January 24, 2011
Up and Coming
The Projectinator Revised, and the Effect of Birth Month

by Iain Fyffe

The Projectinator is something like a year-and-a-half old, and the old girl was due for an overhaul. Before expanding the system to produce projections for players beyond the CHL and USHL, we should make sure the system has a solid basis of comparison. The old 10Y score, which was a rough estimate of a player's value over the 10 seasons after his draft year, is abstract and useful only for ranking players. Much better would be a number that actually tells you something about the player.

As such, the Projectinator has been revised to produce a Goal Versus Threshold (GVT) estimate, which gives you an indication of the player's actual future value to an NHL team. The Projected GVT (PGVT) is intended to reflect the GVT a player will accumulate in the best five of ten seasons after his draft year, on a per-82-games basis. We keep the 10-year horizon, since after that time, a player will typically be an unrestricted free agent, and we use the five best years to evaluate a player based on his peak value.

Determining the peak GVT values for the players in our data set allowed for much tweaking in the Projectinator estimates. For example, the system no longer considers a high number of penalty minutes to be a negative for forwards, and a bit more focus is given to the effect of a player's team on his numbers. In June of 2009, we ran the Projectinator numbers for OHL forwards who were in their age-17 seasons between 1989 and 1998. The top 30 of those players are presented again here, but this time not only with their old fashioned 10Y and 10YE scores, but their brand new GVT and PGVT numbers for comparison.

Age-17 Forwards from 1989-1998, Projected Goals Versus Threshold

Rank	Name			10Y	10YE	GVT	PGVT
1	Lindros, Eric		1.11	1.00	30.3	28.7
2	Thornton, Joe		1.14	.99	21.8	19.7
3	Gratton, Chris		0.75	0.79	8.0	8.7
4	Savard, Marc		0.76	0.76	10.3	9.6
5	Ricci, Mike		0.74	0.71	9.3	9.1
6	Legwand, David		0.81	0.68	9.1	9.8
7	Primeau, Keith		0.78	0.66	15.2	9.0
8	Roche, David		0.35	0.66	2.8	7.2
9	Nolan, Owen		0.88	0.63	17.0	8.4
10	Allison, Jason		0.87	0.63	18.9	8.5
11	Cleary, Daniel		0.59	0.63	5.4	9.1
12	Tkaczuk, Daniel		0.42	0.62	3.8	7.5
13	O'Neill, Jeff		0.77	0.62	10.4	9.2
14	Wren, Bob		0.53	0.62	5.9	5.3
15	Stillman, Cory		0.68	0.61	9.6	6.5
16	Arnott, Jason		0.89	0.58	13.4	5.6
17	Kilger, Chad		0.49	0.58	4.1	5.1
18	Bowler, Bill		0.52	0.58	4.4	4.8
19	Bonsignore, Jason	0.38	0.57	-0.7	2.6
20	Moreau, Ethan		0.49	0.55	5.0	5.4
21	Warriner, Todd		0.46	0.55	4.5	6.7
22	Bell, Mark		0.57	0.55	5.4	5.1
23	Sim, Jonathan		0.49	0.53	6.3	5.1
24	Harvey, Todd		0.54	0.51	7.6	7.7
25	Taylor, Chris		0.46	0.49	5.9	5.2
26	Simon, Chris		0.56	0.49	7.8	4.5
27	Young, Jason		0.43	0.48	0.2	4.5
28	Convery, Brandon	0.49	0.48	4.4	5.6
29	Cheechoo, Jonathan	0.71	0.47	13.3	6.0
30	Seguin, Brett		0.41	0.47	N/A	6.1

For the most part, a player's ranking does not change much from the old system to the new. However, there are some notable exceptions. Players like David Roche and Daniel Tkaczuk, the highest-ranked “misses” under the old system, are still highly-ranked but have been overtaken by superior players like Owen Nolan, Jason Allison, Jeff O'Neill and Todd Harvey. Some high-scoring minor leaguers like Bill Bowler and Bob Wren are also knocked down the list. Famous flop Jason Bonsignore, who was drafted fourth overall but produced a negative GVT value for his career, is now deftly avoided by the Projectinator.

Again, just looking at a few individual players is not a fair way of evaluating the system. Though Bonsignore is better projected by the system than the scouts, there are other players who go the other way (though the system gets it right more often). After expanding the new system to account for NCAA players and Europeans, we'll be able to do a year-by-year review of draft results to see just how powerful the Projectinator is.

You might have noticed that if we simply use a player's expected GVT over 82 games, the top-ranked prospects will always be goaltenders. Goalies play 60 minutes per game, so of course they will produce the greatest amount of GVT per game. When ranking players for the draft, we therefore make a minute (not minute) adjustment. Since top forwards play about 20 minutes per game, a goaltender's PGVT is multiplied by one-third (20 divided by 60) to produce his draft ranking. Similarly, since top defensemen generally play about 25 minutes per game, their PGVT numbers are multiplied by four-fifths (20 divided by 25) to arrive at their draft ranking.

The Effect of Birth Month

A common question is to what degree a player's birth month should be considered when evaluating a draft prospect. It's well known that a player born in the last three months of the year has a distinct advantage over his draft classmates, due to the Entry Draft cutoff date of September 15. A first-time draft-eligible player born in October, November or December has essentially had an additional year of development over one born in March, for example, since minor hockey generally uses birth year only to group players into divisions.

But how much of a difference does it really make? Quite a difference, actually.

To illustrate the dramatic effect that a player's birth month can have on his projection, let's look at a hypothetical forward from the OHL, a prospect for the 2011 Entry Draft. He plays for an average OHL team, is of average size (6'1") and recorded the following statistics: 40 goals, 60 assists and 60 penalty minutes in 60 games. The Projectinator sees this player being worth an average of 8.1 GVT in his prime years, making him a very good prospect; however the projection ranges from 6.0 GVT to 9.9 GVT, depending on when he was born.

The Effect of Birth Month on Projected GVT

Birth Month		Projected GVT
September 1992		6.0
October 1992		6.3
November 1992		6.6
December 1992		7.0
January 1993		7.7
February 1993		8.0
March 1993		8.3
April 1993		8.7
May 1993		8.8
June 1993		8.9
July 1993		9.1
August 1993		9.5
September 1993		9.9

If there's a factor that can cause a swing of as much as 25% in a player's projected value, you'd better be sure to consider that factor. Birth month is a terribly important thing to consider when you're drafting a player. It can make the difference between a great prospect and a merely very good one.