Value & ADP - Quarterbacks

  Posted 7/4 David Yudkin - Exclusive to Footballguys.com

This is the second in a series of articles focusing on an evaluation of historical data to determine where value lies in drafts to better identify historical trends and potentially better understand how things play out AFTER the fact in an effort to construct more effective drafting strategies for future seasons. Since this is more a discussion on the RESULTS than individual players, in most cases I will not be listing the names of the actual players involved. The point IS NOT to debate individual players and where to take them on draft day but rather to look at the results to see WHERE the players were selected and what value scores might be available at different points in a draft.

To clarify, when I mention value, I mean value as discussed in the first article in the series. We could debate the merits as to which is actually better for an individual team . . . drafting the third QB off the board that yields a value of 82 points or selecting the 14th quarterback off the board that produces 41 value points. That is an interesting proposition and a debate in its own right, but I am more interested in the likelihood that players taken in those spots will PRODUCE at those levels rather than which one was a better draft pick.

But first let me explain what it is people will be looking at. What follows are charts of positional ADP data from 1998-2006 for quarterbacks. I listed the value generate from each draft spot for each season. You will notice that I intentionally left off the names of the players that were drafted because for our purposes the actual players don’t matter. The data was compiled from various sources collected from drafts very close to the start of each season. Whether it matches up exactly with ADP from other sources really won’t make that much difference in most cases, as it’s unlikely an individual player’s ADP will vary by all that much. The only data I had available was POSITIONAL ADP and there’s no way to know what round players were taken in. While that would have been a plus, that could have made things more confusing in its own right.

I color-coded the results to make things easier to read and quickly identify. Here is the breakdown by color. Each color represents a scoring tier of that number of value points. If a player had a value of 200 or more points, he will have his score in a red box and so and so on down the line. For example, there were six players taken as top five picks that scored between 75 and 99 value points. The exception to that is the 1+ category, which includes ALL players that had a positive value score.

The second chart reflects a grouping of QBs in five draft pick increments. QB 1-5 means any QB taken in the top five quarterback picks in any season in the years of the study. In this chart, the numbers are indicative of players that actually scored in that range. The totals for each scoring tier are provided at the bottom of each column.

Observations:

Taking a QB early can be EXTREMELY risky, as the “bust rate” of the Top 5 drafted QB is almost 50% (no positive value). Many times that is a direct result of injuries, and while any top pick can get hurt, many top quarterbacks HAVE SHOWN that they get hurt. Unfortunately, taking a QB early can result in a great reward or a total failure with only moderate returns in the middle.

Half of the QBs with value in the QB 1-5 section had value scores of 75 points or more, which is an extremely good ROI from a QB, even if it took a relatively early draft pick to produce those results. Another statistical oddity is that the QB 6-10 range produced fewer busts than the QB 1-5 range (47% vs. 36%). However, while there may have been more players with positive value scores, there were nowhere near as many players in the elite value tiers.

Part of what makes drafting a quarterback early a daunting task is that in four of the nine seasons evaluated there was not a Top 10 drafted QB that had a value score of 100 points. That, in and of itself, could be just cause to avoid drafting a QB early. Another challenge here is that the results are conflicting and indicated opposite trends—the early years in the study are a lot different than the more recent ones.

The past four years have seen a change in scoring patterns compared to the first few years of the study. Of the 40 top ten QBs picks in 2003-2006, only 21 had positive value and only 10 of those had value scores of 50+ points (20%). The huge majority of value came from top 15 picks - 32 of the 42 quarterbacks that had positive value (76%). By comparison, in the 1998-2002 era, there were 20 top ten picks with 50+ value points out of 50 picks (40%).

In the past five seasons, only three QBs had 100+ points in value: Manning twice and Culpepper once (and Culpepper does not appear to be primed for an encore performance). This tells me that there are better middle-of-the-road quarterbacks in the NFL and the baseline of the QB12 has risen. In 1990, the #12 QB scored 200 fantasy points. In 2003, the #12 QB scored 269 points.

The number of higher scoring QBs in the second tier of scoring is on the rise with elite (tier one) QB scoring on the decline. From 1990-1999, there were 35 QBs that scored 300 or more fantasy points (3.5 per year). From 2000-2006, there were 43 (6.1 per year). Other than 2004, which was both a statistical bounty and anomaly, combined scoring of the top three quarterbacks has been down in recent years. The 2003, 2004, and 2006 seasons saw the top three combined score under 1,000 combined points (957, 922, and 986). The last time there were three of four years with the top three scoring under 1,000 combined points was 1991-1993.

This probably boils down to investing in Peyton Manning or waiting several rounds, as Manning has been extremely reliable and has not gotten hurt. Many people will insist that burning an early round (even a first round) pick on Manning is a huge blunder and fantasy faux pas, but his solid performance over the years makes him at the very least a consideration for people struggling with who to take after a bevy of running backs have come off the board.

One popular strategy is to wait until ten QBs have been drafted and then pick two back-to-back. In the nine years accounted for here, there were nine QBs drafted in the 11-20 bracket with value scores of 50 or more (10%). Beyond the top 15 QB selections (and counting only those players that generally were drafted—so those in the QB 16-30 range), there were 19 QBs that returned any value of the 135 players selected in that range (14%). Only six of those players had a value of 50+ points (or 4% of players drafted in that range). Compare that to the QB 1-10 bandwidth, where 30 of 90 players scored 50+ points (33%).

The theory of waiting on a QB early and “grabbing someone just as good” later on may sound good but may not happen as often as people like to think it does. However, the results of the past few years may show that while few quarterbacks taken deep in drafts will yield positive value, there many not be many QBs with value to begin with so it may not be life threatening.

The wait and grab a top flight QB theory was never more prevalent than in 1999 and 2000, when there were six late round QBs that had value scores of 75 or more. There were only 22 players that had scores of 75 or more points in all nine seasons and six of them were late round picks in those two seasons.

Conclusions:

No matter how you slice it, the quarterback position is one of the more difficult ones to evaluate in terms of predicting value. It may be somewhat easier to determine a decent approximation of projected rankings than it is projected value. For example, if we were 100% assured beforehand who would rank as the top quarterback in any given season, that knowledge still might not be the most useful information. Carson Palmer ended up the 2005 season as the top ranked quarterback, but in terms of overall value he ranked 20th for all positions. Peyton Manning has generally been the first QB off the draft board in recent years, and his yearend value ranking (again counting all positions) has been 24, 30, 2, 34, and 4. So Manning was a great pick in two of the past five years but even being a consistent producer didn't help him in the VALUE rankings.

Given the mixed results on drafting a QB relatively early (turned out to be good idea in the first few years, but not so great an idea in recent years), there is not a conclusive set of results to be inclined to either do it or avoid doing it. With injuries somewhat prevalent at the QB position, it may be hard to really compare the value of drafting a RB early, as a QB that put up great numbers in 12 games but missed four others may still have been a hot fantasy commodity.

One thing seems clear: that hoping to get a true difference maker past QB15 does not happen very often. One of the oddities about fantasy drafts is that in middle rounds the options will be a full time NFL starter at QB, a backup at RB, a decent NFL TE, or a NFL WR2. Fantasy scoring wise, the QB will almost certainly score the most fantasy points but could still have no fantasy value.

For the most part, pairing two or three mid round QBs may not net great production from any of them when compared to other top quarterback production. Some people may be able to utilize a QBBC successfully, but weekly scoring could turn out to be very inconsistent and may not come close to matching the top quarterbacks overall.