Fantasy Football Cheat Sheet: Quick Stats and Probability Rules for FPL Managers

Beat the noise: a compact FPL cheat sheet that turns injury updates and stats into fast probabilistic decisions

Struggling to turn Premier League injury bulletins and FPL stat tables into confident transfers, captain calls, or chip plays? You’re not alone. Managers and students of statistics often face two linked pain points: a flood of qualitative team news and an incomplete method to convert that news into numbers. This cheat sheet translates common 2025–26 injury language and FPL metrics into short probabilistic heuristics, expected-value rules, and decision tips suitable for both weekend managers and statistics learners.

What you’ll get (quick)

• Simple probability mappings from injury labels (e.g., “ruled out”, “doubtful”) to play/starting chances
• Compact expected-value (EV) formulas using xG/xA and minutes
• Practical decision heuristics for captaincy, transfers, and chip timing
• Worked examples using a Manchester United vs Manchester City injury snapshot from early 2026
• Practice problems for students learning applied probability and stats

The 2026 context: why this matters now

In late 2025 and early 2026, two trends made data-driven FPL decisions more important: clubs increasingly used GPS load-management and surgeon-wide rotation policies, and AI-driven minute-prediction models became public in many community tools. Those changes mean injury notes and manager pressers now correlate even stronger with actual minutes than before — but the language remains qualitative. Turning that language into probabilities is the high-leverage step that separates good managers from great ones.

Key assumption

The heuristics below assume access to per-90 stats (xG, xA, shots-in-box), basic FPL scoring rules, and the latest injury tag from team news (e.g., “ruled out”, “doubtful”, “trained fully”). We also use simple Poisson approximations for goals from xG (standard in sports analytics) — explained in the EV section.

Injury-report to probability mapping: fast rules

Managers publish team news with many phrases. The following mapping is a pragmatic conversion I use in live FPL decision-making and teach in statistics labs.

1. Ruled out / Long-term injured: P_play = 0. (Bench or transfer only.)
2. Out this week, no update: P_play = 0–0.05 (treat as not available unless training update).
3. Doubtful / Doubt: P_play = 0.3–0.6. Default 0.45 if no training report.
4. Late call / Trained but touch-and-go: P_play = 0.6–0.8. Default 0.7.
5. Likely / Expected to play: P_play = 0.8–0.95. Use 0.9 as baseline.
6. Returned from international duty (AFCON, World Cup qualifiers) within 7–10 days: P_play for starting = 0.5–0.65 (higher rotation risk); minutes multiplier = 0.6–0.8 of normal.
7. Back in full training for 3+ days: treat as Likely unless coach says “managing minutes”.

Rule of thumb: if manager uses the phrase “will be assessed”, reduce P_play by 0.15 relative to the label above. If the player completes full training the day before kickoff, increase P_play by 0.1.

Converting FPL stats to expected value (EV): the formula

Expected value is the core statistic for decision-making: it’s the average points you expect from a choice. We compute EV for a player in a single game as:

EV_play = P_play × E_points_if_play

Where E_points_if_play is estimated from per-90 attacking stats and typical scoring weights. Use per-match conversions (per 90 → per match) and minute scaling for non-90 expectations.

Simple model for E_points_if_play

Use this modular approximation:

E_points_if_play ≈ M_factor + P_goal × G_pts + P_assist × A_pts + P_cs × CS_pts + avg_bonus

• M_factor: baseline minutes-related points (e.g., minutes × (appearance points fraction)).
• P_goal: probability of scoring at least 1 goal in the match ≈ 1 − exp(−xG_match).
• P_assist: approximate from xA per match or correlated with shots/expected assists.
• P_cs: clean-sheet chance (defenders/goalkeepers), estimated from the opponent’s xG allowed.
• G_pts, A_pts, CS_pts: the FPL point values for goals, assists, clean sheets (use the per-position point values). Adjust if your competition uses custom scoring.

Poisson trick for goals

If xG for a player in the fixture is 0.45, the probability they score at least one goal ≈ 1 − e^(−0.45) = 0.36. For small xG (<0.6), the approximation P ≈ xG is also acceptable for back-of-envelope math.

Worked example (striker)

Suppose a forward has per-match stats for this fixture:

• xG_match = 0.5
• xA_match = 0.15
• Minutes expectation if starting = 80

FPL weights (standard): goal for forward = 4, assist = 3. Ignore bonus for now.

1. P_goal = 1 − e^(−0.5) ≈ 0.39
2. P_assist ≈ 0.15 (xA small enough to use as approximation)
3. E_points_if_play ≈ (minutes/90 × appearance baseline) + P_goal×4 + P_assist×3
   • Assume appearance baseline (sub/appearance) ~ 1 point for playing (varies) — include 1.
   • E_points_if_play ≈ 1 + 0.39×4 + 0.15×3 ≈ 1 + 1.56 + 0.45 = 3.01

Adjust if the player is doubtful with P_play = 0.45: EV_play = 0.45 × 3.01 ≈ 1.35 expected points from selecting that player this gameweek.

Injury-adjusted decision rules and heuristics

Translate EV into decisions with these rules of thumb you can apply in the 15 minutes you have before the deadline.

Transfer decision: EV differential

When choosing whether to transfer in a player A to replace B, compute the EV difference:

ΔEV = EV_play(A) − EV_play(B)

• If ΔEV > 1.5 expected points, transfer is justified in single-game thinking.
• For long-term transfers (3+ gameweeks horizon using fixture difficulty), require ΔEV > 0.8/gameweek after accounting for fixture swings.
• Factor in hit points: if you take a -4, require ΔEV > 4 across the horizon (e.g., next 1–2 gameweeks).

Captaincy rule

Captain doubles points, so compare EV_captain_target×2 vs EV_current_captain×1 (assuming you will switch if better):

Choose new captain if 2×EV_play(new) − EV_play(current) > 1.2

Why 1.2? It’s a small buffer to account for bonus volatility and variance: captains involve higher variance, so the threshold avoids chasing marginal leads.

Bench/Auto-sub threshold

If a starter’s EV_play < 1.0 (after injury adjustment) and you have a bench player with EV_play > 1.5 and likely to start, consider swapping the bench into the starting XI — especially if you face multiple injury/doubtful flags.

Chip guidance (Wildcard, Free Hit, Bench Boost)

• Bench Boost: Use only when your whole 15 has starters or when you expect multiple rotation-proof starters. If several of your 15 are doubtful, Bench Boost EV plunges.
• Free Hit: Best during heavy rotation weeks (e.g., immediate post-FA Cup/AFCON return weeks in 2026). If you anticipate >4 starters changing due to injuries/rotation, Free Hit may beat a normal team rework.
• Wildcard: If your average EV of bench players rises after transfers (i.e., you can assemble a plan with average EV > 2.2 across 11 players for multiple GWs), wildcarding is statistically sensible.

Applying the cheat sheet to a real snippet (Manchester United v Manchester City, early 2026)

Team news example (simplified):

• Manchester United: Bryan Mbeumo and Amad Diallo back from AFCON, De Ligt out, Shea Lacey suspended, Mazraoui unavailable.
• Manchester City: Nico Gonzalez doubtful after Friday training; John Stones and Oscar Bobb out; several others long-term.

Step 1 — Convert labels to probabilities

• Nico Gonzalez: doubtful → P_play = 0.45 baseline; if reports say “trained” increase to 0.55.
• Mbeumo (returning from AFCON): returned → starting P_play maybe 0.6; minute multiplier 0.6–0.8.
• Players explicitly out (Bobb, Stones, De Ligt): P_play = 0.

Step 2 — Estimate E_points_if_play from xG/xA

Suppose Gonzalez has fixture xG = 0.25 and xA = 0.10. Using Poisson: P_goal = 1−e^(−0.25)=0.22; P_assist≈0.10. If he’s a midfielder (goal=5, assist=3):

E_points_if_play ≈ 1 (appearance) + 0.22×5 + 0.10×3 = 1 + 1.10 + 0.30 = 2.4

Step 3 — Injury-adjusted EV

EV_play = P_play × E_points_if_play ≈ 0.45 × 2.4 = 1.08 expected points. If he trains fully the day before and P_play → 0.55, EV_play ≈ 1.32.

Decision

If your alternative midfielder on the bench has EV_play = 1.6, keep the bench player in starting XI. If you’re choosing a captain and your current captain EV = 4.0, 2×EV(Gonzalez) = 4.8 (but adjusted: 2×1.32=2.64 actual expectation after injury) — don't captain Gonzalez unless his P_play becomes >0.8.

Statistical exercises for students (with answers)

Use these to practice applied probability and EV. All problems assume independent events and Poisson approximations for goals.

Exercise 1 — Poisson conversion

Player A has xG = 0.4 for a fixture. Compute the probability they score at least once. Then compute their expected goal points if they are a midfielder (goal=5).

1. P_goal = 1 − e^(−0.4) ≈ 0.33.
2. Expected goal points = P_goal × 5 ≈ 1.65.

Exercise 2 — Injury Bayesian update

Prior to manager press conference, a player has P_play prior = 0.9 (likely). The manager says “doubtful” in the presser. Treat the presser as new evidence that multiplies odds by 0.5. Compute new P_play.

1. Odds prior = 0.9 / 0.1 = 9.
2. New odds = 9 × 0.5 = 4.5.
3. Posterior P_play = 4.5 / (1 + 4.5) = 0.818 → 81.8%.

Exercise 3 — Hypothesis test (advanced)

Two forwards have mean xG per match 0.55 and 0.35 across 30 matches with SDs 0.4 and 0.35. Test at α=0.05 whether their xG means differ. (Two-sample t-test — sketch.)

Compute t-statistic and compare to critical value ~2.00. Students should perform t-test; expected conclusion: if t > 2.00, reject null.

Advanced strategies (2026): combining AI minute predictions and injury flags

From late 2025, community tools began offering AI-based minute predictions that combine manager rotation history, fixture congestion, substitution tendencies, and live training reports. Use these responsibly:

• Use AI minute predictions as a prior for minutes, then apply the injury probability mapping above as likelihood to produce a posterior minute estimate.
• When AI predicts a 60-minute expected minutes and P_play (availability) = 0.6, scale E_points_if_play by minutes_ratio = (expected_minutes / 90) × P_play.
• Watch for overfitting: tools can overreact to single-sample manager quotes. Smooth predictions with a 0.2–0.3 weight to season-long minutes baseline.

Quick one-page cheat sheet (printable rules)

• Injury tags → P_play: Ruled out=0, Out=0–0.05, Doubtful=0.3–0.6 (use 0.45), Late call/trained=0.6–0.8, Likely=0.8–0.95.
• EV formula: EV_play = P_play × [appearance + 1−e^(−xG)×G_pts + xA×A_pts + P_cs×CS_pts + avg_bonus].
• Transfer rule: ΔEV > 1.5 → one-game transfer; ΔEV > 0.8/gameweek → multi-week swap; require ΔEV > hit for negative points.
• Captain rule: change captain if 2×EV(new) − EV(current) > 1.2.
• Chip rule: Bench Boost only when entire 15 are likely starters; Free Hit when >4 starters impacted by rotation or injury; Wildcard when you can raise average EV across 11 by >0.5/gameweek for several weeks.

Limitations and best practices

These heuristics simplify a complex system. Key caveats:

• FPL points are high-variance: EV is a long-run average — single-match outcomes can deviate greatly.
• Manager language and last-minute team sheets still matter: always re-check for training updates day-before and morning lineups where available.
• Use Bonferroni-style thinking for multiple doubtful flags: the probability that at least one of three doubtful players is out is higher than any single P_play — adjust squad-level EV accordingly.

Actionable takeaways (use before every deadline)

1. Scan team news and map tags to P_play using the cheat sheet rules.
2. Quick-calc E_points_if_play for each problem player using xG/xA and Poisson for P_goal.
3. Compute EV_play = P_play × E_points_if_play and compare alternatives (bench, transfer target).
4. Apply captain rule: only switch if 2×EV(new) − EV(current) > 1.2.
5. Before chips, ensure squad-level starters are rotation-proof; use Free Hit in weeks with >4 uncertain starters.

Final example: a 3-minute pre-deadline checklist

1. Open team news and mark each flagged player with P_play from the mapping.
2. For captain candidates, estimate xG/xA quickly (or use community per-fixture xG), compute EV and apply the captain rule.
3. If a player is doubtful and their EV drops below an available bench starter by >0.6, swap them out.
4. If >3 starters are doubtful or returning from AFCON with rotation risk, consider Free Hit or a conservative bench boost plan.

Where to go next (tools and data)

For live use combine:

• Official club/team news pages and reputable outlets (e.g., BBC Sport) for the earliest flags and training updates.
• Per-fixture xG/xA providers for your simulation inputs (community sites that updated models in 2025–26 are best).
• AI minute prediction tools as a prior, but always apply the injury-probability adjustments above.

Call to action

Ready to put this into practice? Download our concise printable cheat sheet and try the three practice problems above on your next deadline. If you’re a student, run a Monte Carlo simulation for a full gameweek using these rules and compare outcomes with and without injury-adjustments — you’ll see the statistical advantage immediately. Join our newsletter for weekly, data-backed FPL guides tuned to the 2026 season and access downloadable templates that implement these EV rules in Google Sheets.

Related Reading

• After Meta’s Workrooms: Rethinking Virtual Collaboration for Secure File Exchange
• How to Build a Home Fragrance System That Outlasts Your Tech Discounts
• Why Elizabeth Hargrave Designed Sanibel for Her Dad — What That Means for Board Game Collectors
• Solar-Powered Outdoor Entertaining: Speakers, Lighting, and Power for Backyard Gatherings
• How to Use Heated Compresses to Boost Product Absorption and Relax Facial Muscles