It's all about Value

By Chad Reuter
Senior Analyst
NFLDraftScout.com

April 13, 2008 - "Value" will undoubtedly be of the most often used words during media coverage the NFL Draft April 26^th and 27^th at New York's Radio City Music Hall. But what does that word really mean or represent? Turns out the term has multiple meanings, all of which are important in the draft evaluation process.

Value, not rounds

Most people rate players as first round picks, second round picks, etc. But when NFL teams talk about prospects with "first round value", they are not necessarily referring to the first 32 players taken (31 this year, thanks to the Patriots and "Spygate"). Teams consider 15-25 prospects, depending on the class, as the elite players in the draft truly worth a first round selection.

And, of course, there are other "tiers" of value within that group; in a typical draft, there are 5-7 players at the top, another group or 6-8 prospects rated just below them, and another 10-15 prospects at the next level. A team's breakdown of the top players of the 2008 draft may look something like this:

2008 Prospects with "First Round Value":

This "value tier" concept certainly makes sense considering the arbitrary nature of the number of picks in each round. The number of teams in the league has no correlation with the value of players in a particular draft class. If there were half the number of NFL franchises, would that mean there were 16 top-notch players making up the first round instead of 32? Of course not…but that's how the draft is evaluated by many outside of the league. All first round picks are supposed to become Pro Bowl-caliber players, in the casual fan's mind, because it they are the first selection for their team (barring trades), the highest paid, and the most hyped.

But this is certainly not the case…all first, second, third, etc. rounds are not created equal. The table below displays the differences by breaking down the success of draft picks by ten-pick groups, thereby eliminating the round consideration and focusing on where exactly the players were selected. The three bars per group show the average games played (blue), starts (red), and Pro Bowl appearances (green) per season. Only the first five years of a player's career are analyzed to make easier comparisons across selection groups.

Pro Bowl average figures multiplied by ten for graph visibility. Actual figure for top ten picks is 0.192 appearances per season, or one appearance every five years. The number of Pro Bowl appearances for most late-round picks is so small that they barely register on the graph, even with the adjustment.

Although there is limited information about the quality of players here and the Pro Bowl is often referred to as a popularity contest, it is clear that draft rounds have no real correlation to the value of players available to a team when they select.

As was mentioned earlier, there are definite reductions in value throughout the first round. Top ten picks from the 1994-2003 drafts averaged 11.8 starts over their first five seasons while late first round picks averaged 8.6 starts, a 25% decrease. The elite players were also selected to three times as many Pro Bowls as those selected from picks 21 to 30 (96 to 31). Similar reductions in those figures occur from early to late second round (8.1 to 6.4 starts, 26 to 13 Pro Bowls) and third round (6.1 to 4.2, 19 to 5).

Discussions with NFL personnel men, as well as my own previous research, points to breaks in player value around picks 10, 20, 50, 80, 120, 160, along with a very slight dip at 200. That means there are at least three tiers of value within the first round, a break in the middle of the second, third, and fourth rounds, and again in the late fifth and sixth rounds. Each draft is different in terms of where these value tiers lie, but generally there are changes in player value within a few picks on either side of those points. And the point in the round in which prospects fall has little to do with their appropriate value.

The beginning and ending picks in rounds three through seven depends on the number of free agent compensation picks awarded by the league during the late March owner's meetings; this is yet another example of arbitrariness of draft round delineation.

As would be expected, the difference in value between early tiers and later tiers lessens as the draft progresses. It is very difficult to get a starter once past pick #120, or the mid-to-late part of the fourth round. Even third round selections are not likely to become regular starters, as they average around five starts a year in their first five seasons. That can be loosely translated to mean that about one-third of these picks will become regular starters. The decline in playing time for late third round picks in the round shows they have less of a chance to succeed than early third round picks, where the percentage of potential regular starters increases to about 40%.

Note that the value of mid-to-late third round picks is closer to fourth round picks than second round picks. So the league's decisions to hold only the first two rounds on Saturday not only shortened the more hectic first day, but should also help put more realistic expectations on players selected outside the top 70.

What a team likes

Each team's evaluation of players, and therefore their value tiers, differ based on their evaluation criteria. Teams' draft boards look much more different than one would think. Even the three tiers of first round grade players listed above differ, especially in the 14-24 category, because some teams "value" certain types of players more than others do.

Teams' schemes have great influence over how draft prospects are ranked. Those using 3-4 defensive schemes value players considered "'tweeners" by teams using 4-3 base alignments. That means smaller or leaner college defensive ends with good all-around agility and larger defensive ends who eat up blockers are likely to be more valued by teams relying primarily on the 3-4.

For example, larger ends like Clemson's Phillip Merling and Miami's Calais Campbell may receive a first round grade by 3-4 teams, but USC's Lawrence Jackson garners more consideration in the late first round by 4-3 teams because he has decent size (6'4", 270) but is more explosive. Merling and Jackson also not considered 3-4 OLB prospects because they are not as fluid in space as top 10 prospects Vernon Gholston, Derrick Harvey, and Chris Long…who could all transition from college end to pro linebacker.

Team needs have an underrated part in determining a team's draft board. Those looking for quarterback help may place Matt Ryan in the top 10, as is done above. And they may place QBs Brian Brohm and/or Joe Flacco in the third tier, as well. But some teams not in desperate need of a signal-caller ranked all three as second-round value.

Most personnel men like to think they take the best player available by picking the highest-ranked player off their board when their turn comes around. But the fact is that throughout the evaluation process, scouts and administrators are aware of a team's weaknesses and take them into account when putting the board together…whether consciously or subconsciously. It's hard to imagine a team would not study players at need positions more closely than prospects where they have veteran depth.

The difference in player evaluation among teams negates the idea of a consensus "best player available". Very rarely is there one player truly better than all of the rest still on the board at a particular pick, although it can happen around the value tier breaks outlined above. But throughout the first round, three or four players could be picked at every selection with equivalent chances of becoming great players. In the later rounds, there are dozens of players of equal value waiting to be selected at every slot.

So I guess if can be said that the attributes a team values has a lot to do with the value assigned to each player.

Right place, right time

Draft analysts for ESPN, NFL Network, and all other media outlets use "value" to compare consensus player grades against the point at which the player is selected. Talking heads may say a pick was of "great value" because they had the player rated as a first round pick, but the team found him waiting to be picked in the second round.

On the other hand, a player picked in or within a round earlier than expected may be considered poor value, or a "reach". Sometimes teams do go down their board to find a player who fits an immediate roster need, but most times their high evaluation of a player is due to factors or criteria not recognized by the general public. A player considered a "steal" may have off-field issues or received less-than-glowing reviews from his college coaches to which people outside of teams are not privy. And other times, the consensus player value is simply wrong.

Remembering how value tiers work also helps understand whether a player is truly a reach or of good value. A player picked in the beginning of the second round may be considered a "steal", while some picked at the end of the first round are reaches. But as the data above shows, the likelihood for success of both players is about the same. Both players were probably selected where they should have been, but the first rounder happened to meet a need for teams at the end of the round, where the second rounder did not. The same applies for late second and early third round players, and late third to fourth rounders.

It is non-sensical to call a fifth, sixth, or seventh round player a reach, but still possible to consider them fair value, because the historical likelihood of success for players in that part of the draft is minimal. Certainly there are exceptions to the rule, but the late-round finds like Pats QB Tom Brady are few and far between.

Looking at recent draft pick production, there is a slight bump up for picks 191-200. That is referred to as the "Brady Effect" because he was selected 199^th in the 2000 draft. In fact, without this bump there may not be a significant break in value after pick #160, combining the two final tiers mentioned above.

This is a bit of an oversimplification because other players in the 191-200 range have also performed well, but the effect of a few players causing any visible increase shows the relative lack of success of late round prospects. Assuming your team can trade down and pick up a player like Brady or Terrell Davis, or even a more typical starting player, just isn't realistic.

Getting your money's worth

Then there's another definition of value that most NFL fans closely identify with: getting the most bang for the buck. Discussions about changing rookie contract structures to mirror the NBA's slotting system have been more frequent in recent years because of the extremely high contract amounts awarded to the top selections.

The graph below compares the average annual dollar amount of 2007 rookie contracts to a typical team's draft pick trade value chart (blue bars). That chart assigns a numeric value to each pick in the draft, so that when teams call to inquire about switching selections they are prepared to quickly analyze the benefits and costs of such a move. The top pick, for example, may have a value of 3,000 points while seventh round selections may be "worth" less than 30 points.

Using the trade chart published by Gil Brandt on NFL.com multiple times over the past few years, it's clear that first round picks are simply not as good of a bargain as most second round picks. The value per dollar figures are greater for picks 31-40 and 41-50 than either of the top two selection groups, and the value per dollar for picks 51-60 equals that for selections 11-20. (Note: using average rookie contract values can be problematic due to incentives and other complex contract language for early picks, but it was the best number available for this report.)

Frankly, teams' evaluation of the value of top twenty selections in the trade chart is optimistic at best. There is a 56% decrease in value on the trade chart between the first overall selection (3,000 on the NFL.com chart) and the 10^th overall pick (1,300). So the likelihood of the number one pick being the best player in the draft is almost twice as large as the 10^th pick? That's unlikely.  In fact, the games played, starts and Pro Bowl figures for the 9^th and 13^th overall picks (9^th-14.5, 13.4, 0.20; 13^th-14.8, 11.8, 0.20) are not significantly different from the number one pick (13, 12.2, 0.26). A similar drop occurs between the 10^th pick and the 30^th overall pick (620), which also overstates the difference in value between those players.

Bringing these values a bit closer to reality results in the revised value per dollar figures (red bars) in the graph above. A revised trade value chart was constructed, where instead of nearly a 5-to-1 ratio between the value of the first and last picks in the first round, slightly more than a 2-to-1 difference is used. This results in a much higher relative value per dollar figures for second round picks…and even third round selections are consider better bargains than first rounders in this data.

Top picks are over-valued by teams on their trade chart, however, partially because they need to drive up the price of top picks so the team does not appear to be giving away the farm when taking themselves out of the race for the top players. Even though teams know there is little real difference between the first and fifth picks, they would be lambasted if they only took a fifth or sixth round pick to move down those four spots. And, indeed, the team trying to move up is often desperate enough to pay the higher price, so the team in the higher position should demand more.

But similar precipitous drops also occur in contract amounts for the top one or two players and later mid-to-late first round picks. QB Jamarcus Russell averages about $4.9M over the six years of his deal, while #11 selection ILB Patrick Willis received about $2.5M per year. Even considering the premium placed on quarterbacks that disparity seems a bit wide, especially considering the production of both of those players so far in their young careers.

All this makes second round picks the best value of the draft, which is not surprising considering the contract premium placed on being a first round selection. The average annual contract value for the last pick in last year's first round, WR Anthony Gonzalez, was just under $500K more than that of the first pick of the second round, DT Alan Branch. That's almost $2.5M over five years. The difference between Gonzalez's contract and the penultimate first round pick (Chicago TE Greg Olsen) was only $30K, or $150K over five years. There's not a whole lot of sense behind those dollar figures.

A study by professors Richard Thaler and Cade Massey, called "The Loser's Curse", analyzes how losing teams in the top ten are actually penalized by receiving the top picks because of the large difference between those players' values and the amount of the rookie contract. Most young players simply do not produce enough in their first few seasons to justify the size of the contract, which also hurts the team's ability to improve their team by signing veteran free agents. Teams having top picks may also have a difficult time signing veteran backups at those positions because they know the team invested a lot in the young players.

Versatile verbosity

It's amazing how one word can have so many meanings on draft weekend. Whether it's grading a team's pick, analyzing how a team evaluates a player, understanding the historical trends of the draft, or identifying players that will be bargains due to lower contracts, the term "value" will be repeated throughout network coverage of the draft…and for some time later in media outlets and on the Internet.

Of course, no one will really know the player's worth for at least three years. But in today's analysis-crazy media world, that sort of patience is not a virtue.