clock menu more-arrow no yes mobile

Filed under:

Playing the NFL Draft board using probability

ESPN provides a different lens through which to look at draft strategy

NFL: NFL Draft Gary Vasquez-USA TODAY Sports

The wait is almost over, New York Giants fans. Soon mock drafts will be replaced by the real thing. But until then...

Last week I wrote about 10 mock drafts I conducted to get a feel for when the top four cornerback prospects and wide receiver prospects are most likely going to come off the board and what the Giants’ chances were of having one drop to them at No. 25, where they make their first pick. That was a poor man’s attempt to bring some probability thinking into the subject.

Then after a more recent piece about testing different draft philosophies appeared on these pages, ClayfromBklyn asked me:

You didn’t mention any of the supposed Top Centers (in any of these drafts), so I’m left to wonder when they were taken? Were Tippmann, Schmitz, Oluwatimi & Wypler all taken in between 26 and 56? Or perhaps a couple went there, and a couple went between 58 & 88? In any case, Patterson isn’t nearly good enough to be considered a Day 1 starter at Center, IMO, and that’s the biggest priority for this draft to me. So if any of those 4 were on the board anytime after 25, that would’ve been my pick.

This is a great point, and Clay is exactly right: The top four centers never fell into my best available or greatest need or highest positional value groups when the Giants drafted at No. 25 or at No. 57, but rather were taken in between so I lost my chances at all of them.

I should have waited a few days, because ESPN Analytics has just gone live with their own mock draft simulator. A unique feature of this simulator is that it shows you the statistical chances of particular prospects being taken at specific points in the draft as well as their chances of being available at particular points in the draft. So let’s have some fun (if you’re a nerd like me).

The ESPN Big Board

Any mock draft simulator is only as good as its big board of player rankings is, and we have no idea how good that is, because different people evaluate prospects differently. The actual draft is the result of 32 general managers butting heads, each of them with their own big board that neither we nor their competitors know about (unless someone like Jerry Jones does this at his post-draft presser):

ESPN’s big board has some rankings that may seem unusual at first. Here are their top cornerback prospects:

Data courtesy of ESPN Analytics

Emmanuel Forbes as CB3 and No. 22 overall. Julius Brents and DJ Turner as borderline first/second rounders. And Riley Moss, rarely if ever in the discussion, as low Round 2. As Ed Valentine discusses, though, these aren’t out of line with what some people are saying now, so the virtue of the ESPN big board being released late in the game may be that it reflects more recent trends in player evaluation.

Wide receiver is just as interesting:

Data courtesy of ESPN Analytics

Quentin Johnston barely in the first round, and Jonathan Mingo an early Day 2 prospect, with he and Jayden Reed ahead of Jalin Hyatt. Marvin Mims higher than I would have expected, too.

Finally, the centers (and guards, lumped together):

Data from ESPN Analytics

Steve Avila as the best interior offensive line prospect is unexpected, but Joe Tippmann and John Michael Schmitz are close to each other in the 40s, in between the Giants first- and second-round draft positions.

Playing the odds on the board to get your guys

I’m not going to present another mock draft in this article. Instead, I’m going to use the unique feature of the ESPN Analytics simulator to think about individual draft picks from the standpoint of probability. Here are two questions we can ask of the simulator at picks No. 25 and No. 57:

What players are most likely to be available to the Giants in the first round?

The simulator gives us that information in two different ways - the chance that the Giants take a specific player there, and the chance that different players are available there:

Data courtesy pf ESPN Analytics

Let’s focus on cornerback and wide receiver, the two most likely priorities for the Giants on Day 1 if the draft falls to them favorably. ESPN suggests (top figure) that CB Deonte Banks and WR Jordan Addison are the Giants’ two most likely picks, i.e., most likely to be chosen if they are available. But the bottom figure indicates that they think there is roughly a 50% chance that Addison will be gone and a 60% chance Banks will be off the board.

Those are not completely independent events, since earlier picks have some effect on later picks, but assuming they are independent just to illustrate, that would imply only a 30% chance (.5 x .6) that both are gone at No. 25. Expanding our view, there is about a 65% chance that Zay Flowers is gone and only about a 25% chance that Emmanuel Forbes is gone at No. 25. Throw in the more controversial prospect Quentin Johnston, with only a 40% chance of being gone. So whatever the actual degree of dependency of one pick on a later pick, the bottom line is that ESPN sees the Giants as having a pretty good chance to grab one of the top cornerbacks or wide receivers in Round 1. Of course it all depends on what the Giants’ big board looks like - whether they would be happy with any one of those five players, we don’t know.

If things don’t work out for either position at No. 25, here are the odds of various cornerbacks and wide receivers being available when the Giants next pick at No. 57:

Data courtesy of ESPN Analytics
Data courtesy of ESPN Analytics

This goes against the conventional wisdom, which said that the draft was deep in cornerbacks but shallow in wide receivers, so a team like the Giants can wait on cornerback but should take wide receiver at the first opportunity. What has happened is that in recent weeks the stock of a number of cornerbacks has risen dramatically (Forbes, Brents, Turner), as can be seen in the cornerback big board earlier in this article. Suddenly, players assumed to be available late in Round 2 are not anymore.

On the other hand, some wide receivers are rising on big boards (Jonathan Mingo, Jayden Reed, Cedric Tillman) and are now considered viable draft targets in Round 2.

When do the Giants have to make a move to draft a center?

Here is the answer to Clay’s question, examined in some detail. The ESPN simulator won’t even display the available centers at pick No. 25, meaning they see zero chance of any of them being chosen that early. You have to get to pick No. 32 before they even display John Michael Schmitz and Steve Avila (who started as a center before moving to guard) as having about a 1% chance of being taken, and only at pick No. 37 does Joe Tippmann appear also:

Data courtesy of ESPN Analytics

The problem is that by the time we get to pick No. 57, the odds look like this:

Data courtesy of ESPN Analytics

Now things are getting uncomfortable. The chances of each of these players individually being available at No. 57 is in the 20-37% range. The probability of all three being gone is not that big, something like one in three...given THIS big board and draft model. So in the words of Dirty Harry, “You’ve got to ask yourself one question: Do I feel lucky?”

Depending on how they have center prospects rated, this gives the Giants four possible outcomes, depending on what happened in Round 1:

  1. Take their chances that one or more of the center prospects that they like is still there at No. 57 and plan to draft a center at No. 89 if not.
  2. Draft a center in Round 1, going against ideas of not reaching to fill a need, if all their favorite cornerbacks and wide receivers are gone and they don’t value taking a player at another position, e.g., defensive tackle, instead.
  3. Trade down from No. 25 if none of their cornerback or wide receiver favorites is still on the board at that position and try to get into better position to take a center at a draft position that makes sense.
  4. Trade up from No. 57 to get into the same position.

How far could the Giants trade down and still have a good chance to grab a center they like? At No. 47, all three players have a 50% or better chance of still being on the board:

Data courtesy of ESPN Analytics

Unfortunately Washington has pick No. 47, so they might not want to do business with the Giants, although you never know. Let’s aim for No. 46 (New England). A trade down with the Patriots from No. 25 might look like this, based on the great new NFL Trade Chart Calculator shiny app by @JosephJefe:

Data courtesy of NFL Trade Chart Calculator

The Giants give up No. 25 and get back No. 46, No. 135, and a mid-7th round pick in 2024 from the Patriots. The chart above shows who “wins” the trade in green, using six different trade value charts. Three say the Patriots, three say the Giants, which I interpret to be a fair trade. (That said, the more modern ones say the Giants win.)

A trade up from No. 57 with New England might go this way:

Data courtesy of NFL Trade Chart Calculator

The Giants move up to No. 46 while keeping their Round 1 pick, giving up No. 57, No. 209, and a late (because the Giants are going to the playoffs again in 2023) sixth-round pick to the Patriots. Either way, trade down or trade up, the Giants have a good chance to draft Schmitz or Tippmann.

If the Giants decide to roll the dice on No. 57 instead and it comes up snake eyes, then if they are satisfied with Luke Wypler at No. 89, he is likely to be there according to ESPN. In fact Wypler isn’t even displayed in their graphic until pick No. 96, with an 80% chance of being available:

Data courtesy of ESPN Analytics

Barring that, prospects such as Olu Oluwatimi and Ricky Stromberg come into play, more likely at pick No. 128.

The bottom line

We have no idea how accurate the ESPN big board or draft model are. We have to assume, though, that deep in the bowels of 1925 Giants Drive, the Giants’ analytics department is doing calculations similar to these to inform the strategy of GM Joe Schoen later this week when he has to make decisions in real time. It’s a lot more complicated being down at pick No. 25 than it was last year at picks No. 5 and 7, when Schoen had his list of seven most desired players, two of whom he could be certain to get.

In the meantime, though, the ESPN simulator is a fun way to think about the possibilities. It will be even more fun if it is updated in real time as picks are made this weekend, allowing us to view the changing odds of the Giants acquiring the players each of us longs to see in blue.