Why the Giants can NOT take super star running back Saquon Barkley at number 2.

The Giants have the #2 pick in this weeks 2018 NFL draft. The word they have leaked is that they are strongly considering taking all world running back Saquon Barkley with this pick. In my prior blog, I explained how they could be using Game theory to get the Browns to take Barkley at #1 leaving them Darnold, who I believe is the only quarterback they highly value. 
 
Here we will explain why it would be a very bad idea to take Barkley with the second overall selection.
 
Its all about The Relative Value of Positions
 
For any running back with more than a one year (Franchise) contract (meaning we are excluding the franchised Le’veon Bell for this), the highest paid in the NFL is Davonta Freeman at $8.25m.  By the time you get to 15th highest you are at $4m per year.
 
The Jags signed Blake Bortles for $18m -BLAKE BORTLES!!!!  He happens to be not much better than the 30th best quarterback but 18th highest paid quarterback.  This is largely as a result of some very good quarterbacks still on their rookie contracts. The highest paid QB is now Cousins at $28m per, and the 15th (average for a starter) gets just under $21m.
 
The salary cap is at about $177m.  To compete for the Superbowl you arguably need to produce significantly more value than that.  Let’s say approximately $235m in production (the exact amount could be argued but the concept can’t).
 
If you can draft a top performing quarterback and pay him less than $7m per year, if he performs at just the top half that comes to $21m, a savings of $14m or just above a quarter of the outperformance you will need, from one position.  Paying a running back the $7m that a #2 overall pick will make, if he performs at the very top level, you are still only saving approximately $2m. There is virtually no running back you couldn’t get if you would pay $9m per year. Walter Payton in his prime would likely not get more than $10-11m per year. You hope that Barkley is Walter Payton.
 
The relative value of these positions makes it that you can’t take a running back that high even if you think he will be a top 3 running back in the NFL.  You would have to know you were drafting Jim Brown for that to make sense and even then, the numbers likely would not support it in today’s passing league.
 
The math and replacement value formula makes the argument clear.  It is interesting to analyze whether the entire league has it wrong. Is it possible that the entire league just incorrectly evaluating the relative value of a good running back to a good quarterback?  It’s unclear if the answer matters as you could get just about any running back you want for $9m per year, and therefore even if the league is incorrectly calculating the position, you should still “arb” that in the free agent market and not in the draft.
 
The typical QB is an active part of an offensive play an average of 500 times per year (based on attempts by 15th ranked QB in attempts) versus average running back of 210 carries by same metric. Even accounting for only receptions and the QB can affect materially more plays.  Additionally, the variance around the average quarterbacks’ performance is substantially higher than the variance around a running backs production.  Firstly, there is more variance in yards per attempt between a 90th percentile (top tier) QB versus a 50th percentile one, than there is in yard per carry of similarly stratified running backs (you must assume a certain number of minimum carries when doing the math since unlike QBs some running backs only run very infrequently and their averages are not statistically significant).
Secondly, the impact of the much wider variance on turnovers makes the QB even more important.  The difference between lost fumbles for the most generous of ball carriers (again assume 95th percentile) on a given year versus the average fumbler may be about 1-1.5 lost fumbles per year.  For QB INTs that would be at least 6.  It could be argued that the average turnover would cost a team approximately 40 yards worth of field position.  When that is factored into any yards per play analysis, the impact of a turnover on total value differential is significant, once again making QB a much higher “leverage” position.
 
Seattle had a dominant team when they had Russell Wilson playing on his cheap 3rd round rookie contract. Their ability to win went down materially when they had to start paying him full price. Without a doubt, one of the keys to the Patriots success has been Brady’s willingness to work for what is known to be well below his intrinsic value (I don’t know what kind of deal he and Kraft have worked out for a personal services contract after he retires to compensate for this).
 
Lastly, many of the things they are now saying about Barkley, they were saying about Leonard Fournette. He is/will be a good running back, but he averaged 3.8 yards per carry. If Barkley came in and did well and became a free agent a year from now he would command $8-9m. If a Rosen or Darnold did the same they would command $25m (see Garofalo, Jimmy).
 
You sign a Rosen and if he is only average quality you will get more added value out of him than a top running back. The relative value of the positions is just too lopsided to take a running back that high.
 
So, what should the Giants do?
​
Ideally, they like a quarterback available to them at number 2. If they really do not like the quarterback options available to them (I suspect that they like Darnold who could go to the Browns at number 1), they should look to trade down (although the best deal they would have gotten and the trade I’d have liked was the trade the Jets made with the Colts).
Lastly, if neither is a possibility they should look at Bradley Chubb as their next best option.  It still doesn’t have QB positional value but top defensive ends make double what top running backs make and get a bit closer to the value of a QB.

-Ari Glass