The first three rounds: how tier gaps decide your draft
The first 36 picks fall into four groups split by three gaps. Which group your slot lands in, and why round three is where the room stops agreeing.
By The Parlae deskPublished 5 min read
Four groups, three gaps
The first 36 picks on the September 3 board do not descend smoothly. They come in four groups, and the gaps between the groups are the largest on the board's first three rounds. The tier breaks page lists them: Jonathan Taylor at 7.4 to De'Von Achane at 9.9, a 2.5 pick gap. Drake London at 13.4 to Derrick Henry at 16.0, 2.6 picks. Omarion Hampton at 22.2 to Zay Flowers at 25.6, 3.4 picks. For scale, the typical distance between neighbours in the top 100 is 0.8 picks.
| Group | Picks | Players | Count | Deviation range |
|---|---|---|---|---|
| A | 1.5 to 7.4 | Gibbs, Robinson, Nacua, Chase, Smith-Njigba, McCaffrey, St. Brown, Taylor | 8 | 0.7 to 1.7 |
| B | 9.9 to 13.4 | Achane, Lamb, Cook, Jefferson, C. Brown, London | 6 | 1.7 to 2.7 |
| C | 16.0 to 22.2 | Henry, Rice, A.J. Brown, Barkley, Olave, Pickens, Jeanty, Walker, Collins, Hampton | 10 | 2.3 to 4.2 |
| D | 25.6 to 36.2 | Flowers, Nabers, G. Wilson, Love, McBride, D. Smith, K. Williams, McMillan, Hall, Allen, Bowers, J. Williams, Egbuka, Higgins | 14 | 3.3 to 8.0 |
Source: Parlae board, Sep 3 2026, PPR.
Group A is eight players inside 5.9 picks. Group B is six inside 3.5. Group C is ten inside 6.2. Group D is fourteen inside 10.6, and it is the first group where the room takes a tight end and a quarterback.
The gaps land on the round boundaries
In a 12 team league round one ends at pick 12, round two at 24, round three at 36. Look at where the groups end. Group B's last player, London, goes at 13.4. Group C's last player, Hampton, goes at 22.2, and the 3.4 pick shelf under him is the largest on the board. Group D runs to Higgins at 36.2.
That means the tiers and the rounds nearly coincide, which is rare and which makes the first three rounds unusually easy to plan. Round one is Groups A and B. Round two is Group C, with ten players for twelve picks, so two teams reach into Group D before the round ends. Round three is Group D. Your slot decides which group your first pick comes from, and after that the groups decide for you.
What each slot gets
Picks one to four are inside Group A with room to spare. The room prices Gibbs at 1.5, Robinson at 2.4, Nacua at 2.8 and Chase at 3.9 with deviations under 1.0, which is as close to unanimous as a draft market gets. The pick is not a decision; it is a queue.
Picks five to eight are the end of Group A: Smith-Njigba at 5.5, McCaffrey at 6.4, St. Brown at 6.6, Taylor at 7.4. The 2.5 pick shelf under Taylor means the drafter at pick eight is the last one to get a Group A player, and the drafter at pick nine is the first to start Group B.
Picks nine to twelve are Group B: Achane at 9.9, Lamb at 10.3, Cook at 11.5, Jefferson at 11.7, C. Brown at 13.0, London at 13.4. Six players for four picks, and the pick 12 drafter picks again at 13, so the turn takes two of them.
Round two is Group C for everyone, and it is the round where the position split is cleanest: five running backs and five receivers. Whichever position you took in round one, the other one is available in round two at the same tier.
Round three is where the room stops agreeing
The deviation column is the story of round three. Group A players have deviations from 0.7 to 1.7. Group C runs from 2.3 to 4.2. Group D runs from 3.3 to 8.0, and three players in it have gone more than 20 picks apart across the sample: Josh Allen from 12 to 48, Brock Bowers from 15 to 50, Javonte Williams from 22 to 49.
This is also the round where the positions split. Group D holds seven receivers, four running backs, two tight ends and one quarterback. Through pick 24 the board is only backs and receivers. At 29.1 the first tight end goes, at 33.5 the first quarterback. The round three drafter is the first person in the draft who has to decide whether a positional advantage at tight end or quarterback is worth a Group D receiver, and the deviation column says that different rooms answer that differently every night.
For a drafter, the width of Group D is an opportunity. Fourteen players inside 10.6 picks, with deviations that wide, means that in most rooms at least two of them go past their average. The drafter at the turn, picking 24 and 25, will usually find that the last Group C player is gone and that Group D is fully stocked. That is the best turn on the board.
The decision
Rounds one and two are set by your slot and the groups, and the only mistake available is to reach across a shelf. Do not take a Group B player at pick six because you like him; the room has Group A players priced two picks better, and the shelf is real. Do not take a Group D player at pick 20; Group C still has players.
Round three is the first real decision, and the board makes it clearer than it looks. If you have a back and a receiver, Group D offers four more backs and seven more receivers with tight deviations, and a tight end and a quarterback with wide ones. Take the position where the tier is about to end. On this board the back tier ends at Javonte Williams at 35.4, and the receiver tier keeps going to Tee Higgins at 36.2 and then Ladd McConkey at 38.4, so the back is the one that runs out first.
What to do with this
- Before the draft, open the tier breaks page and write down the three gaps with the players on either side. Your slot puts you in one group in round one; know which.
- In round two, take the position you did not take in round one, from Group C. Ten players are priced inside 6.2 picks, and the room does not distinguish between them.
- Run your slot three times in the mock draft tool and note who is left at your third pick each time. If a Group D player falls to you every time, that is your round three plan; if a different one falls each time, take the running back, because that tier ends first.
Sources
- Parlae board and tier breaks page, Sep 3 2026 (parlae.io/fantasy/rankings, parlae.io/fantasy/tier-breaks, updated 8:23 PM ET)
- Fantasy Football Calculator public API, PPR feed, Sep 3 2026 (7,848 drafts, Aug 27 to Sep 3 window: average, earliest, latest, standard deviation)
- Parlae tier shelf rule (src/lib/fantasy/tier-shelves.ts): a shelf is a gap of at least 3 picks and at least twice the median neighbour gap
By the Parlae deskGuide updated Sep 4, 2026 at 8:00 AM ETHow this is builtReport a wrong number