The Fantasy Football Philosopher: Can You Win Your League at the Draft?
In this series, we tackle the ‘philosophical’ questions of fantasy football. Today, we’ll be asking a simple, popular pre-season query: can you win your league at the draft?
Most of us can agree that draft day is the best day of the fantasy season, and there’s no feeling like walking away from your draft knowing that you nailed it. When you do perform well in the draft, is it reasonable to begin to dream of fantasy glory?
If you listen to the Fantasy Footballers Podcast, you know that Andy, Mike, and Jason like to break up the season into ‘chunks’. The start of the season is, naturally, the ‘draft chunk’: as one might expect, your performance in the first few weeks will largely depend on how the draft went. As the season progresses and things start to shake up a bit (players emerge, trades happen in your league, other players get injured or disappoint) the draft becomes less important, although it still represents a season-long foundation for your team.
From there, a natural question arises: what is the nature of this ‘draft chunk’? Can you really expect to dominate early on following a great draft? Will a top-notch draft, in fact, carry you all the way to a championship, or will you only be able to coast on your draft day performance for a few weeks?
All data, unless otherwise specified, comes from nflfastR and, for ADP values, from the fantasy football calculator.
Methodology
We need to develop a sense of a ‘great draft’, a ‘terrible draft’, and everything in between. There’s actually a relatively simple way to do this: using real ADP data from 2015 to 2020, we can generate a non-snake draft.
As you’re likely aware, drafts will ‘snake’ – reverse order at the first and last picks of a round – so that managers with later picks don’t get the short end of the stick. However, imagine a draft that doesn’t snake: the manager with the first pick also has the 13th pick, the 25th pick, etc., while the manager with the last pick (in a twelve-team league) has the 12th pick, the 24th pick, the 36th pick, etc. That is, the manager picking first has a huge advantage.
Using this paradigm, we can generate lineups for teams drafting 1-12. To simplify things, we’ll just be picking starters for a QB/2RB/3WR/TE roster, and each team will cycle through the positions together (no hoarding one position / trying to stack good RBs, etc.). So, for example, the manager with the first pick will have, according to redraft rankings, the QB1, RB1 and RB13, WR1, WR13 and WR25, and TE1. This year, that lineup would be something like this:
QB: Patrick Mahomes
RB: Chris Carson
WR: Davante Adams
WR: Tyler Lockett
WR: Robby Anderson
TE: Travis Kelce
To say this is a dominant squad would be putting it lightly. CMC is the consensus 1.01, while Kelce and Adams are also first-round picks. Pat Mahomes is no slouch either, going early in the second round. Naturally, this is all a bit unrealistic: it’s hard to imagine getting three 1st round picks (and a high 2nd) on one roster during the draft.
On the flip side, the luckless manager with the last pick would have the QB12, RB12 and RB24, WR12, WR24 and WR36, and TE12. This sad roster would be something like this:
QB: Taysom Hill
RB: Antonio Gibson
RB: Miles Sanders
WR: Terry McLaurin
WR: Ja’Marr Chase
WR: Curtis Samuel
TE: Irv Smith Jr.
This lineup is also unlikely to occur; the best player, Antonio Gibson, is a mid-second round pick, and Miles Sanders / Terry McClaurin can be scooped in the third or fourth round. Both Taysom Hill and Irv Smith Jr. are squarely in late-round flier territory; Ja’Marr Chase and Curtis Samuel aren’t far behind.
Anyways, our goal is to see how these lineups have historically performed. As mentioned, these two rosters above are a little bit ridiculous – it’s hard to imagine walking away from a draft with either of these options. What interests us more is the drafts in the middle…
Draft Chunk
We’ll separate our teams into categories: ‘Great’ drafts get a top-3 player at every position, ‘Good’ drafts get the 4th-6th best player at every position, ‘Bad’ drafts get the 7th-9th best player at every position, and ‘Terrible’ drafts get a bottom three player at every position. We can then get the average points each roster will score over the season:
- Right off the bat, it’s clear that Great and Good drafts score more than Bad and Terrible drafts (we’ll dive more into the nuance soon). This is good news: it means that ADP is reasonably accurate and that players projected to be ‘great’ or ‘good’ are usually better than players projected to be not as good.
- Unsurprisingly, all of these lines are sloping down over time. This is partly because of Bye weeks (I didn’t sit players who are on Bye, largely because it would make things too complicated, but also because it balances out across lineups) and partly because players will get injured, benched, or disappoint and thus become less fantasy relevant as the season progresses.
Let’s take a deeper look at the average roster points scored in the first four weeks of the season. There is a nice intuitive ‘staircase’: Great drafts score the most (by far), then Good drafts, and so on. The dotted line is just the average points here, which is skewed by how impressive the ‘Great’ drafts are.
We see – mostly – the same effect in points scored per player: ‘Great’ drafts have players scoring more than ‘Good’ drafts on average, etc.
Post-Draft Chunk
However, this effect is just for the four weeks right after your draft. Can we expect this sort of dominance to continue? Looking at the total roster points, it seems that, from Week 8 on, ‘Great’ and ‘Good’ drafts are still significantly better than ‘Bad’ and ‘Terrible’ drafts.
However, in terms of points per player, ‘Good’ drafts are now on par with ‘Bad’ and ‘Terrible’ drafts (‘Great’ drafts are still dominant). What’s happening here?
This is a question of sustainability. Active players from ‘Good’ drafts are scoring (roughly) the same amount as players from ‘Bad’ and ‘Terrible’ drafts from Week 8 onwards, but they are more durable, and thus ‘Good’ drafts have more players actually playing late into the season. This can be shown in the chart below, which gives the average number of players that actually posted stat lines (y-axis) vs. time (x-axis). As mentioned before, this number will decrease throughout the season due to injuries, benching, and Bye weeks. Interestingly, players from ‘Great’ drafts actually fall off pretty quickly; this could be a quirk of the data (or the 2018 Le’Veon Bell effect).
Anyways, the point here is that, as one might expect, players from ‘Good’ drafts are going to be fantasy relevant for longer than players from ‘Bad’ and ‘Terrible’ drafts; however, on a per-player basis (among players that are still relevant), they will be scoring a similar amount to players from ‘Bad’ and ‘Terrible’ drafts. The implication is that the advantage from a ‘Good’ draft will fade late in the season. Managers that didn’t draft well can be active in trades, on the waiver wire and with bench players that break out (and which aren’t discussed here). Therefore, even those with ‘Bad’ and ‘Terrible’ drafts will be able to replace the less-durable players in their lineup that aren’t active for a given week as the season progresses and make up the gap. Since managers with ‘Good’ drafts no longer have a point-per-player edge, the advantage will narrow and probably disappear.
Conclusion
As mentioned above, we really should be focusing on the ‘Good’ and ‘Bad’ drafts, since it’s actually feasible to get a ‘solid’ player (ranked 4th – 6th) at every position (or, vice versa, get a player ranked 7th – 9th at every position for a ‘Bad’ draft). ‘Great’ drafts are, sadly, a pipe dream; if you do manage to get a top-3 player in every slot, you can be confident that you are going to dominate all season long!
For just ‘Good’ drafts, though, we’ve seen a distinct scoring advantage – both in points per player and total roster points – early on in the season. This starts to fade later in the season: while players in ‘Good’ drafts are more durable, their per-player edge starts to go away, and active managers who had ‘Bad’ drafts will be able makeup ground via bench breakouts, waiver adds, and trades.
Put simply, the ‘draft chunk’ is legit; perform well on draft day and you should have a strong start to the season. However, as the Footballers emphasize, you can’t let your guard down: the draft is a foundation for the rest of the year, but your edge will dull if you don’t continue to put in the work.
______
Did I miss something? Want to philosophize about something else? Message me on Twitter.
One note: because of hashing problems – not having an ideal ID to link the fantasy points datasets with the ADP datasets – there will be slight hiccups in the data above (incorrect player getting selected, multiple players selected, etc.) but it likely doesn’t have an undue effect and should balance out across lineups.
Comments
Yes, certainly! This article is just looking into the specific nature of the ‘draft chunk’, but you’re certainly right that broader roster construction questions can be answered if we look deeper at each position. Thanks for the comment!
This is great, but the really interesting and actionable info would pertain not to the top-7 guys on the roster (or not only them), but to the backups. I.e., should you prioritize an RB2 or RB3 or WR4 over, e.g., the ADP 8 QB?