Less than a week until the draft begins, and it’s clear that nobody has any clue what’s going to happen at the top of the board (aside from Joeckel, who seems like a decent bet to go #1). Although it’s fun reading supposed “intelligence”, most of its useless and will bear no resemblance to what the reality will be, so don’t get too worked up over things like “sources say the Eagles really like Geno Smith”.
Also, the full NFL schedule was released yesterday. Go have a look if you’re interested. Personally, I don’t think it’s worth that much attention at this stage of the off-season, though playing the Redskins week 1 is nice (likely before RG3 is up to speed).
So what to talk about?
Well over the past few weeks I’ve advanced a number of different ways to view/think about the NFL draft. Today I think I’ve got the clearest illustration yet of what I believe is the “best” way to look at a draft board.
The general idea is that assigning prospects exact values on a 0-100 scale (or any other scale) is ridiculous. It’s foolish to think that any scout or group of scouts can accurately grade NFL prospects with such precision. Conversely, the other method of ranking prospects seems to be a non-numerical hierarchy. That too is useless. It’s not enough to know that Player A is ranked above Player B.
We need to know how big of a gap there is! Otherwise, we have no concept of true value as we go through the draft.
So what can we do better?
We can, rather than assigning specific values, look at VALUE RANGES. Below, I’ve included a few charts that depict the top 20 prospects and assigns each a range of expected values. To come up with the ranges, I simply took the average ratings of NFP, ESPN, and NFL.com (our Consensus Ratings from the TPR system) then used the standard deviation for each prospect to create the range (+- 1 SD).
A few notes: I wish we had a lot more data (ratings). This does not make any adjustment for positional value, it is strictly a depiction of each prospects individual ratings. In practice, each team would incorporate their own rankings and adjust for things like positional value, player personality, fit, etc…
Now let’s look at the charts:
Above are the top 11 prospects (by average rating). As we can clearly see, according to our data, some of these players carry much wider potential ranges than others. There is no “correct” way to read this graph, it all depends on your personal preference.
For example, if we are just looking for “impact” than we would focus on the HIGH end of the ranges. Doing this, we can see that Sharrif Floyd may turn out to be the best player in the draft outside of Luke Joeckel, despite the fact that he’s ranked just 7th overall.
Or, if we are risk averse, we can see exactly why Chance Warmack is such an interesting prospect. Despite playing a positional of relatively little importance, he’s clearly the “least risky” prospect, whose “downside” appears to still be extremely good.
Regardless of what your preference or risk tolerance is, it’s should be pretty clear that a discrete ranking or prospects (i.e. a “big board”) is of little practical value. It’s vital to get a sense of both the potential downside and upside of each prospect, and to EXAMINE THE OVERLAPS.
Dion Jordan and Eric Fisher carry almost identical consensus ratings, but Fisher has a wider range of potential values. Is Fisher’s downside trade-off worth his upside potential? Or would you rather take Jordan, who carries a narrower range of values (but necessarily gives up some upside)?
Here are prospects 12-20:
By average rating, Cordarrelle Patterson is rated higher than Tyler Eifert, but it’s pretty clear you can make the case for Eifert being the better prospect.
Finally, here is a chart of all 20 prospects shown together. I split them up above to make it easier to see.
If I can convince you of just one thing going into the draft, it’s this. False precision is incredibly detrimental to value in the NFL draft. Let’s say you are agnostic as far as positional need, would you rather take Dion Jordan at #4 or slide down and take Star Lotulelei a bit later (assuming you own board looks like the one above). The “value” in that situation is pretty clearly on the side of moving down.
Putting a hierarchical ranking together and just selecting the best available is a very poor way to derive value in the draft.