How valuable are rookie draft picks in a dynasty fantasy football leagues? In these leagues, team managers keep most of their rosters year after year but can draft rookies to replace aging players and fill team needs. Managers trade picks and players frequently, trying to improve their team and gain an edge over the opposition.1 Doing this effectively requires having a good sense of how valuable these assets are, however.

While managers often have a sense of how good players already on NFL teams actually are, rookie draft picks are harder to value. Each one is a black box, and like Brad Pitt in the movie Se7en, we want to know what’s in the box.2 On average, how good of a player should we expect to draft for any particular rookie pick?

This problem is not unique to fantasy. In the NFL and other sports leagues, many have applied quantitative modeling methods to assign an expected value to draft picks to gauge how much of a return is associated with drafting in particular spots.3 The common logic underlying these efforts is that we can examine relevant outcomes for players drafted in each position, averaged over many such players to account for the high variability.

Surprisingly, after searching for similar methods applied to dynasty draft picks, I wasn’t able to turn up much. So I decided to construct my own valuation method.

While I explain the methodology in more depth below, the summarized version is that I use six years of draft data to compare the average draft position (ADP) of rookies and their future dynasty value three years later, producing an Expected Future Value (EFV) for each pick. In this post (Part 1), I begin by constructing these values for different positions separately, using the player’s future ranking (WR1, WR2, etc.) among their position group as the outcome. This allows me to provide some position-specific drafting advice. In Part 2, which I plan on posting soon, I aggregate these position-specific EFVs to a single value, develop a method to determine how many of one type of rookie draft pick equals another, and compare these values to the perceptions of fantasy managers (using information from a crowd-sourced fantasy trade calculator) to demonstrate how these assets are being misvalued.

The post below describes the data used, the methodology, provides plots and a table showing the valuations, and discusses some implications for dynasty strategy. Let’s dive in.

The Data

Expected future values for each dynasty rookie draft pick are created using two pieces of information about the rookies selected, drawn from free and publicly-available data:

  1. At what draft pick (on average) was the player chosen in their rookie draft?
  2. At what draft pick (on average) is the player chosen in dynasty startup drafts in a later year?

To determine which players are chosen at each draft pick, I use rookie draft average draft position (ADP) data from FantasyCalculator.com. These are derived from hundreds of mock draft selections on the Fantasy Football Calculator website, and each player is given an average position in which they were drafted. For example, in 2025 Omarion Hampton’s average draft position is 2.5 (meaning he is usually being selected with either the 2nd or the 3rd pick).

So possessing the 1.02 or 1.03 pick means getting a player like Omarion Hampton. What we want to know is what the return on that pick is, or how valuable such a player becomes. To answer that question, I look at the player’s value in a startup dynasty draft after three years in the NFL.

Why three years? Three years gives players time to establish a record on the field and work through a learning curve. Some positions, like tight end, often take 2-3 years to reach their full potential (e.g., Trey McBride). In rare cases a player might need more than three years to establish their value (for example, Jordan Love played only a single game in his first three seasons with the Packers). But in most cases, after three years in the league we have a pretty good idea of how valuable a fantasy player is going to be. And looking even further down the line would mean shorter-career positions (like running back) might start having declines priced in, plus it would mean having less usable data (we would have to leave out the 2022 rookies, or more). For these reasons three years is like Baby Bear’s porridge: not too hot, not too cold, but just right in terms of having a good balance between enough information versus enough data.

To obtain dynasty startup draft ADPs, I use data from FantasyPros.com, which is sourced from Sleeper and Dynasty League Football and is available from 2017 to 2024. Because FantasyPros does not yet have 2025 dynasty startup data available yet (as of 5/13/2025), for 2025 dynasty startup ADP data I use DraftShark.com.4

After merging data from all of these sources together, we have information on 353 rookies drafted in fantasy leagues between 2017 and 2022, and their dynasty value three years later, giving us a solid data foundation for properly valuing draft picks.

Modeling Expected Future Value by Draft Pick

The method consists of modeling the expected (or average) future value of each draft pick. The analysis is based on a standard 12 team dynasty league with a three round draft, meaning there are 36 draft picks from the 1.01 to the 3.12. For players whose ADP is in each of these positions, we want to know what their future value is.

The simplified, plain English version of what I’m doing: I’m using a statistical method that finds the relationship between where a player was drafted as a rookie and how valuable they became three years later. Imagine a scatterplot showing this relationship. Putting rookie ADP on the x-axis and future draft value on the y-axis, we want to draw a (possibly curved) line that represents the best prediction at any point on the line, an average based on players drafted in similar spots. In the end, this approach gives us a reasonable estimate of what you might expect a 1st, 2nd, or 3rd round rookie pick to be worth three years down the road – while acknowledging that individual players will certainly fall above or below this expected value.

For the more technically inclined: Any number of regression or machine learning methods could be used to model the expected future value at any given draft pickm but ultimately I ended up using a linear regression with natural splines. In choosing a specific method, I was driven by a few considerations:

  1. I chose to model value separately for each of the four main fantasy positions: RB, WR, QB, and TE. If it is easier to project future performance for some positions than others, the expected value might differ quite a bit at a particular draft pick depending on which type of player is chosen.5 Because this is useful information, I wanted to model this explicitly.

  2. The above decision comes at a cost. Instead of fitting one predictive model, we’ll have to fit four. And there are more data points for some positions than others. For example, while the datset includes 144 rookie WRs and 134 rookie RBs, there are only 40 rookie QBs and just 32 rookie TEs.

  3. The final complicating factor is that the relationship between draft position and future value is very noisy. Especially for some positions (spoiler alert: QBs), there are plenty of cases where high draft picks are complete misses (e.g. Johnny Manziel with an ADP of 4.8 or Trey Lance at 6.8) and lower draft picks are massive hits (e.g., Josh Allen at 28 or Carson Wentz at 27.9).

Because of point 1 above, Expected Future Value is defined here as the player’s rank among their position group in dynasty startup drafts three years later (e.g. the highest drafted QB is QB1, the next highest-drafted is QB2, and so on).6 The modeling tool that proved to work best out of the onese I tried (as judged by the mean-squared error) was a linear regression model with natural splines to account for nonlinearities. For the positions with few data points (QB and TE) including a single knot produced the best predictions without showing signs of overfitting (such as suggesting later picks were more valuable than earlier picks in some cases). For the positions with more data points (WR and RB), natural splines with three knots gave the best fit. In R, these models can be estimated using the splines package as:

\[\texttt{lm(FuturePositionRank ~ ns(RookieADP, knots = k), data = rookie_df)}\]

where \(k = 3\) for RBs and WRs and \(k = 1\) for QBs and TEs. These are relatively simple models, but given the small amount of data we have to work with, they should be the most robust.

Analysis

The resulting models for each position are plotted below in Figure 1. What do the models tell us?

**Fig. 1. Expected Future Value by Draft Pick** <br> *Note: small dark points show the player rookie ADP (x-axis) and dynasty startup ADP three years later (y-axis). Large red circles show binned averages for every 6 draft picks. Blue lines show the model fit, while gray-shaded band displays the 95% confidence interval.*

Fig. 1. Expected Future Value by Draft Pick
Note: small dark points show the player rookie ADP (x-axis) and dynasty startup ADP three years later (y-axis). Large red circles show binned averages for every 6 draft picks. Blue lines show the model fit, while gray-shaded band displays the 95% confidence interval.

First, WR and RB rookie picks show similar patterns in expected future value: a steep decrease across picks 1-18, followed by a more gradual decline for picks 19-36. A RB drafted in the first half of the first round of a rookie draft will on average give you a top 20 to a top 34 RB three years from now, though some will be higher-value and some will be near worthless.7 Similarly, a WR drafted in the first half of the first round of a rookie draft will on average give you a top 25 to top 36 WR three years from now, though again there are several higher hits and several complete misses. In both cases getting a player whose expected future value is at least top 48 at their position requires a first round or early second round pick.

Second, drafting rookie QBs is very hit or miss. The line showing the model fit is relatively flat, meaning that there isn’t much additional value to drafting a QB in the first round versus drafting one in the second or third rounds. Three out of the five highest value QBs had an ADP that put them in the first half of the third round, while there are many QBs drafted with a top 18 dynasty rookie pick that are not NFL starters by their 3rd season. Even when using one of the top picks, the expected future value is less than a top-16 QB (a better than average NFL starter).

Third, rookie TEs are relatively predictable to draft. The line showing the model fit is much steeper than it is for QBs. Of the 5 TEs whose rookie ADP put them in the first round of drafts, 4 out of the 5 (80%) were an average to above-average TE after 3 years. On the other hand, no rookie TE with an ADP of 24 or higher became a top 12 TE, meaning if you want a TE that gives you an edge over your opponents in a 12 team league, you need to invest at least a second round rookie dynasty pick. A caveat: the data here are fairly sparse, with only a single TE taken with a top six pick. Projecting forward a bit from data outside of our window, Dalton Kincaid (rookie ADP of 8.5 in 2023) seems likely to be a mid-tier TE by his fourth season, while Brock Bowers (rookie ADP of 5.0 in 2024) has a good chance of being a top-tier TE by his fourth season.

These conclusions are just scratching the surface, but hopefully this illustrates how useful having these EFVs can be. To make it easy for others to reference the EFVs associated with different picks, Table 1 below shows the Expected Future Value for each pick for each position.

Implications for Dynasty Strategy

What are the takeaways of this analysis for drafting rookies and managing a dynasty team? Three obvious ones stand out to me:

  1. Invest Early Picks in RBs, WRs, and TEs (not QBs) – Three of the four positions show fairly steep declines in expected future value across the first 18 draft picks, meaning an early pick has substantially higher returns. On the other hand, QBs are hard to draft with confidence, with examples across nearly the entire span of ADP that are worthless three years later.8 Despite this, quarterbacks with ADPs as low as 28 have become top-tier starting options at QB, meaning there can be relative value in picking QBs in the late second/early third round.
  2. Mid to Late Third Round Picks Are (Nearly) Worthless – There are almost no examples of players after pick 30 returning a high value. The best outcomes, out of 63 players taken in this range, are the WR Jarvis Landry (rookie ADP of 30.8, WR20 three years later) and RB Isaiah Crowell (rookie ADP of 32.8, RB18 three years later). No top 12 TEs or top 18 QBs came out of these picks. While the trade value of a third round pick on its own is minimal, if adding a third round draft pick to a trade deal you otherwise like can make it happen, the analysis here implies you shouldn’t hesitate to give up the pick.
  3. Expected Future Values Can Inform Trade Decisions – Especially if you have a particular team need in mind, these expected future values provide valuable information to guide your decisionmaking. As an example, I had someone in a superflex (2 quarterback) dynasty league ask me about a proposed trade. The fantasy manager had one of their quarterbacks retire and had no one on their roster to replace them. After talking to another manager with a surplus of QBs on their roster, they were offered Kyler Murray for their 1.05 pick and D.J. Moore. My initial instinct (supported by crowd-sourced online trade calculators) was that this asking price was too high for a QB that has underwhelmed in recent seasons. On the other hand, the model suggests that the chance of drafting a QB at a similar tier (with an ADP of QB13 in 2025 startup dynasty drafts) is unlikely even if you had the 1.01 pick. For a QB-needy team, the asking price to move up from a draft pick with an expected future value of QB20 or QB21 (a Jared Goff or Anthony Richardson tier-player) was a WR whose 2025 startup ADP is flex-level (WR25).9 Given this context, the trade actually makes a lot of sense.

Conclusion

This write-up proposes a method for determining the Expected Future Value (EFV) of dynasty rookie draft picks. Using data from six years of dynasty rookie drafts, I estimate the EFV for each pick position and each of the four main player positions. This gives us a data-driven approach to valuing draft picks. The analysis provides some practical advice too: draft QBs in the middle of your draft and other positions early, be generous trading other managers third-round picks.

Just like the real NFL draft, there’s too much to pack into a single day.10 In Part 2 of this series on valuing dynasty rookie draft picks, which I plan on sharing in the near future, I build on Part 1 by:

  1. Developing a way to aggregate draft pick value across all four positions based on how frequently picks in any given range are used to draft a particular position.
  2. Using probability theory to construct a method of valuing rookie draft picks relative to each other (e.g., how many mid-tier 2nd round picks equal a mid-tier 1st round pick?).
  3. Showing that dynasty fantasy managers systematically mis-value draft picks, using values implied by a crowd-sourced fantasy trade calculator.

Stay tuned…

  1. At least, they do if you’re in a fun dynasty league.↩︎

  2. Hopefully things turn out better for us than they did for Mr. Pitt’s hotheaded detective.↩︎

  3. For a few examples, see here, here, and here.↩︎

  4. All draft ADP data is for single QB leagues. The DraftShark data is based on a 0.5 PPR scoring format. For the FantasyCalculator and FantasyPros data, they do not describe what scoring format the ADP data is based on.↩︎

  5. This is exactly what I find when I perform my analysis.↩︎

  6. In cases where the player went entirely undrafted three years later, I assigned a draft value of 33 for QBs and TEs and 65 for RBs and WRs. Any ADPs below these thresholds were given this draft value as well.↩︎

  7. Out of the 29 RBs whose ADP was 6 or below, 6 became a top six RB and 8 had became a RB 48 or lower.↩︎

  8. As a Chicago Bears fan, I nod along knowingly as I write this point out.↩︎

  9. Of course, this assumes that there’s a QB whose ADP is around this level that can actually be taken at the 1.05 pick.↩︎

  10. Or in this case, a single post.↩︎