The 2016 NHL Draft Study

What is the value of any given NHL draft pick? Typically we look at games played to determine the quality of a draft pick, and for forwards perhaps also Points. The drawback is that simple games played does not fully explain the value of a player. The best measure, is salary. The amount teams are willing to pay these prospects is perhaps the best indicator of their true value, and allows us to compare players from different positions on the same level. Yes it has its flaws, with teams making mistakes with contract negotiations, and players can lock in long term at a team friendly price (like a Kyle Turris), however despite those flaws it is still the best measure.


ABOUT THE DATA

The 2016 NHL Draft Study is basically spreadsheet of nearly every NHL draft pick from 2004 - 2013 (goalies and players older than 21 on draft day not included), almost 2000 players total. The data includes production, games played at every level/age, and salary earned between age 18 and 25. The entire sample has only played in the NHL under a salary cap. Statistics from the 2015/16 season have yet to be added.

The salary information was from the cap hit listed at the time on NHL Numbers, and the playing statistics from Elite prospects. I still haven't emotionally recovered from the loss of Cap Geek, but there are inferior alternatives. Cap Friendly is a superior site to NHL Numbers, however Cap Friendly only lists the historical cap hit for a player as if he played the whole season. NHL Numbers actually calculates the real cap hit for players who don’t play an entire NHL season (which varies for players going up and down from the minors on 2 way contracts).

Age: In this study, age refers to how old the player is on Sept 15th of any given season, since that's the draft cut off. It means that the "18 year old season" is the season after the first year of draft eligibility regardless of what month the player was born.


Statistical Career Milestone Probabilities

Below are 3 charts to determine the probability that a player reaches any given career milestone for NHL GP, NHL PTS, and AHL GP. The chart at the bottom is for Expected GP and PTS.

Important Notes:

*It is intended as a cumulative probability, but because of the nature of this time-series analysis, doesn't always behave that way. The data from more recent draft picks are not included in the results for older ages. Sometimes players with at least 1 GP at age 24 aren't included at age 25, so you'll see a number that should only get larger actually get smaller (typically that's only a small decline for larger sample sizes).

Conclusions:

*57% of NHL draft picks won't have played a single game in the NHL at age 25.

*10.2% of NHL draft picks will score 100 NHL points before age 26.

*73% of NHL draft picks will play at least 1 AHL game before age 26.

* The average NHL draft pick will have played 77 NHL games and 109 AHL games before age 26.

*Expected AHL games played starts to shrink at age 22, Expected NHL games at age 23. While some players may still be developing, many busts start to wash out of the system at those ages.

	PROBABILITY OF NHL GP MILESTONES
AGE	0	1	10	100	300
18	97.3%	2.7%	1.9%	x	x
19	90.7%	9.3%	6.5%	1.7%	x
20	80.7%	19.3%	13.5%	4.0%	x
21	70.1%	29.9%	22.3%	7.1%	0.3%
22	61.8%	38.2%	28.9%	12.6%	1.7%
23	57.5%	42.5%	34.1%	17.8%	3.8%
24	56.4%	43.6%	35.8%	20.9%	6.7%
25	57.1%	42.9%	36.2%	22.7%	10.2%
	PROBABILITY OF NHL PTS MILESTONES
AGE	0	1	10	100	300
18	97.6%	2.4%	1.7%	0.1%	x
19	92.4%	7.6%	4.6%	0.4%	x
20	84.7%	15.3%	8.8%	1.2%	x
21	76.2%	23.8%	14.4%	2.2%	0.2%
22	68.4%	31.6%	19.2%	4.6%	0.5%
23	63.4%	36.6%	24.3%	7.3%	0.9%
24	61.4%	38.6%	25.3%	8.9%	1.0%
25	60.9%	39.1%	27.2%	10.2%	1.6%
	PROBABILITY OF AHL GP MILESTONES
AGE	0	1	10	100	300
18	93.2%	6.8%	1.0%	0.0%	x
19	79.7%	20.3%	9.7%	0.2%	x
20	53.0%	47.0%	40.1%	4.2%	x
21	41.1%	58.9%	51.8%	22.1%	x
22	32.8%	67.2%	60.4%	32.3%	0.0%
23	26.7%	73.3%	67.2%	38.5%	1.0%
24	26.2%	73.8%	68.6%	42.1%	3.9%
25	27.2%	72.8%	68.4%	43.9%	6.4%

	STATS/PICKS
	NHL	NHL	AHL	AHL	Euro Elite
AGE	GP	PTS	GP	PTS	GP
18	1.3	0.7	0.5	0.2	3.6
19	3.6	1.5	4.4	2.0	4.6
20	6.2	2.6	19.9	8.5	4.9
21	9.8	4.3	22.7	9.8	4.4
22	12.7	5.4	21.8	9.4	4.9
23	15.0	6.1	17.9	7.4	6.2
24	14.4	5.9	13.1	5.3	8.2
25	14.1	6.0	8.3	3.3	9.5
Sum	77.1	32.5	108.6	45.9	46.2

Salary and Expected Value

The Salary and Expected Value charts below provide some interesting information. The 18 year season is not included in the expected value calculation because only a tiny proportion of the population of draft picks are ready to jump immediately to the NHL, and there is very little variation in the salaries of those that do.

 Then there is the ability to determine what proportion of the sample is earning what salary at any given age.  $0 players are those who play zero regular season games in the AHL or NHL. It's important to note that these are not career earnings below, but rather for that specific age.

Salary: Refers to NHL salary cap hit as listed at NHL Numbers.com, plus up to $50,000 per AHL season. I am not yet aware of a source online for complete and reliable find AHL contract data, so I just pay everyone the same for an AHL game ($50,000/80).

Expected Value: This is the amount a player earned in salary minus the expected salary of a player drafted in that position, at each age. Players returned to junior or playing college after being drafted will have an expected value of zero and are not included in calculating the average salary. While players are not punished directly for returning to amateur hockey, it can have an indirect negative effect for high picks that take slower to get to their 2nd contract.

 Also, there are some false positives and false negatives that can occur in instances where a team gets a player on a favourable contract (like a Victor Headman) or visa versa for teams that pay a player too much (like Rostislav Olesz). I am using salary to measure a player's value, which has some flaws, but might be the single best variable to determining player value.

Important Notes:

 *Salary/Pro: This variable is calculated by total amount of salary divided by the number of players who have turned pro (AHL or NHL). If a player is returned to junior or college but plays a few games at the beginning or end of the season, that does not count for "turned pro" and the Expected Value is zero. Pro Salary earned during amateur seasons is not used in the Expected Value forecast, but still counts as Salary.

 *Note that in the sum at the bottom of SALARY INFO chart, the salary from age 18 is not included.

 Conclusions:

* Only 9% of all draft picks will earn pro money as 18 year olds, and only 1.7% will earn over a million dollars.

*On average, NHL draft picks will earn $3.2M from age 19 to 25.

*3.8% of NHL draft picks will earn over $4M at age 25.

*At age 26, NHL teams will pay draft picks an average of $37,104 per NHL game played.

*The lowest average amount NHL teams will spend for each NHL point occurs at age 21.

	SALARY INFO
		% of players earning this salary range
AGE	Salary/Pro	$0	< $1M	> $ 1M	> $4M
18	$1,644,717	91.0%	7.4%	1.7%	0.0%
19	$790,128	78.4%	18.0%	3.6%	0.0%
20	$226,862	51.2%	44.7%	4.1%	0.0%
21	$320,391	40.7%	53.8%	5.5%	0.9%
22	$393,609	36.1%	56.9%	7.0%	1.8%
23	$473,669	39.6%	49.2%	11.2%	2.8%
24	$488,906	49.4%	37.6%	13.0%	3.3%
25	$522,247	60.4%	25.3%	14.3%	3.8%
sum	$3,215,812
	SALARY AND EXPECTED VALUE
AGE	Salary-E[V]	E[V]	Sum(E{V])	Salary/GP	Salary/PTS
19	$26,008	$764,120	$764,120	$28,984	$67,128
20	$480	$226,382	$990,502	$24,279	$57,428
21	$478	$319,914	$1,310,416	$23,969	$54,935
22	$564	$393,045	$1,703,460	$25,971	$61,279
23	$44	$473,625	$2,177,085	$30,144	$73,818
24	-$0	$488,906	$2,665,992	$33,735	$82,131
25	$0	$522,247	$3,188,238	$37,104	$87,369
sum	$27,574

Development Matrix

There are many different paths to take to get to the NHL. It can be fascinating to explore these different paths for different samples of prospects. If you would like to see how a sample of draft picks advances through different leagues and skill levels, welcome to the Development Matrix (or more correctly, matrices). If a draft pick played at least 1 game at any level last year, where is he likely to go next year? My spreadsheet contains more information than you'll find below, which was compressed for posting purposes.

Important Notes:

* KHL = Russian Elite, EurMA  = Euro Mainland Elite (Czeck, Slovakia, Germany, Swiss), ScanA = Scandinavian Elite (Sweden and Finland), EurB = All of the 2nd tier European leagues, EurJr = All European junior leagues,  CHL = Canadian Major Junior, WJC = Under 20 World Junior Championship, Tier II = All North American amateur that doesn't ruin NCAA eligibility, USDP = United States Development Program

* Age 17 is the first year of NHL draft eligibility.

* Charts meant to be read left to right, not up and down.

* Do not sum the numbers in any direction. It is what percentage of the sample played at least 1 game at any level and many play at more than one level. They are not intended to sum to 100%. You can't sum the columns because the denominators are different.

Conclusions:

* Nearly 7% of draft picks played for the USDP in their draft year. Of those players, 82% go on to play in the NCAA, 11% to Canadian Major Junior.

%  of	Played Here		Then played here at	AGE 18
Sample	at Age 17	CHL	NCAA	AHL	NHL
2.0%	KHL	10.0%	0.0%	10.0%	5.0%
3.0%	EurMA	54.2%	0.0%	10.2%	0.0%
5.1%	ScanA	8.9%	0.0%	5.9%	5.9%
9.2%	EurB	19.3%	0.6%	4.4%	1.7%
1.9%	NCAA	10.8%	86.5%	5.4%	2.7%
44.6%	CHL	96.0%	0.0%	12.4%	4.8%
6.8%	WJC	32.3%	9.0%	15.8%	24.1%
16.0%	Eurjr	23.1%	0.0%	2.8%	0.9%
24.7%	Tier II	9.4%	40.7%	0.6%	0.2%
6.6%	USDP	11.5%	82.4%	0.8%	0.8%
%  of	Played Here		Then played here at	AGE 19
Sample	at Age 18	CHL	NCAA	AHL	NHL
2.7%	NHL	13.0%	0.0%	29.6%	94.4%
6.8%	AHL	66.9%	0.0%	52.6%	21.1%
3.2%	KHL	3.2%	0.0%	12.7%	9.5%
1.5%	EurMA	26.7%	0.0%	6.7%	10.0%
7.7%	ScanA	3.9%	0.0%	13.8%	9.9%
8.8%	EurB	10.4%	0.6%	11.6%	5.2%
16.3%	NCAA	7.8%	88.8%	11.5%	5.0%
49.3%	CHL	83.5%	0.0%	25.6%	10.3%
19.2%	WJC	27.5%	13.0%	29.1%	24.3%
10.5%	Eurjr	12.1%	1.4%	8.7%	2.9%
15.1%	Tier II	7.0%	71.1%	1.3%	0.0%
%  of	Played Here		Then played here at	AGE 20
Sample	at Age 19	CHL	NCAA	AHL	NHL
9.1%	NHL	0.0%	0.0%	58.9%	89.9%
17.1%	AHL	4.7%	0.0%	85.7%	41.5%
1.2%	ECHL	8.7%	0.0%	69.6%	4.3%
3.4%	KHL	0.0%	0.0%	19.0%	14.3%
1.6%	EurMA	0.0%	0.0%	21.4%	7.1%
7.6%	ScanA	0.0%	0.7%	36.2%	15.2%
6.7%	EurB	0.0%	0.0%	16.1%	6.5%
25.8%	NCAA	2.7%	86.9%	14.8%	5.9%
44.4%	CHL	26.3%	0.0%	64.4%	17.6%
23.6%	WJC	2.8%	16.1%	52.2%	31.5%
5.8%	Eurjr	0.0%	1.9%	18.9%	5.7%
4.3%	Tier II	2.4%	75.6%	2.4%	0.0%
%  of	Played Here		Then played here at	Age 21
Sample	at Age 20	NCAA	ECHL	AHL	NHL
18.4%	NHL	0.0%	0.7%	64.1%	89.2%
43.8%	AHL	0.0%	16.2%	88.1%	47.6%
11.0%	ECHL	0.0%	47.7%	74.4%	6.8%
3.1%	KHL	0.0%	6.1%	16.3%	8.2%
2.7%	EurMA	0.0%	4.3%	8.5%	2.1%
6.3%	ScanA	0.0%	6.7%	33.7%	13.5%
6.4%	EurB	0.0%	5.8%	9.6%	1.0%
26.1%	NCAA	81.2%	4.7%	28.1%	8.4%
12.1%	CHL	0.0%	41.9%	50.2%	3.9%
2.2%	Eurjr	0.0%	10.5%	26.3%	2.6%
1.2%	Tier II	42.9%	19.0%	9.5%	0.0%
%  of	Played Here		Then played here at	Age 22
Sample	at Age 21	NCAA	ECHL	AHL	NHL
27.5%	NHL	0.0%	0.8%	52.8%	85.5%
51.7%	AHL	0.0%	17.8%	78.9%	46.9%
13.8%	ECHL	0.0%	54.9%	67.6%	5.9%
3.1%	KHL	0.0%	2.0%	18.4%	18.4%
3.0%	EurMA	0.0%	0.0%	11.1%	8.9%
4.9%	ScanA	0.0%	1.3%	14.5%	11.8%
5.7%	EurB	0.0%	0.0%	4.7%	2.3%
22.1%	NCAA	55.4%	18.8%	45.8%	12.6%
%  of	Played Here		Then played here at	Age 23
Sample	at Age 22	NCAA	ECHL	AHL	NHL
33.1%	NHL	0.0%	0.2%	41.9%	85.0%
51.1%	AHL	0.0%	18.8%	68.7%	40.7%
15.0%	ECHL	0.0%	56.3%	47.4%	3.2%
3.3%	KHL	0.0%	0.0%	15.9%	9.1%
3.7%	EurMA	0.0%	4.3%	4.3%	10.6%
5.2%	ScanA	0.0%	0.0%	11.8%	8.8%
6.2%	EurB	0.0%	3.7%	2.4%	3.7%
12.6%	NCAA	24.2%	37.3%	59.6%	8.1%
%  of	Played Here		Then played here at	Age 24
Sample	at Age 23	KHL	ECHL	AHL	NHL
34.2%	NHL	4.0%	1.7%	33.9%	83.9%
42.7%	AHL	4.1%	20.7%	62.4%	33.9%
16.2%	ECHL	0.0%	50.3%	39.7%	0.6%
5.0%	KHL	78.9%	0.0%	5.3%	7.0%
4.6%	EurMA	1.8%	0.0%	7.1%	12.5%
5.8%	ScanA	7.7%	1.5%	4.6%	7.7%
8.0%	EurB	2.1%	11.7%	5.3%	3.2%
3.1%	NCAA	0.0%	36.1%	52.8%	8.3%
%  of	Played Here		Then played here at	Age 25
Sample	at Age 24	KHL	ECHL	AHL	NHL
30.8%	NHL	3.7%	1.8%	25.6%	83.9%
30.7%	AHL	5.1%	13.8%	52.0%	28.4%
14.1%	ECHL	0.9%	45.6%	23.7%	0.0%
6.7%	KHL	75.8%	3.2%	3.2%	8.1%
6.7%	EurMA	12.3%	1.5%	4.6%	7.7%