8.4 9.1 5.3
14 12.4 8.1 8.9 5.1
15 12.0 7.9 8.9 4.8
20 9.8 6.8 7.6 3.9
25 9.2 6.0 6.4 3.2
30 8.2 5.3 6.0 2.6
35 7.8 4.7 5.4 2.1
40 7.3 4.2 4.8 1.6
45 6.8 3.8 4.5 1.3
50 6.2 3.4 3.7 0.9
60 5.4 2.7
70 4.9 2.1
80 4.1 1.6
90 3.4 1.2
100 2.6 0.8
And now we have a snapshot of the market at each position at each GVT level and its respective GVS value. It's from this point on however that we unfortunately lose the scientific aspect due to scouting reasons and have to start creating arbitrary numbers. Our test subjects will be some of the top forwards in the 2011 Draft.
With no real consensus at the top of the draft, let's try comparing Sean Couturier and Gabriel Landeskog. As per Iain Fyffe's projections, Sean Couturier has a projected Peak GVT of 11.6 and Landeskog has a projected Peak GVT of 7.6. However, based on how scouts are looking at both of them and projecting their output and likeliness of reaching that output could alter their value based on certain variables.
Sean Couturier for example gets a lot of knocks for his poor skating ability, yet he does have the skill set to be a first line forward if he can get that tool to be a bit better or find ways to manage when he gets to the pro level. With that in mind, let's say a team's scouting department can make an educated guess that Couturier had a 45% chance of hitting his projected peak (11.6 GVT), a 50% chance of being a second liner (6.0 GVT) and a 5% chance of being a bottom-six player (2.5 GVT).
Gabriel Landeskog has gotten high praise for his intangibles and is about as sure a thing as there is in the draft to be an NHLer, but with questionable upside. So let's say the scouting department makes an educated guess that Landeskog has a 70% chance of hitting his projected peak (7.6 GVT), a 20% chance of being a bottom six player (2.5 GVT) and a 5% chance of exceeding his projected peak value by a half-win (10.5 GVT).
There are obvious flaws with this method, one being that the percentages are completely arbitrary with no evidence other than scouting knowledge and another being that Iain's Peak value is the peak as per the player's projected career as opposed to an absolute upside. For the former, scouting essentially is a subjective field so this is nothing new. On the latter, that was just a way to explain the example; in actual practice, the player's true upside could very likely be different than his projected Peak GVT.
Staying with the example, let's say we had some sort of trust in the projections and likelihood of each projection; we could weigh the projections based on the GVS at each production point and get a value for that prospect. So for the two players:
0.45 * 7.5 + 0.50 * 3.35 + 0.05 * 0.78 = 5.09 GVS
0.70 * 4.65 + 0.20 * 0.78 + 0.05 * 7.35 = 3.78 GVS
This method isn't meant to be scientific, but rather to help analyze upside and projections from a scouting perspective. Different people may get different projections on players and thereby get different numbers at the end. Too many times there are debates over ceiling versus likelihood of being an NHLer when it comes to draft prospects and by simply putting a number, even a fully arbitrary one on a scouting report, it could help at least somewhat support who is right or wrong.
Corey Pronman is an author of Hockey Prospectus.
You can contact Corey by clicking here or click here to see Corey's other articles.