Orton's Arm

I love the idea of the Bills taking a quarterback. I'm a little leery of Tannehill. I've heard mixed reviews about him. Also, he only averaged 7.1 yards per attempt in college. I feel a QB needs to average 7.2 - 7.4 yards per attempt in the NFL to be considered franchise. On the other hand, I've never actually watched him play. Maybe his yards per attempt stat understates the quality of his accomplishments. If the Bills do pull the trigger on a first round QB, I want them to be reasonably certain he's The Guy. The Guy needs to have a high level of accuracy, he should be able to see multiple reads quickly and process information quickly, and he should be good at pre-snap reads. As a bonus, I'd like him to be able to hit receivers in perfect stride. I'm less concerned about physical attributes, but I certainly wouldn't complain if he added mobility and arm strength to the above list of traits. If Tannehill meets these criteria, then the Bills should be willing to trade up a lot higher than seventh overall to get him. I don't think there's been a single instance in NFL history in which a team said, "even though we got a young franchise QB from this trade, we regret having made it, because the price was too high." If you get a franchise QB, the price you pay will not be too high! But if Tannehill isn't the answer, then the Bills should look at non-QBs in this draft. One possibility is to take a WR or CB at tenth overall, then to trade back into the first round for a LT. Trading away picks from next year's draft should not be an option. This means that trading back into the first will probably require the second and third round pick from this year's draft. (Yes, I understand that trades require a willing trading partner, and that the existence of such a partner is not necessarily a given.) The Bills might use a fourth or fifth round pick on a SS/LB tweener, with the intention of turning that tweener into a full fledged LB. A draft like this would address two of the Bills' three biggest non-QB needs (which are LT, CB, and WR), while also improving the linebacking corps and team depth.

A number of good points have been made in this thread. I agree with Cash that you don't want to draft a player in the first or second round if you know no one else will take him before the third. Cash has made some other good points as well. But Ramius has a good point too: a team has to evaluate best player available on the basis of information available at the time. Based on that information, a number of analysts--not everyone, but a large number--felt that Spiller was the best player available when the Bills picked. I'm not a big fan of using first round picks on RBs, as anyone who's read my posts may notice! But I wrote that if Spiller became the next Thurman Thomas, he'd justify ninth overall. By that I meant that Spiller isn't going to justify ninth overall by rushing alone: he also needs to make a Thurman-like contribution to the passing game. I'm encouraged by the flashes Spiller showed late last season, as well as by the fact that the Bills had sometimes lined him up at WR! The fact that I've set the success threshold for a RB higher than I would have for, say, a LT, is deliberate. If a RB and LT are roughly equal as football players, a team should take the LT. Most teams "get" the concept of position value, which is why it is literally inconceivable that the Colts will choose DeCastro over Andrew Luck. And yet, there are those who would have us believe that all 32 NFL teams would choose DeCastro over Luck if they felt DeCastro will be slightly better at OG than Luck will be at QB. As a Bills fan, I would love it if the first nine teams in the draft were that stupid!

I've always felt that where a player is taken in the draft should roughly correspond with his achievements in college, his performance at the combine, how well his physical and mental attributes project to the NFL, and other factors. For someone using this definition, "reaching" consists of taking a player at a much earlier point in the draft than his college play and other factors would warrant. Obviously, there are different methods by which players can be evaluated. For example, the Bills thought of Trent Edwards as a first round talent, whereas the other teams in the league were content to let him "fall" to round 3. How do you know whose player evaluation techniques are the best? Judge by results, because nothing else matters. By analyzing results, one can compare draft analysts both against each other and against NFL GMs. The GMs and draft analysts with the best past track records will generally produce the most accurate future predictions. A "more accurate than average" player evaluation technique will still result in some misses, just as a "less accurate than average" technique will still result in some successes. When someone diverges from the conventional wisdom--be that the conventional wisdom of draft pundits or the conventional wisdom of the majority of NFL GMs--one should not automatically conclude that the person diverging is wrong. It's possible that the analyst or GM in question has developed a new player evaluation technique, and that this technique is at least as good as the techniques being used by those who create conventional wisdom. Then again, it's also possible that the person who drifted away from the pack had no idea what he was doing. This form of uncertainty can be resolved with the benefit of hindsight. The best draft analysts have better track records than the worst NFL GMs. What's the good of a sophisticated player evaluation technique if the results it produces consist of Whitner, McCargo, and Lynch? Marv did not have a single success story not named Kyle Williams. All his free agent signings were bad, including Peerless Price, Melvin Fowler, Tuten Reyes, Robert Royal, the backup DT we signed from the Colts for starter money, Greg Jerman, Derrick Dockery, etc. His head coaching hire was a mistake. His first, second, and third round picks were all busts. He didn't get anything from the later rounds either, except of course for Kyle Williams. I'm confident that if you were to give a reasonably good draft analyst two years' worth of draft picks, including two picks in the top-12, that this analyst would probably achieve more in those two years than just Kyle Williams.

Good post, and I agree with it. Fitz is an upgrade over Trent Edwards, and is okay as an NFL starter. Better than some, worse than others. Unfortunately, nine out of ten of the last Super Bowls have been won by teams with franchise QBs. One can be reasonably confident that the Bills will not win a Super Bowl with him under center.

Good post. I sense you're not feeling as frustrated with this discussion as I've become, which may be why you've expressed yourself better than I would have had I addressed the points you've made. Instead, I'll take this opportunity to step back, and address a few of the points which have been made. Regarding Cutler, I wanted him before the 2006 draft--so much so that The Dean began derisively referring to me as Cutler's Arm. Regarding Losman, I was unhappy with the pick from the beginning, because I felt he hadn't demonstrated himself to be a good, polished pocket passer at the college level. I felt a first round pick was used on him because of his physical gifts. I would never use a first or second round pick on a QB with that as a primary basis. K-9 has repeatedly asked me what data sets NFL teams lack, that they would need to make optimal decisions. I don't know why he's asking this. He already knows that I have no inside knowledge of front offices' workings. How can I know what data sets they do or don't lack? In answer to Cynical's post, the fact that QBs and LTs are far more likely to be taken in the top 5 or top ten than are interior OL or non-pass-rushing LBs proves only that NFL teams are more likely to draft players at some positions (the ones I've labeled tier 1 and tier 2) than they are at positions I've labeled tier 4. If a good QB or RDE is harder to find than a good OG, it may be because the good QB or RDE is far more valuable than the good OG. Initial fan attempts to quantify that difference will be subjective (not arbitrary), and a well run front office should seek more sophisticated measurements of the relative values of positions. (The alternative is to assume that, because relative position values cannot be exactly measured, they should be ignored and labeled "arbitrary.") Some in this thread have made inaccurate statements about my formula. To resolve the resulting confusion, it may be better to explain its intended meaning with words. 1. If you have a pick in the top-15 of the draft, your best bet is to see if there's at least one player from a premium or close-to-premium position with whom you feel comfortable. If there is, you should probably take that player, even if there's a player at a non-premium position who has a moderately higher player grade. In this draft, for example, I feel the Bills should be looking closely at the available LTs, WRs, CBs, and QBs, to see if there's someone from those categories with whom they feel comfortable. 2. If you know that a player isn't going to be with your team for a long period of time, either because his position is associated with a short career, or because your team has adopted a policy of allowing first round CBs to go first-contract-and-out, then you devalue his position accordingly. If a successful CB is expected to remain with your team half as long as a successful LT, then you multiply the CB's value by 0.5 when comparing him to the LT. 3. The purpose of the first two steps isn't to tell you which specific player you should take. The purpose of those steps is to allow you to focus on a relatively small subset of players. If three or four guys have roughly similar values according to the formula, it's okay to choose from among the three or four based on need and on your overall strategy for building the team. It's also okay to incorporate a gut level "feel" for whether a player will be successful. The Bills have shown us why the first two steps are necessary. Over the last 40 years, they've used 25% of their first picks of the draft on RBs, and another 25% on DBs. 0% - 3.8% have been used on QBs (depending on how you want to do the counting), and 5% have been used on OTs. A team which jumped straight to step 3 would be liable to make these kinds of mistakes. The formula is largely a response to these errors in drafting. It's intended to stop teams from ignoring premium positions to focus on RBs (as the Bills have traditionally done). It also states that if you're going to let your best DBs go first-contract-and-out, then you don't take a DB in the first round.

Fitz is part of what's wrong with the Bills' deep passing game. Fitz isn't an accurate QB in general, and that lack of accuracy is accentuated on longer throws. Problems with the OL--especially later in the year--also hampered the deep passing game. As you pointed out, the lack of a speed receiver is another facet of the problem. Unfortunately, the Bills won't be able to solve all three of those problems in the upcoming draft. But even solving one or two of them would be a step in the right direction. Finally, the fact that Belichick has given up on Mallett is a strong indication that he, at least, does not think Mallett has the mental tools necessary to be a franchise QB. Belichick has reached this conclusion after having a year to evaluate him with the Patriots. I was very leery of exactly this even before the draft; which is why I would not have drafted him in the first three rounds.

That's the kind of logic which causes the Bills to take a first round RB every 3.5 years. Aaron Maybin was a one year wonder. The lesson to be learned there is that you don't take a 220 lb DE at 11th overall on the basis of one good year.

Your posts are so disconnected from reality anyway that at this point, moving even farther from reality won't make much difference. As for the specific example of running back, I've said that a RB can move to a higher tier if he's an integral part of the passing game, like Marshall Faulk or Thurman Thomas. If you don't believe I've said this before, go back and reread my posts. > What part about that hypothetical G being a hypothetically better FOOTBALL player > than that hypothetical LT don't you understand? Again, if it's a choice between a top-15 LT and a Pro Bowl OG, you take the LT. The fact that you flat-out disagree with this causes me to know--not just think, but know--that you're not speaking for NFL GMs. Most teams aren't run the way you say they're run. Over the last decade, many more LTs have been taken in the top-10 of the draft than OGs. Many more QBs have been taken in the top 5 than non-pass rushing LBs. NFL GMs are taking position value into account. You are trying to tell us they are not, you are wrong, and the NFL draft history proves you are wrong. > If you were building a team you'd constantly be stuck reaching to fill a need. If you have a choice between a LT and a RB, and if the RB has a slightly better grade, taking the LT is not "reaching." Stop altering the meaning of words in your futile effort to win the argument. > Meanwhile, in the real world, good teams would be stocking their rosters with > the best players available and adding superior FOOTBALL players to their rosters. Superior football players, but at inferior positions. They'd have great OGs, RBs, non pass rushing LBs, etc. Their QB, LT, CBs, and RDEs will stink. How many games do you think a team like that will win? > According to your formula, the only thing that matters is the relative worth of the position. Go back and reread the formula. Relative position worth is only one of three equally weighted variables in the formula. > But if your draft board lists best QBs first, followed by the best LTs, followed by the > best DEs, etc., good luck finding the best players. That is not the result the formula would produce. I am also relying on GMs to exhibit some intelligence when applying the formula's results. > Marv's drafts were terrible NOT because of the positions they drafted. They were > terrible because the players they took just weren't good enough. Marv's drafts were terrible for both reasons. He had a gaping hole at quarterback. Picking at 8th overall, he chose to pass up Cutler for Whitner. That was a terrible decision not just because Whitner failed to justify a first or second round pick, but also because the lack of a quarterback greatly diminished what the team could otherwise have accomplished. > But I have a feeling that simple concept might escape you given your intractability on the subject. Lenin said to always accuse your enemies of that which you yourself are guilty. If it's been your goal to be as intractable on this subject as possible, you've succeeded magnificently. > This is why debating the issue with you is worthless. The reason you've failed to convince me of your point is that you've contributed nothing to this discussion beyond your own blind, dogmatic insistence that relative position values should be ignored on draft day. You have yet to provide a single logical reason why this is the case. You are one of an (unfortunately large) group of people who seem to think they don't have to provide logical arguments to sway others to their view, as long as they insist on their own view loudly enough. Anyone who won't be swayed by their loudness is automatically given negative labels. That persuasive technique has been your main contribution to this discussion. > I provide you with a typical and common scenario (some might say I 'added to the > discussion' even) and you immediately say that in order for it to be valid it has > to be quantified statistically. Once again you have distorted what I've written. The example you gave was of a LB who dropped back into coverage and forced a QB to go to his second read. Because the QB was forced to hold the ball a little longer than normal, the SS was able to tip the ball into the hands of a cornerback for an INT. I never said your example was invalid. I did say I'd like to see it quantified. The reason for this quantification is to prevent the sort of vague, overly optimistic "oh, I'm sure he'll justify his very early draft position with lots of forced interceptions," type thinking of which you are guilty. Count the number of interceptions caused by good coverage linebackers and then decide where in the draft you want to take a coverage LB! > The discussion has reached it's [sic] end because we have gone back and forth on it long enough > and I have no more to add. You had nothing constructive to add in the first place. > Oh, I think it was pretty easy for you to disrespect my opinion all along. You loudly and confidently expressed disagreement with my opinion, without stating logical reasons why. You gave no indication that you'd thought deeply about anything which was being discussed. Instead you cavalierly and arrogantly dismissed what I'd written, as in the below example. > I'm calling you arrogant for suggesting that your calculations, based on a > completely arbitrary set of numbers, can even remotely predict the career > trajectory of a player entering the draft. Instead of entering into a constructive discussion about how my formula might be modified to better represent a player's long term contribution to the team, you simply dismissed it. Thinking about how to improve a formula takes work. Negatively labeling someone who disagrees with you is easy. Given a choice between contributing something useful to the discussion on the one hand, or negatively labeling those who disagreed with you on the other, you did not hesitate. > "Do us ALL a favor?" Don't you mean just you? No, I mean everyone. No one benefits when people employ the persuasive techniques you've adopted. Even if you're right--which you're not--but even if you were, no one here is going to learn much of anything new if your only contribution to a discussion is a dogmatic insistence on your own perspective, coupled with one or two extremely shallow arguments in support, and plenty of attacks on those who disagree. > Or can you speak for everyone on the board? Of course not! On the other hand, I haven't heard anyone say, "We need to start replacing logical, football-related discussions with petty personal squabbles." Until people start endorsing the techniques you use, don't consider those techniques endorsed.

I performed those calculations because you asked me to. Now you're calling me arrogant for having performed them? The next time you ask a favor of me, I'll assume you're asking in bad faith. > That does NOT suggest you never take a RB in the first round. Depends on the RB. You seem to flat out reject that. I do not flat-out reject that. Go back and reread my posts. > But building teams is NEVER a linear process. If you don't have one > of those premium positions filled and a blue chip player is available, > you take him. But you DON'T reach for them. Agreed. > You seem to suggest that even though a player like Keuchly may be the highest rated > player available, it makes more sense to take a lesser talent at one of the more > 'important' positions. That's flat out wrong and I don't know how else to put that. If you are not taking into account the relative value of positions, then it is you who are flat-out wrong. As I stated earlier, a top-15 LT is worth more than a Pro Bowl guard. If it's a choice between the two, you take the LT. A team which ignores the relative values of positions on draft day--or which embraces misguided notions about those values--will tend to have a worse record than a team which correctly takes position value into account. > But you won't see a tiered ranking system because player grades are assigned according > to ABILITY, regardless of position. I find it impossible to believe that all NFL teams are run as badly as you've just described. If you're claiming to have enough inside knowledge to definitively state that they are all run this way, then I simply don't believe you. I'm willing to accept the notion that there are some teams which ignore position value on draft day. If a poorly run team like that had a choice between a RB and a LT, and if the RB had a slightly higher grade, it would take the RB. I could easily imagine TD or Marv doing something boneheaded like that, either because they're ignoring position value, or because they (erroneously) believe RB to be a premium position, or because they want a quick fix. > What if a LB drops into a coverage and forces the QB to go to his second read > and this hesitation causes him to throw a ball that's tipped by the SS and into > the hands of the CB for an INT? If you want to quantify that, look at the best coverage LBs in the league today, and count the number of INTs they typically create over the course of a season. The resulting number will help you determine the value a good coverage LB can contribute to your team. Quantifying that contribution helps avoid flights of fancy, vague thinking, or wishful thinking. > I think we've reached the end of this discussion. Right. Because obviously a discussion ends when you've had the last word. > I think perhaps Matthew's Bag is right. You're making it very hard for me to respect your opinion. > In the future I'll try not to let your over-intellectualization of the draft, > the game, or anything else associated with it get in the way of your fun. Thank you. Your initial responses to my posts clearly lacked any intention of adding anything constructive or useful to the conversation. While you have occasionally thrown in a football-related tidbit, your overall tone has been obnoxious and argumentative. Next time do us all a favor and don't "contribute" anything at all.

I have little or no knowledge about the inner workings of front offices. Lacking this knowledge, I independently considered the question of how one might run a football team. If independent thought seems arrogant to you, it might mean you're using the word "arrogant" as a synonym for "non-herd-like thinking." > I thought your 3.5 was a self-deprecating reference to a legendary thread on PPP It was intended to be humorous, but not self-deprecating. The concept of expected value is widely recognized, the expected value of a die roll is 3.5, and anyone who says otherwise is just making stuff up. See example 1. I'm not adverse to self-deprecating humor. But my joke about 3.5 wasn't intended to be at anyone's expense, either my own or anyone else's. > When I think of the thousands of man-hours spent by both Blesto and National > in compiling their player rankings . . . I cringe at the arrogance you display here at times. I have on multiple occasions acknowledged that I haven't watched any college football this past year; and have invited comment from those with more pre-draft knowledge than me. The fact that you're nevertheless trying to pin the "arrogant" label on me requires some very serious explaining on your part. > I can say with certainty that nobody I know currently or in the past, has a tiered positional value chart ranging from 1-4 On the other hand, I have heard that better run front offices prefer to use their early picks on premium positions. The four premium positions are QB, LT, RDE, and CB. If you accept that a quarterback is more valuable than an equally good LT or CB, then that justifies the first two of my tiers. As for some of the other tiers, Mike Shanahan once said that if you have good tackles and a good center, you can get by with decent guards. Comments like that suggest Shanahan is placing a higher value on centers, RTs, and LTs than on OGs. I find it extremely difficult to believe that Shanahan is the only one in the NFL who values some positions more highly than others. A tier system just a formalization of this practice. > Perhaps my biggest gripe is that NFL personnel evaluators seem to be acutely aware > of one thing that you and all your moneyball number crunching seems to take for granted: > that these players are young, flawed, human beings and no matter the amount of preparation, > there is simply no compensating for that with a scientific formula. I acknowledge that they're young and flawed. Show me anything I've written where I've said otherwise! The draft is a guessing game, but some guesses have a higher probability of being right than others. With a large enough data set, a good mathematical formula will increase both the number and the degree of your team's successes. Conversely, if a mathematical formula does not cause you to make better choices than you otherwise would have made, it's not a very good or useful formula! > I was being honest when I suggested you start a draftnik site using these various formulas > because people will probably buy it. You chose to accept that as an insult I wasn't offended by that. > Here's a word of advice if you're gonna talk to NFL teams: they don't mind 'subjective' so much as 'arbitrary'. Obviously, I haven't done even 3.5% of the research or analysis necessary to create something I'd send to NFL teams! Nor do I intend to. So I'd like to begin by taking that unrealistic option off the table. The position values I've mentioned are subjective, and there's certainly room for revision to make them more accurate. But they are not arbitrary. The New York Times did a regression analysis which showed that yards per pass attempt is three times as important as yards per rushing attempt; and that interception percentage is as important as yards per rushing attempt. In addition, nine of the last ten Super Bowl winners had franchise QBs. Based on these and other data, I tended to assign positional importance on two primary bases: 1) the degree to which the position was associated with good pass offense or good pass defense, and 2) the opportunity a good or elite player would have to substantially change the overall dynamic of pass offense or pass defense. As an example, Bruce Smith changes the dynamic of your pass defense, because he forces the other team to double team the RDE. He adds +1 to the numbers game. In contrast, it's unlikely a LB will be so good in coverage that he can be put into one-on-one coverage against a good pass catching TE. So there's less chance of him adding +1 to the numbers game on passing downs.

A lot of the resistance to taking a CB in the first round is that during the post-Polian era, the Bills have traditionally allowed their first round CBs to go first-contract-and-out. Then additional first round picks would be used to replace the guys who'd just departed. Over the last 40 years, the Bills have used 25% of their first picks of the draft on RBs, and another 25% on DBs. Clearly that's unacceptable, and is a major reason why this franchise has a losing record. But if the Bills take a CB 10th overall, if he plays at a level consistent with his draft position, and if he stays here his whole career, I'd be fine with the pick.

With a guy like Mallett, my biggest concern wouldn't necessarily be either the drugs or whether he was healthy. It would be whether he could read defenses and quickly process information. I've read that once the ball was snapped, Brady could process the situation over a full second faster than Bledsoe. Give me someone who processes information at Brady speeds, not Bledsoe speeds. (Even though Bledsoe had the better arm.) If I had good reason to believe that Mallett was good at reading defenses and quickly seeing multiple reads, then and only then would I be tempted to take a chance on him. As you pointed out, he very clearly has the physical tools needed to be a franchise QB. If he has the mental tools also, not much stands in his way! If he doesn't have those mental tools, he'll never be more than a backup or marginal starter.

You're acting as if you know a lot more than you actually do. This false knowledge on your part is the reason why your post is arrogant, annoying, and contributes nothing to the discussion. Even though you don't know the extent to which front offices take position value into account, or how they take it into account if they do, you're acting as if you're sure my method would be completely novel to all 32 of them. In the future, do not pretend to know more than you actually do. > How is quality of play derived again? And what constitutes reasonable? I answered these questions in my previous post. Go back and reread it. > And the position values, how are they determined? The numbers I used represent my subjective estimate. In the unlikely event you have something constructive to contribute to a discussion about how those subjective estimates might be made more accurate, I'd be happy to listen. If, on the other hand, you're planning on writing something along the lines of "subjective = worthless," then don't bother. You've wasted enough of our time already.

Yours is a reasonable request. There are two ways of creating a reasonable best case estimate for a player's value. One is to watch the player yourself. The other is to base your opinion on draft experts. If there are ten experts whose opinions you trust, and if at least one of those experts thinks a player has Pro Bowl potential, then Pro Bowl becomes a reasonable best case scenario for that player. I haven't watched college football this year, which leaves me to rely on the scouting reports I've seen. To be honest, I haven't gone through that many scouting reports either. If someone with more pre-draft knowledge wants to amend my player predictions, I certainly won't stand in that person's way! That being said, below are best case scenarios for the players in question: Kuechley. Reasonable best case quality of play = 85. Position value = 3. Best case years with team = 11. Reasonable best case = 85 x 3 x 11 = 2805. Gilmore. Reasonable best case quality of play = 80. Position value = 7. Best case years with team = 11. Reasonable best case = 6160. Barron. Reasonable best case quality of play = 60. Position value = 3.5. Best case years with team = 11. Reasonable best case = 2110. Floyd. Reasonable best case quality of play = 80. Position value = 6. Best case years with team = 11. Reasonable best case = 5280. The worst case scenario for the four players is that they're all busts.

Before we go any further, I think we need to clarify what's meant by "draft value." If "draft value" is taken to mean whatever Kiper and others like him says it means, then your point is correct. We shouldn't assume that the NFL team which made the pick is wrong, or that Kiper is right. But there are other ways one can define draft value. I'd like to propose the following: Draft value = player value / draft pick value Draft pick value. To obtain this, you start with an NFL draft value chart, and modify it according to the talent pool available that year. Player value = the quality of the player's play x the number of years with the team that picked him x the value of his position Using the above equations, let's look at Marshawn Lynch's draft value. Player value = quality of play (40 for Lynch) x years with the team (3, I think) x value of his position (3.5) = 420 Compare that to a hypothetical QB at or near the Pro Bowl level. Player value for QB = quality of play (70) x years with the team (12) x value of his position (10) = 8400. Lynch represents 5% of the player value of this hypothetical QB. Because Lynch's player value was low, and the value of the 12th overall pick is high, he represents a bad draft pick value. One can use the benefit of hindsight to obtain reasonably firm numbers for the Lynch calculations. If one were to perform a similar calculation regarding a player who was just drafted, many of the numbers used would be estimates. The two biggest unknowns would be the expected quality of play, and whether a player will be allowed to go first-contract-and-out. By plugging different numbers into those unknowns, one can create reasonable best-case, worst-case, and expected case scenarios for early draft picks. That kind of analysis would have shown that even in a reasonable best case scenario, Marv did not get great value from the 8th overall pick (Whitner) or the 12th overall pick (Lynch).

Good response. I certainly don't object to your pointing out examples of players who'd been called reaches, who went on to justify their draft positions! In another thread, I wrote about different tiers of player value. Tier 1: QB Tier 2: LT, RDE, CB Tier 3: DT, RT, WR, C. Tier 3.5: pass catching TE, S, LDE Tier 4: OG, RB, non-pass-rushing LB. The RB can move to a higher tier if he's a pass catching, all-purpose RB like Thurman Thomas or Marshall Faulk. Of the three players you mentioned, one was a LT, one was a RDE, and one was a DT. That's two tier 2 guys, and one tier 3 guy. The data size is small, so any conclusions one might draw are merely conjectural. One possibility is that reaching for a higher tier player is safer--and associated with more upside--than reaching for a lower tier player. Another possibility is that when a front office of above-normal competence reaches for a player--as Polian did when he took Freeney--there's a strong chance of him being right. But when a front office of below-average competence defies conventional wisdom by reaching, odds are that front office is wrong. Maybe both suppositions are true. Certainly, Marv's front office was of below-average competence. Two of Marv's three first round reaches (Whitner and Lynch) played lower tier positions. Edit: I should also throw in something about the difference between a perceived reach and an actual reach! An actual reach is when you take a player higher than he deserves, based on his college performance and performance at the combine. A perceived reach is when pundits label it a reach. It's quite possible that Marv's perceived reaches were actual reaches, while the perceived reaches of Freeney, Levi Brown, and Alualu were not actual reaches.

Bob Chalmers has blown this thread out of the water. With his having expressed things so lucidly, there's little work for me left to do. As he pointed out, there's little point to taking that punter in the first round if you're reasonably certain no one else will take him before the fourth round. That's just a waste of scarce draft resources. Another reason not to take a punter in the first round is that a punter is considerably less valuable than an equally talented starter on offense or defense. In your example talent levels are not equal: the punter is HoF, while the available offensive and defensive starters are not. But which would you rather have: a Hall of Fame punter, or Stevie Johnson? A Hall of Fame punter, or Andy Levitre? I'd much rather have Johnson or Levitre over the punter, even though the former two guys aren't Hall of Fame players. More generally, I'd propose the following formula to get a (very rough) approximation for player value. Player performance x position value = player value The above formula is far from perfect, and I'm sure someone who's put more time and thought into this than I have could come up with something better. But the formula is a starting point, and is better than nothing. For the Hall of Fame punter, the formula is player performance of 100 (out of 100) x position value of 1 (out of 10) = player value of 100. For Levitre, the formula is player performance of 55 x position value of 3.5 = player value of 192.5. The above logic applies to all positions, not just kickers and punters. A top-15 LT is worth more than a Pro Bowl OG, for example. I mentioned the absurdity of taking a punter in the first round to illustrate a larger point: that the value of a player's long term contribution to your team is affected not just by the quality of his play, but by the relative value of his position. Another point to remember is that if you take a punter in the first, second, or even third round, you create a situation in which the punter has to play at or near a Hall of Fame level to justify the pick. There's very little upside--that is, there's very little room for him to give you more value than his draft position would have indicated. However, Hall of Fame level punters are very rare. There's an excellent chance the guy will fail to live up to that Hall of Fame success threshold which had been set for him. For every Hall of Fame punter, there are probably ten or even a hundred Pro Bowl level punters. Statistically, this guy is much more likely to be in latter category than the former. Finally, someone in this thread characterized the word "reach" as a made-up word. The pundits called Whitner a reach, and they were right. They called McCargo a reach, and they were right. They called Trung Canidate (a RB drafted in the first round by the St. Louis Rams to replace Marshall Faulk) a reach, and they were right. Of the players I remember having been called reaches, every last one has turned out to be a bust. Maybe that's just my memory being imperfect. But until there's at least some evidence to suggest that a player called a reach has as much chance of succeeding as anyone else, I, for one, am not going to dismiss the collective wisdom of football analysts. Any of those analysts would have done a much better job as Bills' GM than Marv did!

Eli has been playing like Peyton. For a while now, in fact. In the early years of his career, Eli didn't put up Peyton-like numbers or play at a Peyton-like level. He didn't play like Peyton during the season of his first Super Bowl. (Which is an important reason why the Giants only went 9-7 during that regular season.) But then something interesting happened during that postseason: for those few postseason games, he looked like a franchise QB, and was an important reason why the Giants won those postseason games. His playing at that elite level was a harbinger of things to come. In 2009, Eli averaged 7.9 yards per pass attempt. He followed that up by averaging 7.4 and 8.4 yards per pass attempt in 2010 and 2011, respectively. Those are three very solid seasons, even by his brother's standards. Peyton's career average is 7.6 yards per pass attempt. From 2007 to 2010, Peyton averaged 7.8, 7.2, 7.9, and 6.9 yards per attempt.

I wasn't being sarcastic, and I didn't contradict myself. The problem Bill Walsh was addressing was that scouts were beginning by trying to slate players into particular rounds. As I stated earlier, that's not a good starting point. The right starting point is to clearly visualize the team you want to have, and then to ask whether the player you're looking at has a place on that team. If the answer to that question is no, then you move onto the next player, without trying to pin a round onto the player you've just decided to pass up. If the answer is yes, then you start figuring out where you need to be in the draft to take the player you want. Joe Montana would have been great draft value even if he'd been taken first overall. The reason he would have been great draft value at first overall is because there weren't other, better players available in that (or just about any other) draft for the 49ers to take! But taking a Pro Bowl punter first overall is a boneheaded move--and represents bad draft value--because there are going to be plenty of other players in that draft who could have contributed more to your team than the punter. While it's impossible to perfectly predict how any given drafted player will perform, every draft pick you make should represent a good faith effort to add whichever player will contribute the most long term value to your team.

I agree with you about McNabb not seeming elite. When I'd think about the top QBs in the league, I'd think about Brady, Manning, etc. McNabb seemed one to two steps down from that. As for Kurt Warner, he definitely deserves to be in the Hall! No question. Over the course of his career, Kurt Warner averaged 7.9 yards per pass attempt. That's not just good. That's not just elite. That's truly frightening. To put that number into perspective, Peyton Manning's career average is 7.6 yards per attempt, Tom Brady's is 7.5 yards per attempt, and Joe Montana's is also 7.5 yards per attempt. Kurt Warner led teams to three Super Bowl appearances, including one victory. To those who say he didn't play enough to deserve to be in: he has 32,000 career passing yards, compared to 40,000 for Joe Montana. That's about two seasons' worth of difference. Warner and Montana have very similar TD:INT ratios. Warner's postseason stats are equally impressive. His postseason average of 8.55 yards per attempt is higher than any other QB in NFL history. I'm not a big fan of passer rating, but for what it's worth his average postseason passer rating of 102.8 is the second highest in league history. Warner's three Super Bowl appearances represent the three highest passing yardage totals in Super Bowl history. During the regular season, he achieved a higher percentage of 300+ yard games than any other QB in NFL history. (Kind of the anti-Trent Edwards.) He was the fastest player in NFL history to reach 10,000 passing yards, and tied with Dan Marino to be the fastest player to reach 30,000 career passing yards. Warner and Johnny Unitas are the only two QBs in NFL history to have four consecutive games with a passer rating of over 120. Not only does Warner deserve to be a first ballot Hall of Famer, he should be considered better than most other QBs in the Hall of Fame.

Draft value is not a moronic concept by any possible stretch of the imagination! Getting a Pro Bowl punter in the 7th round is good draft value. Getting a Pro Bowl punter in the top ten picks of the draft is not. Bill Parcells once said that you start by figuring out which players you want, and then you figure out which draft picks you need to get them. I think that's also what Walsh was getting at. Scouts were beginning discussions by describing the round in which a player should be taken. That's the wrong starting point, as Walsh recognized. You begin by envisioning each player's future role with your team.

In answer to your first question, I would not agree that the Bills could take anyone but Tannehill and be successful! The Bills need to avoid the error of blowing the 10th overall pick on some random RB, non-pass rushing LB, or OG! As for whether a player contribute immediately: Aaron Rodgers contributed almost nothing to the Packers during his first few years with the team. Do you think that the Packers regret having chosen him? The point I'm making here is that a player's long term contribution matters a lot more than what he can do in his first year! "What can you do for me immediately?" type thinking is a lot of what caused Marv to squander his early picks on "quick impact" players like RBs, SSs, and LBs. As for Tannehill, I've never seen him play, and have formed no firsthand opinion about whether he'll be successful as an NFL QB. Some analysts, like Gil Brandt, have him ranked among the top ten available players. There are other analysts who disagree. But if Tannehill becomes a franchise QB--which I have no idea if he will be or won't be--then the Bills would be fools not to take him! If your team doesn't have a franchise QB, and if there's one there for the taking, you take him. Period. Not only that, but you pay whatever (reasonable) draft day price you have to pay to trade up and get him! If you're against Tannehill because you think he'll be a bust, fine. There are plenty of experts who'd agree with you, and in a few years we'll know if they're right. But if you're against taking him because he's a quarterback, then you're wrong! Quarterback should be at the absolute top of the Bills' preference list for their 10th overall pick!

Bennett played in a 3-4 defense, and had much more of a pass rushing role than any LB will have in Wannestedt's 4-3 defense. Will Kuechley become a good football player? Very probably. Will his career be good enough to justify the 10th overall pick? Almost certainly not. Over the last decade, this team has often used early picks on easier-to-fill positions, while leaving holes at harder-to-fill premium positions. That needs to stop! I love the thought of adding players like Kuechley to the team. Nothing wrong with upgrading one's linebacker corps. What I don't understand is this willingness to neglect premium positions--positions which should be addressed with a top-10 pick, if possible--in order to draft a non-pass rushing LB. Getting a top-10 pick and coming back with a good, non-pass rushing LB is like getting a check for $200,000 with which to buy a new car and coming back with a Buick. Maybe you needed that Buick, and are better off with it than you would have been without it. But you didn't need it at $200,000!

Kurt Warner came into the 1999 season as the Rams' second string QB, but he did okay that year. Even in high school, Joe Montana displayed uncanny accuracy. Fitz bounced around the league for several years before coming to the Bills, and he's been with this team for a few years now. Trent Edwards was released early in the 2010 season. Fitz was the unquestioned starter for almost all of 2010, and all of 2011. He's already had his first full off season as the designated starter. And yet he still has accuracy issues. In the first game against the Patriots, he threw several passes which should have been interceptions. Fortunately for him, a combination of the Bills' WRs playing good pass defense, and the Patriots' DBs not being very good at intercepting, prevented those passes from getting picked off. Defenses started catching up to him--and to Chan's Fitz-oriented offense--even before his rib injury. In his two most recent seasons, Fitz has averaged 6.7 and 6.8 yards per attempt. This compares to a career average of 6.5 yards per attempt for Trent Edwards, and a needed career average of 7.2 - 7.4 yards per attempt to be considered franchise. The reason for Fitz's pedestrian average is his lack of accuracy, and odds are heavily against that lack getting corrected at this late stage in his career. If Joe Montana could figure out how to be a very accurate high school quarterback, why can't Fitz figure out how to be accurate after all his years in the NFL? No matter what the Bills do on draft day, Fitz will go into this season as the unquestioned starter. He'll have every opportunity to prove me wrong. For the sake of the team, I hope he succeeds. Also Fitz is a good guy, giving me another reason to hope for his success. But hope and belief are not the same thing. This team needs a franchise QB if it's going to win the Super Bowl. Not just a guy who's "good enough to get you by." Thus far, Fitz has not played at or near a franchise QB level. Unless that changes, the Bills will not win a Super Bowl with him under center. We don't have the talent at other positions to be the Ravens of 2000, and that's not going to change any time soon.

I appreciate the tone of your response. Now we just have to get the other stuff sorted out. I think that we're in agreement that the New York Times' regression analysis does support the concept that pass offense and pass defense are several times as important than run offense and run defense. Given that disparity in importance, I think the next obvious question is, if you pick a player 10th overall, what is he going to do to help your pass offense or pass defense? Hopefully we're on the same page at least as far as this. As an aside: I don't object to a player helping the run offense or run defense. But I'm averse to the idea of using that help as the primary basis for taking a player at 10th overall. If your rushing offense or defense need help, that's what later round picks are for. You are correct to say that LBs have responsibilities in pass defense. If there's at least one LB on the field, and if the QB drops back to pass, then obviously that LB isn't supposed to just stand there with his hands in his (nonexistent) pockets! I'm confused by why you seem to think I'm unaware of this. Just as it wouldn't make sense to draft a CB at 10th overall if you were only ever going to use him in zone coverage, the same logic also applies to a non-pass rushing linebacker. For a player to make a 10th overall-type contribution to pass defense, he either needs to be a good pass rusher, or he needs to be able to cover a good WR or pass catching TE one-on-one. One or the other. If all you need him for is zone coverage, you can get good zone coverage guys later in the draft. This is typically what Wannestedt likes to do. He's the sort of guy who will take a hard hitting college safety in the fourth or fifth round and convert him into a linebacker. That gets him a guy who can cover, and who's also good at tackling. Bryan Scott is like this, and the Bills have already made plans to move him to a sort of safety/linebacker hybrid position. It wouldn't shock me if they asked him to add a few pounds, and moved him completely to linebacker after he did so. Option 1 is to use lower round picks to obtain good, solid linebackers to fill in any holes which may exist in the Bills' linebacking corps. Option 2 is to use the 10th overall pick to upgrade one of the LB positions from that envisioned by option 1. The question is: does option 2 represent the highest value use for the Bills' first round pick? I don't see how it possibly could. The lower round linebackers Wannestedt is likely to covet will provide pretty good zone coverage already. It's hard to see how Kuechley's zone coverage will be enough better than that for him to represent a bigger upgrade to the team than, say, a solid LT, #1 WR, #1 CB, or some other players whose contribution to pass offense or defense is both obvious and substantial.