FanPost

Number of Starters Not Necessarily an Indicator of Draft Success

As usual, OCC has presented us with some excellent, thought-provoking analysis – this time in his post about team by team draft success over the past three years. He states that it’s the goal of every team to pick immediate or eventual starters in the first three rounds – which I would agree is true. Then he presents a very logical premise that the better a team is at drafting, the more starters they’ll have (in the first 3 rounds).

My first reaction to this premise is to immediately agree. And when I take a look at the data presented in the post, it sure seems to support most people’s opinion that the Cowboys have drafted well the past three years.

But then I noticed the other teams that also scored high (using OCC’s draft success rate). Most of them just aren’t very good teams. And the thought occurred to me; if a team isn’t very good, then they probably have more holes to fill – therefore likely to start more of the players they draft.

That seems logical, too. The lower quality of players you have on a team (i.e., the worse off your team is), then your bar will be set lower than a higher performing team. For example, a drafted player that starts for Cleveland or Jacksonville might not necessarily start for New England. Conversely, a backup on New England’s team might be a starter on half a dozen other teams.

To test this theory, I compared OCC’s list of teams and their associated "success rates" with another measure of success; their win percentage over the past three years (assumes that a higher winning percentage correlates to a higher bar or less needs for a team).

Team

Picks in Rds 1-3

Primary Starters

Success rate

Win

Carolina

8

7

88%

31%

Seattle

8

7

88%

52%

Dallas

7

6

86%

46%

Cleveland

12

10

83%

29%

Tampa Bay

10

8

80%

44%

Buffalo

9

7

78%

33%

Kansas City

12

9

75%

40%

Denver

12

9

75%

52%

Minnesota

7

5

71%

40%

Atlanta

7

5

71%

75%

Miami

9

6

67%

42%

Tennessee

9

6

67%

44%

Washington

6

4

67%

44%

Oakland

7

4

57%

42%

Chicago

7

4

57%

60%

New England

14

8

57%

81%

Detroit

9

5

56%

42%

Baltimore

9

5

56%

71%

St. Louis

11

6

55%

34%

Arizona

8

4

50%

38%

Philadelphia

10

5

50%

46%

Indianapolis

10

5

50%

48%

San Francisco

9

4

44%

64%

Jacksonville

7

3

43%

31%

New York Jets

7

3

43%

52%

New Orleans

8

3

38%

65%

Cincinnati

12

4

33%

48%

Houston

9

3

33%

58%

New York Giants

9

3

33%

58%

Green Bay

9

3

33%

75%

San Diego

10

3

30%

50%

Pittsburgh

9

2

22%

67%

You can see from the chart that only 3 of the top 10 teams (ranked by their draft success rate) managed a winning record over the past three years, averaging a 44% win rate. Meanwhile, the bottom 10 included 8 teams with winning records, averaging 57%.

This seems to support my theory that the percentage of draft picks that end up being a starter is more a factor of the team’s existing quality of players (or lack thereof) - as measured by the quality of the teams record.

I have not done the analysis on this, but there could be another factor in play. Since teams drafting fewer starters also have better records, they likely are picking lower in the draft. So over three rounds, a team like Cleveland might average 20 or more spots ahead of someone like New England. That has bound to translate into higher quality players – and possibly more starters.

Again (in theory), it seems that a team’s record has more to do with how many starters they draft in the first three rounds – and not just that they are better at drafting.

Ok, so just how DO you measure drafting success? I’d say the ultimate goal of a drafted player is to help their team win. Yes, starting for the team is one way to do that, but not all starters are equal - and that view discounts the contributions of backups, which in many cases can be significant towards helping their team win games.

I probably don’t need to ask this, but does anybody else have any other ideas?

Another user-created commentary provided by a BTB reader.