Guide
Handcuffs and backfield committees by the numbers
Quick answer: Pull every running back with /players?position=RB (4 requests), group weekly snap shares by team, and look at the weeks each team's RB1 didn't play. In 2025 that happened 63 times, and the back who took over averaged 61% of snaps and 12.5 PPR. That's a starter's line. Committees are the other half of the story: 12 teams had no RB above 60% of snaps in at least 9 of their 17 games.
- Requests
- 4
- Plan
- Basic
- RB1-out games (2025)
- 63
- Fill-in average
- 61% / 12.5 PPR
Is a handcuff worth a roster spot in the fantasy playoffs?
Sometimes, and snap share tells you which times. A handcuff is only as good as the role he inherits. Some backups step into a 75–90% share and score like an RB1. Others split the work three ways and hand you 6 points. The difference is visible in the data before you need it, because every backfield has already shown you what it does when its starter sits.
This guide measures that directly. For every team we find the RB1 (the best average snap share among backs who played a meaningful number of games), list the weeks he missed while his team played, and record who led the backfield instead, with his snap share and PPR.
How do I measure every backfield with snap share?
One sweep of /players?position=RB returns every running back with his weekly snap rows and weekly stats for each season, so the whole league costs 4 requests. Pass the season as an argument:
import os, sys
from collections import defaultdict
import requests
API = "https://api.gridirondata.com/api/v1"
H = {"x-api-key": os.environ.get("GRIDIRON_API_KEY", "YOUR_KEY")}
SEASON = sys.argv[1] if len(sys.argv) > 1 else "2026"
ALIASES = {"LA": "LAR", "JAC": "JAX", "WSH": "WAS"} # snap rows vs stat rows
def all_rbs():
"""Every RB in one paginated sweep: about 4 requests."""
players, cursor = [], None
while True:
params = {"position": "RB", "limit": 100, **({"cursor": cursor} if cursor else {})}
r = requests.get(f"{API}/players", headers=H, params=params, timeout=60)
r.raise_for_status()
body = r.json()
players += body["players"]
cursor = body.get("next_cursor")
if not cursor:
return players
def ppr(row):
for k in ("fantasy_points_ppr", "points", "fantasy_points"):
if row and row.get(k) is not None:
return float(row[k])
return None
# team -> week -> [(snap_pct, name, ppr)]
backfield = defaultdict(lambda: defaultdict(list))
shares = defaultdict(list) # (team, name) -> weekly snap shares
for p in all_rbs():
block = p.get("seasons", {}).get(SEASON, {})
weekly = (block.get("snap_counts") or {}).get("weekly") or {}
stats = block.get("weekly_stats") or {}
season_pts = sum(ppr(r) or 0 for r in stats.values())
if weekly and season_pts / len(weekly) < 2:
continue # fullbacks and deep backups: snaps, no touches
for wk, row in weekly.items():
pct = row.get("offense_pct") or 0
if pct <= 0:
continue
team = ALIASES.get(row.get("team"), row.get("team"))
backfield[team][int(wk)].append((pct, p["player_name"], ppr(stats.get(wk))))
shares[(team, p["player_name"])].append(pct)
fill_ins = []
for team in sorted(backfield):
weeks = backfield[team]
# RB1: best average share among backs who played a quarter of the team's weeks
enough = [n for (t, n), v in shares.items() if t == team and len(v) >= max(2, len(weeks) // 4)]
rb1 = max(enough, key=lambda n: sum(shares[(team, n)]) / len(shares[(team, n)]))
rb1_weeks = {w: pct for w, rows in weeks.items() for pct, n, _ in rows if n == rb1}
committee = [w for w, rows in weeks.items() if max(rows)[0] < 60]
line = (f"{team}: RB1 {rb1} {sum(rb1_weeks.values()) / len(rb1_weeks):.0f}% avg, "
f"committee weeks {len(committee)}/{len(weeks)}")
for w in sorted(set(weeks) - set(rb1_weeks)): # team played, RB1 didn't
pct, name, pts = max(weeks[w])
fill_ins.append((pct, pts))
line += f"\n Wk {w}: RB1 out -> {name} {pct:.0f}% of snaps, {pts if pts is not None else '-'} PPR"
print(line)
if fill_ins:
pts = [p for _, p in fill_ins if p is not None]
print(f"\n{len(fill_ins)} RB1-out games: the fill-in averaged "
f"{sum(p for p, _ in fill_ins) / len(fill_ins):.0f}% of snaps and "
f"{sum(pts) / len(pts):.1f} PPR")
Three choices keep the output honest:
- Fullbacks and deep backups are dropped when they average under 2 PPR per game with snaps. Fullbacks log real snaps without touches, and counting them would make a workhorse look like part of a committee.
- RB1 is the best average share among backs who played at least a quarter of the team's weeks. Using total snaps instead would crown whoever filled in longest.
- An "RB1 out" week requires the team to have played. A bye isn't an absence. Team codes are normalized (
LAandLAR,WSHandWASappear in different rows).
What happened when the RB1 sat in 2025?
Here's an excerpt of python handcuffs.py 2025, run against the production API, with the league-wide summary line at the bottom:
KC: RB1 Isiah Pacheco 52% avg, committee weeks 10/17
Wk 9: RB1 out -> Kareem Hunt 78% of snaps, 12.5 PPR
Wk 11: RB1 out -> Kareem Hunt 80% of snaps, 13.2 PPR
Wk 12: RB1 out -> Kareem Hunt 71% of snaps, 20.0 PPR
Wk 18: RB1 out -> Brashard Smith 53% of snaps, 7.8 PPR
NE: RB1 Rhamondre Stevenson 59% avg, committee weeks 6/17
Wk 9: RB1 out -> TreVeyon Henderson 75% of snaps, 12.7 PPR
Wk 10: RB1 out -> TreVeyon Henderson 84% of snaps, 28.0 PPR
Wk 11: RB1 out -> TreVeyon Henderson 88% of snaps, 32.3 PPR
SF: RB1 Christian McCaffrey 83% avg, committee weeks 0/17
TB: RB1 Bucky Irving 62% avg, committee weeks 5/17
Wk 5: RB1 out -> Rachaad White 80% of snaps, 23.1 PPR
Wk 6: RB1 out -> Rachaad White 77% of snaps, 17.6 PPR
Wk 7: RB1 out -> Rachaad White 90% of snaps, 8.4 PPR
Wk 8: RB1 out -> Rachaad White 55% of snaps, 8.1 PPR
Wk 10: RB1 out -> Rachaad White 72% of snaps, 10.4 PPR
Wk 11: RB1 out -> Rachaad White 60% of snaps, 8.2 PPR
Wk 12: RB1 out -> Rachaad White 60% of snaps, 5.2 PPR
WAS: RB1 Jacory Croskey-Merritt 39% avg, committee weeks 15/17
...
63 RB1-out games: the fill-in averaged 61% of snaps and 12.5 PPR
That summary is the headline: across 63 games without their RB1, the back who led the room averaged 61% of snaps and 12.5 PPR. Twenty of those games reached 15 PPR. The best fill-ins didn't just split the work; they took all of it:
| Fill-in | Team | Starter out | Snap share | PPR |
|---|---|---|---|---|
| TreVeyon Henderson | NE | Wks 9–11 | 75%, 84%, 88% | 12.7, 28.0, 32.3 |
| Kareem Hunt | KC | Wks 9, 11, 12 | 78%, 80%, 71% | 12.5, 13.2, 20.0 |
| Rachaad White | TB | Wks 5–8, 10–12 | 55–90% | 5.2 to 23.1 |
| Kimani Vidal | LAC | Wks 6–11, 13 | 52–93% | 3.0 to 22.8 |
| RJ Harvey | DEN | Wks 11, 13–18 | 47–68% | 4.3 to 22.1 |
Two lessons. First, a handcuff who takes 75%+ is a legitimate starter, and Henderson's 28.0 and 32.3 were league-winning weeks off the bench. Second, even a full workload isn't a steady workload: Rachaad White went from 23.1 to 5.2 across a seven-game run on 55–90% of snaps. The role is the floor you're buying. Touchdowns decide the ceiling.
A surprise in the numbers: the fill-in's line barely depended on how dominant the starter had been. In backfields where the RB1 averaged 60%+ of snaps, fill-ins played 65% and scored 11.8 PPR. Where the RB1 averaged under 60%, they played 60% and scored 12.9. Don't assume only workhorse backfields produce useful handcuffs.
Which backfields are committees?
The same run counts committee weeks (no back at 60% or more). In 2025, 12 teams were committees in at least 9 of 17 games: Arizona, Baltimore, Chicago, Cleveland, Denver, Green Bay, Houston, Kansas City, Pittsburgh, Seattle, Tennessee and Washington. Seattle and Washington hit 15 of 17.
At the other end, San Francisco (Christian McCaffrey, 83% average) and Atlanta (Bijan Robinson, 78%) had zero committee weeks, and Indianapolis (Jonathan Taylor, 83%) had one. Those are the backfields where the starter's absence leaves the biggest hole, and where his backup is the most direct insurance you can buy.
Committee status isn't permanent. Run the script weekly and watch the count move; our weekly usage reports track the same 60% line, and the snap share guide explains why share moves before points do.
How do I run it on the current season?
Run python handcuffs.py 2026. It uses the same 4 requests. The current season's snap rows update the day after games at about 11:00 am ET, so run it Tuesday once Monday night's game is in.
Here's what it found on November 20, 2026, with stats through Week 10 plus Thursday night's Week 11 game. Across 39 RB1-out games so far, the fill-in averaged 55% of snaps and 11.0 PPR. The standout is Kansas City: Kenneth Walker III sat Week 10, and Emmett Johnson took 88% of the snaps for 35.1 PPR. Carolina, Cincinnati, Dallas, Detroit, Indianapolis and Kansas City haven't had a committee week yet; Green Bay (9 of 10) and New Orleans (9 of 9) have barely had a week without one. Early in a season, an "RB1 out" week can also be a starter who opened the year hurt, so read each team's list, not just the average.
Then do three things with the output:
- For your own RB1: look at his team's "RB1 out" history. If the fill-in took 70%+, roster him before the playoffs. If the work went three ways, your insurance is a waiver claim, not a bench spot.
- For everyone else's RB1: backups behind workhorses with a 70%+ fill-in history are the cheapest upside on your wire. Check
percent_ownedon the same rows. - For committees: don't pay for "the handcuff" at all. When nobody in a room clears 60%, the absence of one back mostly spreads snaps around.
Pair this with injury statuses on game day. The daily 11:00 am ET refresh and the Sunday runs before each slate tell you when a starter is ruled out; the injury changes guide shows how to get only the changes.
How many requests does it cost?
Four per season you analyze. Running 2025 and 2026 back to back is 8, and a weekly re-run all season is well inside Basic's 5,000 requests a month ($9/month). The weekly snap and stats rows are not on the free plan, which returns profiles only. Every field used here is in the API reference, and more tools are in the guides.
Frequently asked questions
What is a handcuff in fantasy football?
The backup running back who inherits the starter's role if the starter misses a game. Rostering him insures your RB1, or gives you a cheap starter when someone else's RB1 goes down.
How do you define a backfield committee?
Any week in which no running back played 60% or more of his team's offensive snaps, with fullbacks excluded. In 2025, 12 teams were committees in at least 9 of their 17 games.
How much does a handcuff score when the starter sits?
In the 63 games where a team's RB1 was out in 2025, the back who led the backfield played 61% of snaps on average and scored 12.5 PPR. About one game in three reached 15 PPR.
How many requests does the script use?
Four. /players with position=RB fits every running back into four pages of 100, and each row already carries weekly snap shares and weekly stats for the season you ask about, on Basic and above.